• Author / Uploaded
• Emilio Vicente

Citation preview

An Introduction to Metametaphysics

How do we come to know metaphysical truths? How does metaphysical inquiry work? Are metaphysical debates substantial? These are the questions which characterize metametaphysics. This book, the first systematic student introduction dedicated to metametaphysics, discusses the nature of metaphysics – its methodology, epistemology, ontology, and our access to metaphysical knowledge. It provides students with a firm grounding in the basics of metametaphysics, covering a broad range of topics in metaontology such as existence, quantification, ontological commitment, and ontological realism. Contemporary views are discussed along with those of Quine, Carnap, and Meinong. Going beyond the metaontological debate, thorough treatment is given to novel topics in metametaphysics, including grounding, ontological dependence, fundamentality, modal epistemology, intuitions, thought experiments, and the relationship between metaphysics and science. The book will be an essential resource for those studying advanced metaphysics, philosophical methodology, metametaphysics, epistemology, and the philosophy of science. TUOMAS E. TAHKO

is a University Lecturer in Theoretical Philosophy and a

Finnish Academy Research Fellow at the University of Helsinki. He is the editor of Contemporary Aristotelian Metaphysics (Cambridge, 2012) and the author of numerous articles in journals, including Mind, The Philosophical Quarterly, Erkenntnis, and Thought.

An Introduction to Metametaphysics TUOM AS E . TAHKO University of Helsinki

University Printing House, Cambridge CB2 8BS, United Kingdom Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781107434295 © Tuomas E. Tahko 2015 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2015 Printed in the United Kingdom by TJ International Ltd. Padstow Cornwall A catalogue record for this publication is available from the British Library Library of Congress Cataloguing in Publication data

Tahko, Tuomas E., 1982– An introduction to metametaphysics / Tuomas E. Tahko, University of Helsinki. pages cm Includes bibliographical references and index. ISBN 978-1-107-07729-4 (hbk) – ISBN 978-1-107-43429-5 (pbk) 1. Metaphysics. 2. Ontology. 3. Knowledge, Theory of. 4. Science–Philosophy. I. Title. BD111.T28 2015 110–dc23 2015022580 ISBN 978-1-107-07729-4 Hardback ISBN 978-1-107-43429-5 Paperback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Contents

Preface

page ix

1 Why should you care about metametaphysics? 1.1 Metametaphysics or metaontology? 1.2 How to read this book 1.3 Chapter outlines 1.4 Further reading

1 3 6 7 11

2 Quine vs. Carnap: on what there is and what there isn’t 2.1 On what there is 2.2 Plato’s beard 2.3 Enter Meinong 2.4 External and internal questions 2.5 Language pluralism

13 15 18 20 27 35

3 Quantification and ontological commitment 3.1 The meaning of the existential quantifier 3.2 The existential quantifier and ontological commitment

39 41

3.3 Quantifier variance and verbal debates 3.4 Beyond existence questions

49 57

4 Identifying the alternatives: ontological realism, deflationism, and conventionalism 4.1 Ontological realism and anti-realism 4.2 Ontological deflationism 4.3 Towards extreme conventionalism 4.4 A case study: Sider’s ontological realism 4.5 Taking stock

45

64 65 71 76 83 90

v

vi

Contents

5 Grounding and ontological dependence 5.1 Ontological dependence: a fine-grained notion 5.2 Identity-dependence and essential dependence 5.3 Is grounding ontological dependence? 5.4 Formal features of ground

93 94 98 104 106

5.5 Grounding, causation, reduction, and modality 5.6 Grounding and truthmaking

112 116

6 Fundamentality and levels of reality 6.1 The ‘levels’ metaphor 6.2 Mereological fundamentality 6.3 Further specifications: well-foundedness and dependence 6.4 Generic ontological fundamentality 6.5 Fundamentality and physics

120 124 127

7 The epistemology of metaphysics: a priori or a posteriori? 7.1 A priori vs. a posteriori 7.2 Modal rationalism and a priori methods 7.3 The epistemology of essence 7.4 Modal empiricism and the status of armchair methods 7.5 Combining a priori and a posteriori methods

151 152 155 163 167 172

8 Intuitions and thought experiments in metaphysics 8.1 Specifying ‘intuition’ 8.2 Intuitions and experimental philosophy 8.3 Experience-based intuitions 8.4 Rational intuition

177 179 185 188 190

8.5 Scientific thought experiments 8.6 Philosophical thought experiments 9 Demarcating metaphysics and science: can metaphysics be naturalized? 9.1 Autonomous metaphysics 9.2 Fully naturalistic metaphysics

133 136 141

194 197

203 206 211

Contents

9.3 The Principle of Naturalistic Closure and the Primacy of Physics 9.4 Methodological similarities 9.5 Moderately naturalistic metaphysics Glossary Bibliography Index

217 225 231 236 243 255

vii

Preface

Metametaphysical issues, or methodological issues pertaining to metaphysics, have been central in my work for about a decade. My dissertation was called The Necessity of Metaphysics – I have always been optimistic about our ability to overcome the many methodological challenges that metaphysical inquiry faces. Although my views regarding many specific questions have changed over the years, my general attitude towards metametaphysics has remained largely unchanged: I still think that realism is worth defending, that modal epistemology is of particular methodological importance in metaphysics, and that we cannot do metaphysics without relying on at least some a priori reasoning, whatever the correct account of the relationship between a priori and a posteriori turns out to be. In recent years I have taught several courses in Helsinki on the topics of this book. There’s no doubt that preparing for these courses and discussing the material with my students has helped me to better articulate many of the central questions of metametaphysics. Much remains to be done before this young area reaches the conceptual clarity that one might desire, but I hope that this book goes at least some way towards this goal. I would like to express my gratitude to the following people, who read and commented on the material of this book: Hanoch Ben-Yami, Francesco Berto, Matti Eklund, Guglielmo Feis, Marcello Oreste Fiocco, Joachim Horvath, Markku Keinänen, James Miller, Matteo Morganti, Donnchadh O’Conaill, Olley Pearson, Paavo Pylkkänen, and Anand Vaidya. Their valuable feedback saved me from many errors and omissions; any remaining errors are my own. I have discussed the material of this book with too many people to list here as well as presenting the papers related to the book at numerous seminars and conferences. I appreciate the feedback received at these events. I would also like to thank Hilary Gaskin at

ix

"# $%# "&'# &()

x

Preface

Cambridge University Press for support throughout the process of writing this book, as well as an anonymous reader for helpful comments. My greatest debt, however, will always be to my mentor and PhD supervisor, the late E. J. Lowe. In Jonathan’s work I first discovered the metametaphysical attitude that continues to guide my work today. Finally, most of the work for this book has been made possible by various grants from the Academy of Finland. I have drawn on the following previously published and forthcoming material, although it has been extensively reworked for the purposes of this book. In addition, some material in Chapter 9 is based on joint ongoing work with Matteo Morganti. I’d like to thank him for the permission to use that material in this book. ‘A New Definition ofA Priori Knowledge: In Search of a Modal Basis,’ Metaphysica 9.2 (2008), pp. 57–68. ‘A Priori and A Posteriori: A Bootstrapping Relationship,’ Metaphysica 12.2 (2011), pp. 151–164. ‘In Defence of Aristotelian Metaphysics,’ in T. E. Tahko (ed.), Contemporary Aristotelian Metaphysics (Cambridge University Press, 2012), pp. 26–43. ‘Counterfactuals and Modal Epistemology,’ Grazer Philosophische Studien 86 (2012), pp. 93–115. ‘Boundaries in Reality,’Ratio 25.4 (2012), pp. 405–424. ‘Truth-Grounding and Transitivity,’Thought: A Journal of Philosophy 2.4 (2013), pp. 332–340. ‘Boring Infinite Descent,’ Metaphilosophy 45.2 (2014), pp. 257–269. ‘Natural Kind Essentialism Revisited,’ Mind 124.495 (2015), pp. 795–822. ‘Ontological Dependence,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2015 edn); see http://plato.stanford.edu/archives/spr2015/

entries/dependence-ontological/ (with E. J. Lowe). ‘Empirically Informed Modal Rationalism,’ in R. W. Fischer and F. Leon (eds.), Modal Epistemology After Rationalism, Synthese Library (Dordrecht: Springer,

forthcoming). ‘The Modal Status of Laws: In Defence of a Hybrid View,’ The Philosophical Quarterly (forthcoming), doi:10.1093/pq/pqv006.

‘Minimal Truthmakers,’Pacific Philosophical Quarterly (forthcoming), doi:10.1111/ papq.12064 (with Donnchadh O’Conaill.)

1

Why should you care about metametaphysics?

This introductory chapter deals with the motivation for studying metametaphysics and its importance for metaphysics more generally. The relationship between metametaphysics and metaontology is clarified, some guidance for reading the book is given, and chapter outlines are provided. In addition, the chapter contains suggestions for further reading, divided between introductory material and more advanced material. Since you’ve opened this book, it is probably safe to assume that you have an interest in metaphysics. Perhaps you think that metaphysics is an interesting area of philosophy and want to know more about it or maybe you’re a student or a professional philosopher specializing in metaphysics. Alternatively, you might be suspicious of metaphysics and its value or contribution within philosophy (and outside it). Perhaps you think that metaphysics is not a substantial area of philosophy because it focuses on pseudo-problems or merely conceptual, linguistic disagreements. You may be coming to philosophy from another discipline, such as the natural sciences, and you might be suspicious of the methods of philosophy, especially when compared with the rigour of your own discipline. Or perhaps you work in a different area of philosophy, wondering how on earth metaphysicians could possibly justify their outlandish claims about the structure ofreality… All of the above attitudes are metametaphysical attitudes. Just as with any kind of attitude, if you hold a metametaphysical attitude you ought to be able to justify why it is that you hold it. The reason might be simply because you haven’t seen much discussion about what metaphysicians are really up to or of how they think they arrive at their various metaphysical positions. If that’s the case, you’ve opened the right book. If you’re inclined to be dismissive about the value of metaphysics or think that its methods are spurious because you have read all the great metaphysicians and found their work wanting in this regard, you’ve also opened the right book. In 1

2

Why should you care about metametaphysics?

contrast, if you consider metaphysics the heart of philosophy and can’t get enough of it, if you enjoy comparing different theories and judging their relative merits, then – you guessed it – you too have opened the right book. The author of this book is a metaphysician working in the tradition that is usually called analytic metaphysics. The analytic vs. continental distinction is not – the author feels – particularly helpful, but for want of a more descriptive account, it should be made clear that this book is focused on the analytic tradition. The author of this book also has a particular metametaphysical attitude. This attitude is a type of ontological realism, which we will look into in detail later in this book. But as a reader, you should be aware that the author is biased in favour of certain types of realist metaphysics and towards the view that metaphysics does have both intrinsic value and an impact throughout philosophy and the sciences. This is not an uncommon attitude amongst metaphysicians, but it certainly requires justification. However, this is not a research monograph defending a particular position, so space will be given to various positions. Metaphysicians are a defensive lot; they hold their metaphysical views dear and their metametaphysical views perhaps even dearer, despite the fact that they don’t always explicitly express the latter. So you will notice that the present author sometimes takes a defensive attitude. Accordingly, this introduction to metametaphysics is ‘opinionated’ – someone with a more dismissive attitude towards metaphysics would no doubt write a very different account. In any case, since it is still much too early to talk about a fully established set of metametaphysical views, despite certain clear patterns, anyone writing a book on metametaphysics has to make some difficult choices on how to lay out the various positions and indeed even what to call them. Similarly, the precise area that a book on metametaphysics – introductory or otherwise – should cover is certainly open to debate. This book, if anything, errs on the side of covering too much, since at times the reader may feel that the discussion has turned to first-order metaphysics instead of the promised meta-analysis of metaphysics. This is largely because it is very difficult, impossible even, to discuss the various metametaphysical issues without resorting to a battery of examples of first-order metaphysical debates. The reader will soon notice that there are two themes not obviously included under metametaphysics, but discussed extensively throughout this book. They are epistemology and (philosophy of) science. Although it

1.1 Metametaphysics or metaontology?

is true that these topics are not obviously metametaphysical in themselves, it would be difficult to avoid them altogether when discussing metametaphysics. The reason for this is quite simple. A central, perhaps the central question of metametaphysics is: How do we acquire metaphysical knowledge? Here is an alternative formulation of essentially the same question: How does metaphysical inquiry work? These are very clearly epistemic questions, having to do with metaphysical knowledge. Science and its philosophy enter the picture very quickly after these initial questions, for one popular answering strategy to epistemic questions in metaphysics is that metaphysical knowledge and inquiry have something to do with scientific knowledge or inquiry. Of course, not everyone would accept this answer and even if one does, difficult questions remain concerning the exact relationship between metaphysical and scientific knowledge. At any rate, most metaphysicians today would readily propose that there is either some sort of important parallelism or else some continuity between metaphysics and science. At the same time, metaphysics is also one of the last frontiers of philosophy where pure ‘armchair reasoning’ without any connection to experimental methods may seem a perfectly acceptable method of inquiry. So there is also a tension here, one that strongly divides opinions. Given all this, it is difficult to see how any book concerning metametaphysics could remain completely silent about epistemic or scientific matters – this one certainly doesn’t.

1.1 Metametaphysics or metaontology? Most readers interested in metametaphysics are no doubt familiar with another, closely related term, namelymetaontology. The title of this book is a conscious choice: we can distinguish between metametaphysics and metaontology. The usage of these terms is not entirely standardized, but roughly put, it could be said that metametaphysics is the broader of the two terms. More precisely, metametaphysics encompasses metaontology, but covers other issues as well. This type of distinction can also be made between metaphysics and ontology. The term ‘metaphysics’ has an Aristotelian srcin: according to the usual story, the ‘meta’ (‘beyond’, or ‘after’) refers simply to the fact that in certain collected works of Aristotle some works appear after his works concerning physics. So ‘metaphysics’ does not really refer to the content of these works, but rather their

3

4

Why should you care about metametaphysics?

srcinal relative locations. The term ‘ontology’, however, has a more content-oriented Aristotelian srcin, as the Greek

ουτα

(onta) refers to

‘being’. So ontology is the study of being (or being qua being – being as it is in itself, as Aristotle might add). Note however that Aristotle did not use these terms; they have been adopted later on. Ontology emerges as a somewhat more well-defined, albeit extremely general, area of study, whereas metaphysics is typically conceived as concerning reality or the structure of reality, in an even more general sense. The distinction between metaphysics and ontology is, however, vague at best, since many authors use the terms interchangeably. Accordingly, similar vagueness affects the distinction between metaontology and metametaphysics. But what is metaontology? The first systematic use of the term is usually credited to Peter van Inwagen’s 1998 article of the same title. 1 In van Inwagen’s usage, the term ‘metaontology’ has Quinean connotations. Quine considered the central question of ontology to be ‘What is there?’ – something that we will discuss in Chapter 2. But van Inwagen points out that if we wish to consider what it is that we are asking when we say ‘What is there?’, this seems to go beyond ontological questions, hence metaontology. Van Inwagen defines a fairly strict sense of the term:metaontology in Quine’s sense concerns quantification and ontological commitment (these will be discussed in more detail in Chapter 3). This turns out to be a fairly narrow understanding of metaontology, but note that on this definition the metaontological question could be different for someone other than Quine, who might think differently of the task of ontology. In any case, largely because of this srcinal usage of the term, metaontology is typically understood in this relatively narrow sense. In passing, we might note that one alternative understanding of the task of ontology, perhaps closer to the Aristotelian line, would be to give a central position to so-called ‘formal ontology’. This term of art does not refer to ontology conducted with formal methods (although it could involve formal methods); rather, it refers to the study of ontological form, which involves the structures and relations in which ontological elements (such as objects) stand.2 More generally, ontology understood in this fashion involves an examination of the categorical

1 2

Peter van Inwagen, ‘Metaontology,’ Erkenntnis 48 (1998), pp. 233–250. The terminology has Husserlian srcins, see Barry Smith and Kevin Mulligan, ‘Framework for Formal Ontology,’Topoi 3 (1983), pp. 73–85.

1.1 Metametaphysics or metaontology?

structure of reality – a task which goes back to Aristotle’s Categories. One contemporary example of the systematic study of ontological categories in this sense is E. J. Lowe’s four-category ontology.3 Hence, it is not difficult to see that ‘metaontology’ understood from this point of view could amount to something quite different than when understood from the Quinean point of view. Partly for this reason, the title of this book contains the broader term, namely ‘metametaphysics’, for we wish to account for both of these views. One source of confusion regarding the term ‘metametaphysics’ may derive from the fact that the best-known work containing the word in its title – the 2009 Metametaphysics anthology edited by David Chalmers, David Manley, and Ryan Wasserman– is by and large focused on the project of metaontology as van Inwagen defines it (with some exceptions).4 Indeed, the subtitle of the anthology is ‘New Essays on the Foundations of Ontology’. In fact, the terms ‘metaontology’ and ‘metametaphysics’ are also often used interchangeably. But let us attempt to define the term ‘metametaphysics’ as it is used in this book. Metametaphysics =df The study of the foundations and methodology of

metaphysics.

Here, ‘metaphysics’ is understood to encompass ontology, so metametaphysics will also involve the study of the foundations and methodology of ontology. Accordingly, metaontology is to be understood as a subspecies of metametaphysics. Chapters 2 and 3, and to some extent Chapter 4, could roughly speaking be said to concern metaontology in van Inwagen’s sense, although they do not do so exclusively. Subsequent chapters (Chapters 5 to 9) concern the methodology of metaphysics in a much broader sense; they also involve a great deal of epistemology. However, the reader is advised to not put too much weight on these distinctions, as they are indeed vague. The guiding thought in this book is to be inclusive and the suggested definition of metametaphysics certainly allows this. Both terms, ‘metaontology’ and ‘metametaphysics’, are used in this book, roughly in the sense suggested here; that is, metaontology refers primarily to the 3

4

E. J. Lowe, The Four-Category Ontology: A Metaphysical Foundation for Natural Science (Oxford: Clarendon Press, 2006). D. Chalmers, D. Manley, and R.Wasserman (eds.), Metametaphysics (Oxford University Press, 2009).

5

6

Why should you care about metametaphysics?

study of existence, quantification, and ontological commitment, whereas metametaphysics encompasses these areas and also broader issues in the methodology of metaphysics.

1.2 How to read this book This book is aimed at relatively advanced undergraduate and graduate students with at least some prior knowledge of metaphysics and related fields. However, being the first of its kind, it will also prove helpful to professionals working for example on metaphysics, epistemology, or philosophy of science. While some prior knowledge of metaphysics is assumed, prior knowledge of metametaphysics is not necessary. An introductory course in metaphysics ought to be sufficient to follow the book, at least if the reader supplements this book with some of the further material referenced within it. The recommended background is an advanced course in metaphysics and basic knowledge of philosophical logic, although formalism is kept to a minimum. It should perhaps be emphasized that the reader is certainly advised to read some of the primary material referred to in the book, for it is impossible to do justice to all the topics that we will cover. Partly for this reason, a fairly extensive bibliography for an introductory book is provided. Emphasis is given to some of the most recent literature in metametaphysics, with the hope that even experts in the field may find the book useful. The book also includes a glossary with short definitions of some of the most important technical terms. The glossary is not exhaustive and the reader is also advised to consult the index for the full context of each term, but the glossary can be used as a quick reminder. There are no particularly important guidelines regarding the process of reading this book. The book has been written with the assumption that most readers will proceed from the beginning to the end and this is indeed the advisable order for those not very familiar with the topics of the book, but each chapter can certainly be read on its own. Typically, when some prior knowledge of relevant concepts, views, or tools is assumed, this is indicated in the text, with reference to the chapter where the concept/ view/tool was first introduced. More advanced readers should have no trouble jumping ahead to topics that interest them. If the book is used for a course in metametaphysics, the teacher may decide to pick individual

1.3 Chapter outlines

chapters to supplement other material. The same can of course be done with a normal metaphysics course, as many introductory courses in metaphysics now contain lessons on the methodology and foundations of metaphysics. One aspect worth mentioning here is the number of examples from the sciences that the reader of this book will encounter. In many cases, certain metaphysical positions are illustrated with examples from the natural sciences, physics and chemistry in particular. It is assumed that most readers will have some familiarity with many of the examples from previous metaphysics or philosophy of science courses, but they are generally laid out in such a way that no prior knowledge is necessary. There are a few exceptions, though. For instance, certain examples from fundamental physics may be difficult to understand without any prior knowledge of physics, even though they are not presented formally. However, in these cases, the reader will not miss anything absolutely crucial if they decide to skip the more detailed examples. A final note on the system of referencing used. Full bibliographical detail is provided in footnotes and also in the final bibliography. In each chapter, the first reference includes the full bibliographical detail; later instances use the short-title system.

1.3 Chapter outlines A brief outline of each chapter is provided below. The purpose of these outlines is to give the reader a general idea of the topics discussed in each chapter. Note however that technical terms and various ‘isms’ are not defined in the outlines, so the reader should not be too concerned about being able to understand the relevant views – that’s what the chapters themselves are for. Although each chapter can be read on its own, some of them are thematically connected. This is the case especially with Chapters 2 and 3, and partly also Chapter 4. These three chapters focus on metaontology as it was defined earlier, although no particular attempt is made to stay strictly within metaontology. Chapters 5 and 6 are somewhat technical, as they introduce the metaphysician’s ‘toolbox’ – concepts and tools of formal ontology that are used in metaphysics and metametaphysics. Both chapters also apply these tools. Chapters 7 and 8 turn to epistemic matters, discussing the methods of metaphysical inquiry. Chapter 9 concludes

7

8

Why should you care about metametaphysics?

the book with a discussion of the relationship between metaphysics and science, taking advantage of much of the material of the earlier chapters.

Chapter 2: Quine vs. Carnap: on what there is and what there isn’t The historical srcins of metametaphysics are typically traced back to the debate between W. V. Quine and Rudolf Carnap in the 1940s and 1950s. In this chapter, an overview of that debate will be given, but the historical details and the srcinal context of the debate will not be the main subject. One central topic is the status of existence questions such as ‘Do numbers exist?’ Are such questions substantial or merely conceptual, to be settled by linguistics rather than genuine metaphysics? Carnap was famously sceptical about the metaphysical import of such questions, arguing that there is nothing substantial at stake when we ask such questions. The resulting view is a type of language pluralism, according to which we can choose our ontological framework – our preferred language – liberally. Alexius Meinong’s views on the matter and the problem of ‘Plato’s beard’ – dealing with non-existence – will also be discussed. Carnap’s distinction between internal and external questions is outlined and some of its modern applications discussed.

Chapter 3: Quantification and ontological commitment This chapter continues to discuss existence questions, but the focus shifts towards quantification: the status and meaning of the existential quantifier, including its history and name, are discussed. In particular, the question of the Quinean criterion of ontological commitment, according to which we are ontologically committed to those entities that we quantify over, is critically examined, also with reference to its modern counterparts. Moreover, the possibility of so-called ‘quantifier variance’ is discussed, as defended by Eli Hirsch and opposed by Ted Sider, among others. Quantifier variance is the thesis that there is no single (best) quantifier meaning. The thesis is closely related to Hirsch’s view that ontological debates concerning physical objects are ‘merely verbal’. Finally, Kit Fine’s alternative metaontological position, which attempts to undermine the importance of existence questions, is discussed.

1.3 Chapter outlines

Chapter 4: Identifying the alternatives: ontological realism, deflationism, and conventionalism This chapter surveys various metametaphysical positions, some of which have already been discussed in previous chapters. The main contenders are ontological realism, ontological anti-realism, deflationism, and conventionalism. It will become clear that some terminological clarification is needed in order to correctly identify the various subspecies of these views. The debate concerning quantifier variance between Hirsch and Sider will be discussed again, but from a slightly different point of view. Sider’s version of ontological realism will receive further attention and is considered as a case study, with reference to an example from physics.

Chapter 5: Grounding and ontological dependence It is time to introduce some further metaphysical tools: grounding and ontological dependence. The notion of ‘ground’ stormed into contemporary analytic metaphysics at the beginning of the twenty-first century, but the roots of the notion go all the way to Aristotle. At its simplest, grounding may be understood as ‘metaphysical explanation’. To be more precise, when some x is grounded in some y, it is usually thought that y explains x. On the face of it, grounding expresses a relation of ontological dependence. Ontological dependence is a family of relations and different versions of dependence will be discussed in some detail. The question whether grounding is indeed a version of ontological dependence or not will also be examined. The formal features of ground and some related notions as well as applications are outlined. These include causation, reduction, modality, and truthmaking.

Chapter 6: Fundamentality and levels of reality This chapter concerns the view that reality comes with a hierarchical structure of ‘levels’. This type of view has a long history and it remains very popular. Our everyday experiences as well as scientific practice seem to strongly support such a view, since scale is a major factor in both of them. Usually, the reference to scale becomes apparent when talking about parts and wholes – which are studied in mereology: we talk about subatomic

9

10

Why should you care about metametaphysics?

particles constituting atoms, atoms constituting molecules, and molecules constituting everything we see around us. We can express this in terms of ontological dependence, which is covered in Chapter 5: the wholes depend for their existence on their parts. Fundamentality comes in when we ask whether there is an end or a beginning to this hierarchical structure, or equivalently to the relevant chain of dependence. Much of the discussion in this chapter will concern the analysis of ‘metaphysical foundationalism’, which states that there is an end to the chain of dependence, and ‘metaphysical infinitism’, which states that chains of dependence can continue infinitely. These views are also discussed with reference to physics.

Chapter 7: The epistemology of metaphysics: a priori or a posteriori? The epistemology of metaphysics, which is the topic of this chapter, is a broad area. The discussion starts from the a priori vs. a posteriori distinction, which turns out to be more controversial than one might have thought. Various options to clarify the distinction are considered. The bulk of the chapter deals with modal epistemology: our knowledge of possibility and necessity. This will be our case study of the epistemology of metaphysics. Much of metaphysical knowledge seems to involve modal elements, so we need an account of how we are able to acquire modal knowledge. The two main competitors here are ‘modal rationalism’ and ‘modal empiricism’. At first glance, they seem to reflect the a priori vs. a posteriori distinction regarding the source of modal knowledge, but the situation is more complicated than that, as ‘pure’ a priori or a posteriori knowledge appears to be scarce. Therefore, a view according to which both types of knowledge are needed becomes somewhat attractive. Such a view and its prospects are studied.

Chapter 8: Intuitions and thought experiments in metaphysics This chapter continues on epistemic themes. Intuitions and thought experiments are considered important sources of metaphysical knowledge, but there is much controversy surrounding them: how reliable are they as sources of evidence? One problem is that often it is not clear what is even meant by ‘intuition’. This chapter examines a variety of ways to understand metaphysical intuitions and their role in metaphysical inquiry. These

1.4 Further reading

include ‘rational intuitions’, which are typically associated with a priori faculties, and experience-based intuitions. The chapter also discusses the relationship between thought experiments and intuitions as well as the differences and similarities between philosophical and scientific thought experiments.

Chapter 9: Demarcating metaphysics and science: can metaphysics be naturalized? This chapter concerns the relationship between science and metaphysics. There are a variety of options in this regard. One of them is that metaphysics is autonomous and able to tell us something about the world on its own, despite the complications introduced by, say, modern physics. On this view, metaphysical inquiry into the fundamental structure of reality can uphold scientific realism. Another option is at the other extreme: fully ‘naturalistic’ metaphysics. According to this view, metaphysics cannot hope to say anything about reality independently of science. There are also less extreme versions of this view. One such version is due to James Ladyman and Don Ross; the view can be outlined by discussing two principles: the Principle of Naturalistic Closure and the Primacy of Physics . The upshot of these

principles is that metaphysics has a role to play, but it is severely limited, primarily unificatory. The possibility of building a methodological bridge between science and metaphysics is also examined: even if the subject matter of the two disciplines is distinct, perhaps there are some similarities in

their method? Finally, a more modest, reconciliatory view of the relationship between science and metaphysics will be proposed: ‘moderately naturalistic metaphysics’.

1.4 Further reading Although at the time of writing this is the only textbook dedicated to metametaphysics, there are certainly important discussions of metametaphysics, metaontology, and the methodology of metaphysics in various textbooks and collections of papers. Themes relevant to metametaphysics are also discussed, sometimes extensively, in books dealing with philosophical methodology or metaphilosophy more generally. The following is a list of such material that, while far from complete, may

11

12

Why should you care about metametaphysics?

prove useful to the reader of this book. The material is divided between introductory and more advanced material. Here the focus is exclusively on books, but the reader is encouraged to browse the final bibliography for further material.

Introductory material Francesco Berto and Matteo Plebani, Ontology and Metaontology: A Contemporary Guide (London: Bloomsbury, 2015).

Chris Daly, Introduction to Philosophical Methods (Peterborough, ON: Broadview Press, 2010). Alyssa Ney, Metaphysics: An Introduction (Abingdon: Routledge, 2014). S. Overgaard, P. Gilbert, and S. Burwood, An Introduction to Metaphilosophy (Cambridge University Press, 2013). David Papineau, Philosophical Devices: Proofs, Probabilities, Possibilities, and Sets (Oxford University Press, 2012).

More advanced material A. R. Booth and D. P. Rowbottom (eds.), Intuitions (Oxford University Press, 2014). Albert Casullo and J. C. Thurow (eds.), The A Priori in Philosophy (Oxford University Press, 2013). D. Chalmers, D. Manley, and R. Wasserman (eds.),Metametaphysics (Oxford University Press, 2009). Fabrice Correia and Benjamin Schnieder (eds.), Metaphysical Grounding: Understanding the Structure of Reality (Cambridge University Press, 2012).

R. W. Fischer and Felipe Leon (eds.), Modal Epistemology After Rationalism, Synthese Library (Dordrecht: Springer, forthcoming). Matthew C. Haug (ed.), Philosophical Methodology: The Armchair or the Laboratory? (Abingdon: Routledge, 2014). Matteo Morganti,Combining Science and Metaphysics(New York: Palgrave Macmillan, 2013). T. E. Tahko (ed.), Contemporary Aristotelian Metaphysics (Cambridge University Press, 2012). Timothy Williamson, The Philosophy of Philosophy (Oxford: Blackwell Publishing, 2007).

2

Quine vs. Carnap: on what there is and what there isn’t

The historical srcins of metametaphysics are typically traced back to the debate between W. V. Quine and Rudolf Carnap in the 1940s and 1950s. In this chapter, an overview of that debate will be given, but we will not dwell on the historical details or the srcinal context of the debate.

1

Roughly put, the whole metametaphysics debate is sometimes reduced to the question of whether existence questions such as ‘Do numbers exist?’ have any deep meaning or, on the other hand, are merely conceptual, to be settled by linguistics rather than genuine metaphysics. Carnap is famously considered as having been sceptical about the metaphysical import of such questions, arguing that there is nothing substantial at stake when we ask them. The resulting view is a type of

language plural-

ism , according to which we can choose our ontological framework

– our

preferred language – liberally. 2 This, at any rate, is symptomatic of the view of some contemporary philosophers who identify themselves as

1

2

Readers interested in a more historically oriented account may refer to Matti Eklund, ‘Carnap’s Metaontology,’ Noûs 47.2 (2013), pp. 229–249. As Eklund makes clear, the typical narrative of the Quine–Carnap debate is highly controversial. But to simplify matters, we will focus on the way that the debate has been presented in much of contemporary literature. We use the notion of ‘language pluralism,’ following Eklund’s ‘Carnap’s Metaontology,’ for what is sometimes called ‘ontological pluralism’; compare with Matti Eklund, ‘Carnap and Ontological Pluralism,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.), Metametaphysics (Oxford University Press, 2009), pp. 130–156. This is to distinguish Carnap’s view from a different view which is, somewhat confusingly, also sometimes called ‘ontological pluralism’. This more recent usage of the term refers to the idea according to which there are many ‘ways of being’. We will return to this issue (in passing) in Chapter 6. See also Kris McDaniel, ‘Ways of Being,’ in Chalmers, Manley, and Wasserman (eds.), Metametaphysics, pp. 290–319.

13

14

Quine vs. Carnap: on what there is and what there isn’t

(neo-)Carnapian. 3 As a result of the dismissive view of metaphysics associated with Carnap, Quine has emerged as the somewhat unlikely hero in defence of the possibility of metaphysics. Quine’s famous article ‘On What There Is’ (see Section 2.1 ) is considered the definitive work against the Carnapian view that ontological questions are a matter of (linguistic) preference. 4 Quine’s article is commonly considered to have once again made ontology a respectable discipline after the devastating effects of logical positivism, whose proponent Carnap was. However, it should quickly be pointed out that Quine’s attitude towards ontology in the context of the present book would seem to be much more

deflationary –

that is, leaning towards the view that ontological questions are not substantive – than the typical narrative of the debate between Quine and Carnap might suggest. The various options in this regard are discussed in Chapter 4. In any case, there are good reasons to think that the disagreement about ontology between these two figures was not as deep as it is typically played out to be, one of these reasons being that Carnap himself apparently did not think that there is a deep disagreement about ontology between Quine and himself.

5

Be that as it may, the influence of

their debate for contemporary metametaphysics cannot be denied:

the

definitive Metametaphysics anthology edited by David Chalmers, David Manley, and Ryan Wasserman , which contains many of the modern classics of the literature, focuses almost exclusively on Quine and Carnap, with a few notable exceptions. 6 There are a number of interesting aspects in the contemporary literature surrounding the Quine–Carnap debate. We cannot discuss all of them, but perhaps the most central one is the role of existence questions in ontology. Another is the role and interpretation of the existential quantifier. We will postpone some of the relevant discussion until the next chapter, which focuses on ontological commitment, as we will need some background

3

4

5 6

For a prominent example, see Eli Hirsch, ‘Physical-Object Ontology, Verbal Disputes, and Common Sense,’ Philosophy and Phenom enological Research 70 (2005), pp. 67–97. Note however that Hirsch limits his discussion to composite physical objects. W. V. Quine, ‘On What There Is,’The Review of Metaphysics 2 (1948), pp. 21–38; reprinted in his From a Logical Point of View (Cambridge, MA: Harvard University Press, 1980), pp. 1–19. Again, see Eklund, ‘Carnap’s Metaontology.’ Chalmers, Manley, andWasserman (eds.), Metametaphysics.

2.1 On what there is

information before we engage with the topic in more detail. That background will be provided in this chapter, but we will also get to engage with some of the more controversial aspects of the Quine–Carnap debate, including the contemporary incarnations of their respective views. This chapter starts with a brief recap of the role of Quine’s famous article in the first section and continues with a discussion of one of its central themes, the problem of ‘Plato’s beard’, in the second section. After that, in the third section, we make a small detour and consider the views of Alexius Meinong, whose import on the metaontology debate is often ignored. We will see that Meinong’s work is very much relevant both for the Quine–Carnap debate and for a number of contemporary views. In thefourth section we will introduce Carnap’s distinction betweenexternal and internal questions, which also survives, in one form or another, in many contemporary views. From this, (neo-)Carnapian language pluralism emerges, to be discussed in the fifth section.

2.1 On what there is If ontology is the study of what exists, then it would seem that the answer to the ontological question is not very difficult: everything exists. This is the rather anticlimactic starting point of Quine’s ‘On What There Is’. It might appear that this answer is immediately unsatisfactory. For one thing, what are we to say of Sherlock Holmes and other such entities that, on the face of it, do not exist, but nevertheless appear to be something? This is a topic which we will pick up below, but before that we ought to spend a moment discussing the underlying motivation of Quine’s question. In particular, is it really the purpose of ontology simply to list all the things that there are? Surely not: for instance, we are also interested in the relationships between existing things. Yet, it would not be fair to Quine to reduce his position to the over-quoted opening of his article. The important metaontological aspect of the Quinean position is the thesis that being is the same as existence, that is, there are no things that do not exist. Peter van Inwagen, who is a card-carrying Quinean, summarizes the thesis as follows: This thesis seems to me to be so obvious that I have difficulty in seeing how to argue for it. I can say only this: if you think that there are things that do not exist, give me an example of one. The right response to your

15

16

Quine vs. Carnap: on what there is and what there isn’t

example will be either, “That does too exist”, or “There is no such thing as that”.7

On this view, saying that ‘There are numbers’ is on a par to saying ‘Numbers exist’. Van Inwagen’s statement serves to illustrate the key problem that Quine focuses on, namely the status of things that supposedly do not exist, but are nevertheless talked about or referred to. Aswe’ll see in the next section, this problem of ‘nonbeing’ continues to puzzle metaphysicians. Another important aspect of the Quinean view, which van Inwagen lists as one of the Quinean metaontological theses, is that all things that exist do so in the same fashion,univocally. So when we ascribe existence to material objects such as tables and chairs and to abstract objects such as numbers or sets, we supposedly mean the same thing by ‘existence’. The supposed argument in favour of this view is simple:it is the same thing, or at least close enough, to say that unicorns do not exist and that the number of unicorns is zero. This connection between existence and number, says van Inwagen, points towards univocalism about existence. Note however that one could consider the univocalism in question to concerncounting rather than numbers (and existence):If we say that the number of apples in the basket is five and that eating an apple is one of your ‘five a day’ (portions of vegetables or fruit), are we using the numbers in exactly the sameway? A third point about ‘existence’, which we’ll mention only in passing here, is that the Quinean typically considers the existential quantifier ‘ ∃’ to capture the presumed univocal sense of existence. We will return to issues surrounding quantification and the interpretation of the existential quantifier in the next chapter, but for now we can take the Quinean picture at face value. The upshot of this picture – the Quinean method – is that the core questions of ontology can be answered with a simple formula: (1) Take your best scientific theory and assume that what this theory says is true. (2) Translate the sentences of your theory into a formal language, typically first-order predicate logic. (3) The domain of (existential) quantification in your translated theory will give you the ontological commitments of that theory.

7

Peter van Inwagen, ‘Meta-ontology,’ Erkenntnis 48 (1998), p. 235.

2.1 On what there is

Note the mention of scientific theories here; for Quine, the central task of philosophy is to assess and assist scientific theory choice – to determine the ontological commitments of scientific theories. Of course, this three-step formula is certainly not the whole of Quine’s import on the topic, but merely a rough simplification. While we are on the topic of Quine’s ‘On What There Is’, a further issue to bring into attention is that in the famous article in question, Quine was perhaps more interested in defending his version of nominalism rather than a particular methodological or metaontological point. Quine wanted to show that when we say that there are red houses, red roses, and red sunsets, we do not need to commit to the existence of any universal redness, which all of these red things exemplify. This is simple to show by following the method described above: when we translate ‘Socrates is mortal’ into a formal language, we get ‘∃x M(x)’, where ‘M’ stands for the predicate ‘mortal’ and the value of variable ‘ x’ is ‘Socrates’. Here the domain of the existential quantifier includes Socrates, which we are quantifying over (in this formula, ‘Socrates’ is a bound variable). But we are not quantifying over ‘mortal’ – to express a commitment to the universal ‘mortality’ we would have to resort to second-order logic.8 Now, Quine himself wishes to strongly resist this type of quantification over predicates, where the predicate is understood as ranging over things such as properties or universals. As we will see, Quine strives to keep his ontology as sparse as possible, which means that quantification over such suspicious entities as universals, for instance, is to be avoided at all costs. Moreover, Quine has independent reasons to avoid ascending to second-order logic – he famously considered second-order logic to be set theory in disguise, or somewhat more poetically, ‘set theory in sheep’s clothing’.9 What Quine meant by this is that second-order logic is not logic, properly speaking, but rather a branch of mathematics (namely a part of set theory). On Quine’s understanding, first- and second-order quantification turn out to have very different statuses. The first quantifies over objects, the existence of which is 8

9

For more on second-order logic, and to see how this could be done, see Herbert B. Enderton, ‘Second-order and Higher-order Logic,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Fall 2012 edn); see http://plato.stanford.edu/archives/fall2012/ entries/logic-higher-order/. W. V. Quine, Philosophy of Logic, 2nd ed. (Cambridge, MA: Harvard University Press, 1986), Ch. 5.

17

18

Quine vs. Carnap: on what there is and what there isn’t

not in any doubt. But the second purports to be much more powerful, as it quantifies over the realm of universals, properties, concepts … The existence of these more abstract things can, however, be doubted, and Quine himself was particularly doubtful about them. It should beadded though that Quine’s reason to resist quantification over abstract entities derivesfrom his view regarding the role of quantification itself, namely that quantification is intimately linked to ontological commitment. More precisely, the quantified sentences of a theory express the ontological commitments of the theory. We have arrived at the heart of one of Quine’s most famous slogans: ‘To be assumed as an entity is, purely and simply, to be reckoned as the value 10 of a [bound] variable’. This is, in sum, what Quine considers to constitute

ontological commitment. We will assess the Quinean criterion of ontological commitment primarily in thenext chapter; let us now turn to the problem of nonbeing, which continues to enjoy a central role in the metaontology debate.11

2.2 Plato’s beard Besides an unwarranted commitment to universals, Quine was concerned about what has come to be known as the Platonic riddle of nonbeing: Nonbeing must in some sense be, otherwise what is it that there isnot? This tangled doctrine might be nicknamed Plato’s beard; historically it has proved 12 tough, frequently dulling the edge of Occam’s razor.

The historical srcin of this puzzle is Plato’s Sophist, where the views of Parmenides are discussed. It is proposed that to say something or to think something is to speak or thinkof something. The idea is that when Plato thinks or says that Pegasus does not exist, he must be thinkingof Pegasus. Hence, there exists something that Plato is thinkingof, namely Pegasus! The upshot, aptly summarized by Christie Thomas, is that ‘The problem of nonbeing has the perplexing consequence that negative existence claims quite

10 11

12

Quine, ‘On What There Is,’ p. 13. For further discussion on ontological commitment and Quine’s criterion, see also Philip Bricker, ‘Ontological Commitment,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2014 edn); see http://plato.stanford.edu/archives/win2014/entries/ ontological-commitment/. Quine, ‘On What There Is,’ p. 2.

2.2 Plato’s beard

13 generally seem to require the being of the very objects they deny’. Occam’s

razor is supposed to lead us towards a sparser ontology by denying the existence of unwanted entities, but the doctrine seems to be in trouble ifthat very denial commits us to the existence of those unwanted entities. The sparse ontology that Quine prefers is appropriately illustrated with his metaphor of a ‘desert landscape’. The idea is exactly that of Occam’s razor, namely that we should avoid making ontological commitments beyond those that are necessary. For Quine, the desire for desert landscapes is closely connected with the project of avoiding the apparent ontological commitments of ordinary language by paraphrasing – expressing the problematic sentence in an ontologically non-committing manner, where appropriate. Quine’s proposed solution to the riddle of nonbeing relies on a paraphrasing strategy familiar from Bertrand Russell’s theory of descriptions. Quine wants to steer well clear of the numerous alternative solutions that he discusses, such as unactualized possibilities and impossible entities – these would lead us towards a more bloated ontology instead of the desertified one that he prefers. By employing Russell’s theory, Quine can paraphrase problematic statements in such a way that no commitment to non-existing things is required. For instance, the statement ‘The present King of France is bald’. can be paraphrased as ‘Something is the King of France and is bald, and nothing else is the King of France’. This solution relies on Russell’s theory of definite descriptions. Here the definite description is ‘The present King of France’, but note that this description works quite differently from a proper name such as ‘Socrates’ in the sentence ‘Socrates is bald’. The resulting logical forms for these sentences are thus also different, which enabled Russell to solve various puzzles.14 Quine believed that here we also have the tools to solve the Platonic riddle of nonbeing. On Quine’s view names, descriptions or even predicates do not entail ontological commitment; only quantifying over something does. So if you can paraphrase away the quantification over non-desirables like unicorns or, in Quine’s case, universals, then your theory can be maintained without 13

14

Christie Thomas, ‘Speaking of Something: Plato’s Sophist and Plato’s Beard,’ Canadian Journal of Philosophy 38.4 (2008), p. 633. Thomas goes on to provide a historically detailed analysis of this case, arguing that Plato himself need not be committed to the existence of things like Pegasus, and hence: ‘Plato’s beard is not quite as tough on Ockham’s razor as Quine and others have led us to believe,’ p. 667. See Bertrand Russell, ‘On Denoting,’ Mind 14 (1905), pp. 479–493.

19

20

Quine vs. Carnap: on what there is and what there isn’t

a commitment to the existence of such problematic entities. Moreover, Quine suggests that those who do wish to commit to such entities can very easily do this: by saying that there is something (using existential quantification and a bound variable), which Socrates and Aristotle have in common, namely that they are mortal, we may express a commitment to the universal ‘mortality’. Similarly, we commit ourselves to the existence of numbers by saying that there is something which is a prime number larger than a million. The point that Quine insists on is that there is no other way of expressing ontological commitments than this use of quantification and bound variables.

2.3 Enter Meinong Before moving on to Carnap, we would do well to spend a moment discussing the often ignored but important contribution of the Austrian philosopher Alexius Meinong to the debate at hand. Meinong was active before Carnap and Quine and is best known from Russell’s critical commentary, but his influence is in fact very much present in Quine’s ‘On What There Is’. Quine’s article discusses the views of two ‘fictional’ characters, McX and Wyman, who bear remarkable similarity to John McTaggart and Meinong, respectively. Wyman is the more radical of the two: Wyman, Quine says, ‘is one of those philosophers who have united in ruining the good old word “exist”’.15 Meinong is not in agreement with Quine about ‘there is’ or ‘there are’ and ‘exists’ being univocal. For Meinong, there are some things, such as Pegasus, that do not exist. But, of course, Quine does not want to say that there are things that do not exist. Instead, Quine insists that there is no such thing as Pegasus. So the debate is exactly about the Platonic riddle of

nonbeing: Meinong thinks that some things are merely unactualized – they do not have the property of existence – but we can still say that, in some sense, there are such things as Pegasus. One way to illustrate the disagreement between Quine and Meinong, then, is to ask whether existence is a real property – something that an object can either have or lack. This distinction can be further analysed with the help of Meinong’s distinction between Sein – the existential status of entities – and Sosein – the properties (or ‘nature’) that an entity has. The distinction is at the root of Meinong’s 15

Quine, ‘On What There Is,’ p. 3.

2.3 Enter Meinong

principle of independence, which states, roughly, that an entity may have any

properties whatsoever, independently of its existence. According to this principle, non-existent objects can have all the properties that we ascribe to them: we can (truly) say that non-existent golden mountains are made of gold and that round squares are round. For Quineans everything exists, for Meinongians not everything exists. Meinong himself probably considered existence itself to be ‘said in many ways’, that is, to be equivocal rather than univocal, but many later Meinongians would agree with Quineans about there being just one sense of existence. So there are several different aspects to the debate between Quine and Meinong on the one hand and contemporary Quineans and Meinongians on the other. Similarly, where Quine rejects the distinction between ‘there is’ and ‘exists’, Meinong himself might endorse this distinction and make use of it in addressing the problem of nonbeing, but not all contemporary Meinongians would do so. For instance, Graham Priest, whose views we’ll discuss in more detail below, would be closer to Quine on this particular point even though his view can be called a type of Meinongianism.16 Accordingly, we should keep in mind that Meinong’s srcinal position, which is partly open to interpretation, may be different from the position of some contemporary philosophers who have adopted the label ‘Meinongian’. The same is no doubt true of Quine and ‘Quinean’. But the core of the debate between these various versions of Quineanism and Meinongianism is the disagreement about the status of existence as a property that can be distinguished from the object itself – this is an issue where Priest would be squarely in the Meinongian camp. Leaving the historical details aside, we may distinguish between various further positions regarding existence and quantification among contemporary Meinongians. For instance, some would consider existence to be univocal, with Quine, but some would deny this. Some would distinguish between different senses of ‘being’ for existents and non-existents, whereas some would deny this distinction and insist that non-existents have no being whatsoever. Let’s borrow a passage from David Lewis to further illustrate the various options that now emerge; here is Lewis on ‘controversial entities’ that some would consider to exist and some would not:

16

Graham Priest, Towards Non-Being: The Logic and Metaphysics of Intentionality (Oxford University Press, 2005).

21

22

Quine vs. Carnap: on what there is and what there isn’t

These controversial entities include past and future things, the dead who have ceased to be and those who are not yet even conceived; unactualized possibilia; universals, numbers, and classes; and Meinongian objects, incomplete or inconsistent or both. An expansive friend of the entities who says that all these entities exist may be called an allist. A tough 17

desert-dweller who says that none of them exist may be called a noneist.

At first glance, it might seem that Meinong is an allist on Lewis’s terminology, since he says that thereare all sorts of controversial entities. But note that Meinong does not claim that these entities exist. This is in fact a key point for Meinong: there are different ways for different kinds of entities to be. Some entities exist, but some of them only ‘subsist’. This ‘subsistence’ is ascribed to something that does not exist, that perhaps is outside of being altogether. But for Meinong, the being or nonbeing of an entity is not part of its very nature. Meinong has another principle that captures this idea: the principle of indifference. This principle says that an entity isindifferent to being by its very

nature, even though in any given case the entity is either in a state of being or of nonbeing. However, the entity’s nature may dictate its nonbeing:a round square has a Sosein (nature) which dictates that it could never exist, but the principle of indifference states that the entity’s nature does not include its being or nonbeing. All this opens up a number of puzzles, some of which would require discussing and interpreting Meinong much more thoroughly than we can do here. Hence, we may be better off relying on some of the interesting secondary literature that Meinong’s work has inspired. In the secondary literature, Meinong sometimes appears to come closer to a type of noneism, as described by Lewis. That seems at least half right; or rather, as Richard Woodward puts it, ‘Noneists hold that Meinong was half right: right to distinguish existent objects from non-existent ones, but wrong to distinguish those non-existents that have being (the ‘subsistent’) from those that do not’.18 Allists, on the one hand, are with Quine at least to the extent that they hold ‘existence’ to be univocal. On the other hand, as we saw earlier, Quine famously prefers desert landscapes when it comes to ontology, whereas the allists embrace any and all of the controversial 17 18

David Lewis, ‘Noneism or Allism?’, Mind 99.393 (1990), p. 23. Richard Woodward, ‘Towards Being,’ Philosophy and Phenomenologica l Research 86.1 (2013), p. 183. See also Tatjana von Solodkoff and Richard Woodward, ‘Noneism, Ontology, and Fundamentality,’ Philosophy and Phenomenological Research 87.3 (2013), pp. 558–583.

2.3 Enter Meinong

entities. In contrast, the noneists may be desert-dwellers when it comes to existence, but that does not compel them to accept the Quinean doctrine of ontological commitment, which emphasizes quantification. Instead, a noneist could hold that we can quantify over non-existents without much trouble.19 Hence, it seems that the question of being vs. existence is separable from the riddle of nonbeing, as suggested by Tim Crane: ‘whether there is a distinction between being and existence is irrelevant to the truth of the claim that there are things that do not exist’.20 Meinongians would say that controversial as many of the entities mentioned by Lewis may be, that does not mean that they are unable to have properties. For instance, Pegasus has a certain set of properties – being horse-like and having wings in particular – but lacks the property of existence. The Quinean objections to this idea are two-fold: either the Meinongians have misunderstood the notion of existence altogether or, if there is a genuine ontological disagreement here, then the Quinean position is obviously correct, since the Meinongian position is nearly self-contradictory. So the Quinean reaction is rather severe. Quine raises the objection that the Meinongian doctrine of unactualized possibilities is highly implausible, citing examples like the round square cupola on Berkeley College or other such logically impossible entities. The Meinongian will acknowledge that such entities are indeed impossible, but insists that this only means that they could not exist, not that there are no such (impossible!) entities. This is of course a significant price to pay, although the idea of impossible entities – or entire impossible worlds – may not be completely alien to us. 21 Moreover, on the face of it, the Meinongian line enables us to show that anything whatsoever exists: we can take any set of properties X and in accordance with the Meinongian principle of independence it appears that there could be an object with those properties. Take the

properties of ‘being golden’, ‘being a mountain’, and ‘existence’ – if there is an entity with these properties, then we may conclude that a golden mountain exists! 19

20 21

See Richard Routley, ‘On What There Is Not,’ Philosophy and Phenomenological Research 43 (1982), pp. 151–578 and Priest, Towards Non-Being. See also David Lewis’s commentary on Routley in Lewis, ‘Noneism or Allism?’ Tim Crane, The Objects of Thought (Oxford University Press, 2013), p. 24. For discussion, see for instance Daniel Nolan, ‘Impossible Worlds:A Modest Approach,’ Notre Dame Journal of Formal Logic 38.4 (1997), pp. 535–572.

23

24

Quine vs. Carnap: on what there is and what there isn’t

The Meinongian position has certainly not enjoyed as much support as the Quinean position. This may be partly due to prejudice going back to the seemingly devastating criticisms of Meinong’s position by Russell and Quine. Yet contemporary Meinongianism is at least as sophisticated as contemporary Quineanism.22 In fact, Meinongian metaontology and related positions, such as noneism, have only recently started to receive the attention they undoubtedly deserve. Both of the above-mentioned problems for Meinongianism had already been raised by Russell, and they are indeed genuine problems for Meinongians, but there have been some interesting attempts to address them.23 One of these attempts, building on the work of Terence Parsons, is based on identifying so-called ‘nuclear properties’ – a privileged set of properties which is allowed in our set X of properties that an entity could satisfy. Importantly, ‘existence’ is not among such nuclear properties; rather, it is an ‘extranuclear’ property. The obvious problem that follows is how to specify the nuclear/extranuclear distinction, so the Meinongian is not home free yet. Richard Routley’s ‘characterization postulate’ makes a similar attempt, suggesting that objects have the characterizing properties used to characterize them. But Routley’s approach likewise suffers from arbitrariness, for it is not clear who is supposed to determine these characterizing properties. There are, however, further strategies. For instance, Priest revises the nuclear or characterizing properties approach, proposing that the non-existent objects have these properties in the possible (or impossible) worlds where they exist.

24

Further, some properties, such as

being a golden mountain, are ‘existence-entailing’. What this means is that the golden mountain, for instance, does not

actually exist, but

since the property of ‘being a golden mountain’ is existence-entailing, the entity does exist in those worlds which contain entities having that property. It is not entirely intuitive to determine which properties

22

23

24

are

For a survey of the current debate, see Maria Reicher, ‘Nonexistent Objects,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Summer 2014 edn); see http://plato. stanford.edu/archives/sum2014/entries/nonexistent-objects/. See for instance, Francesco Berto, ‘Modal Meinongianism and Fiction:The Best of Three Worlds,’ Philosophical Studies 152 (2011), pp. 313–334, and Terence Parsons, Nonexistent Objects (New Haven, CO: Yale University Press, 1980), as well as the already cited work by Routley and Priest. Priest, Towards Non-Being, p. 86.

2.3 Enter Meinong

existence-entailing in this sense. For instance, Priest himself holds that ‘being round’ is not existence-entailing, whereas ‘being round or square’ might be. 25 Accordingly, this solution may suffer from some of the same problems that the nuclear/extranuclear distinction does. Moreover, the solution obviously violates Meinong’s principle of independence, since it explicitly suggests that the properties that an object has may have a bearing on its existence. While Priest’s solution may be an improvement over Meinong’s and Routley’s solutions, it is still an open question which properties should be considered as existence-entailing. This issue is picked up by Crane, who is of the opinion that there may be significantly more such properties than Priest suggests. Crane also thinks that one can get rid of the possible world machinery, which admittedly complicates some of the other proposed solutions.

26

On Crane’s view, it is the natures of things that determine which properties are existence-entailing, as these natures determine which conditions must hold for predications to be true of those things. In contrast, non-existent objects cannot have certain properties, such as being golden, because it is in the nature of golden things that they exist. This entails, as suggested above, that the supposedly non-existent golden mountain does exist in some worlds, even though not in the actual world. However, non-existents, such as Sherlock Holmes, can have certain properties that are not existence-entailing, such as being fictional. This is a property that the non-existent Sherlock Holmes can have even in the actual world. Crane’s suggestion as well requires something further, namely an account of what it is to have a property and how exactly the natures of things determine which properties are existence-entailing. But we will leave this issue aside here, as it would require heavier metaontological machinery than we have at our disposal so far (some of that machinery will be provided in subsequent chapters). Whether or not Crane is able to overcome the srcinal Russelian challenge, it is clear that his view is another interesting example of an anti-Quinean position. Crane’s view does differ somewhat from both the Meinongian and the noneist lines, though. Crane’s starting point is that he wants claims about non-existents to come out as true, but this is

25 26

See Reicher, ‘Nonexistent Objects.’ Crane, The Objects of Thought, p. 60

25

26

Quine vs. Carnap: on what there is and what there isn’t

inconsistent with the standard approach to existence and quantification. Let’s briefly consider one of his examples: (S) Some characters in the Bible existed and some did not.27

(S) appears contradictory, because the domain of the quantifier ‘some’ includes both existents and non-existents. So it cannot come out as true on the standard understanding of quantifiers like ‘some’. But Crane wishes (S) to come out as true without interfering with the basic logic of quantification. Instead, Crane offers a reinterpretation of domains of quantification, suggesting that they may include non-existents as well. But why think that domains of quantification could not include non-existents in the first place? Crane has a theory, based on the popular conception of domains of quantification regarded as sets: I suspect that resistance to the idea of domains ‘containing’ non-existent objects comes from a certain metaphysical picture of what these domains involve. The picture is that sets are collections of things and these things must be real if we are going to understand how language hooks up with reality at all. The appeal of thinking in terms of sets is that sets are precisely defined entities, and have a clear criteria of existence and identity. On this picture, all members of sets must exist, and so if domains are understood in terms of sets, then domains containing non-existents do not make any sense.28

Instead of conceiving domains of quantification as sets, Crane proposes that quantifying over things is simply a way of talking about things, to have them in our discourse. Crane’s interpretation of (S) is that to say ‘there are biblical characters who did not exist’ should not be understood as asserting the existence of these non-existent biblical characters. Rather, it is to say that some biblical characters did not exist – the quantifier ‘there are’ should not mislead us. So existential sentences, on Crane’s view (reflecting that of Louise McNally), are not used to assert the existence of anything, but rather to introduce new referents into the discourse. The upshot is that the existential claim ‘there are Fs which are G’ is not obviously equivalent to the claim ‘there exist Fs which are G’, rather than simply ‘someFs are G’, 27

Ibid., p. 30. For a classic discussion relating to this example, see Saul Kripke, Reference and Existence (Oxford University Press, 2013).

28

Ibid., p. 37.

2.4 External and internal questions

where the natural language quantifier ‘some’ is not necessarily ontologically committing.29 To conclude this tightly-packed section, it seems clear that we would do well to take the Meinongian approach and its variations seriously. Meinongianism has been undergoing something of a renaissance in recent years and this trend can be expected to continue. We will now return to the heart of the Quine vs. Carnap debate.

2.4 External and internal questions Quine’s ‘On What There Is’ was partly aimed at Carnap, since Quine was suspicious of Carnap’s commitment to properties, propositions, and meanings. For Carnap, these are as unproblematic as numbers: there is nothing wrong in our commitment to such entities. Carnap’s position relies on his distinction between internal and external questions: [W]e must distinguish two kinds of questions of existence: first, questions of the existence of certain entities of the new kind within the framework, we call them internal questions; and second, questions concerning the existence or reality of the system of entities as a whole, called external questions. Internal questions and possible answers to them are formulated with the help of the new forms of expressions. The answers may be found either by purely logical methods or by empirical methods, depending upon whether the framework is a logical or a factual one.30

Internal questions have an absolute priority in Carnap’s system, because it turns out that only internal questions are to be understood as factual. These are questions such as ‘Are there carbon molecules larger than C60?’ to which we can, sometimes quite easily, provide a true or false answer based on the empirical methods available in our scientific framework. Questions like ‘What is the smallest prime number?’ would also qualify as internal, but in this case the answer is analytically true or false. For instance, Carnap would consider all mathematical truths to be analytically true, so true in virtue

29

30

Ibid., p. 46 ff. For an in-depth discussion of quantifiers such as ‘some,’ see also Hanoch

Ben-Yami, Logic and Natural Language (Surrey: Ashgate, 2004). Rudolf Carnap, ‘Empiricism, Semantics, and Ontology.’ Reveue Internationale de Philosophie 4 (1950); reprinted in his Meaning and Necessity: A Study in Semantics and Modal Logic (University of Chicago Press, 1956), p. 206.

27

28

Quine vs. Carnap: on what there is and what there isn’t

of meaning, similarly to the famous ‘All bachelors are unmarried’. In contrast, external questions concern the framework itself: they are philosophical or ontological questions about the nature and, perhaps, justification of the whole framework. We might think that external questions should be raised when introducing a new framework, such as a specific scientific theory. Hence, we could ask:‘Is quantum mechanics committed to the objective existence of the wavefunction?’ or something similar. Or, again, we could ask: ‘Do numbers exist?’ Philosophers certainly ask such questions in earnest, but Carnap considers this a mistake: ‘[T]he introduction of the new ways of speaking does not need any theoretical justification because it does not imply any assertion of reality’.31 The attitude that Carnap is advocating here is liberal, but also dismissive. It is liberal in the sense that we are free to adopt novel frameworks without a lengthy assessment of their theoretical justification (but note that they may still need a pragmatic justification). In this sense, external questions could be considered a matter of pragmatic preference, without true/false answers. But Carnap’s approach can also be seen as dismissive in the sense that the value of a philosophical analysis of different frameworks and indeed the possibility of objective comparison between their theoretical commitments seems to be undermined. To fully understand Carnap’s approach, we ought to say something about what is being meant by ‘framework’ here. This turns out be a tricky issue. Scientific theories were mentioned in this connection, but Carnap’s point seems more general: the system of numbers and the system of points in space-time can also be understood as frameworks.32 Since the notion of a framework is clearly supposed to apply to a very wide variety of things, including both abstract and concrete things, we could perhaps understand the notion of a framework as any linguistic system (or a language fragment) that is constructed in order to speak about new kinds of entities. This way a framework could apply just as well to the system of numbers and the system of concrete things. The language of a scientific theory dealing with novel entities is certainly one example of such a linguistic framework (think about the names that scientists give to these entities and the concepts that they introduce), but any novel ‘language’, regardless of its

31

Ibid., p. 214.

32

For further discussion on ‘framework,’ see Eklund, ‘Carnap’s Metaontology,’ pp. 231 ff.

2.4 External and internal questions

function, would presumably qualify. With the internal/external distinction and the notion of a framework in place, it is not difficult to see that, at least on one interpretation, Carnap’s view would seem to lead towards a certain type of relativism: if the factivity of any existence claim can only be assessed relative to a given linguistic framework, then the type of universal existence claim that philosophers are often interested in loses much of its value. Indeed, the most radical interpretation of this view suggests that asking factual questions in the external sense – like philosophers perhaps sometimes do – is completely misguided, devoid of cognitive content.33 But it should be immediately noted that this type of framework-relativity does not necessarily arise from the interpretation of frameworks as linguistic systems. The relativism that does emerge may not be any more serious than a type of linguistic relativism: there are different possible languages and the truth-values of sentences are relative to languages. So, whether this gives rise to any kind of scepticism about ontology, as Carnap’s work is often considered to do, is controversial. Indeed, if Matti Eklund is correct, the type of sceptical result regarding ontology that is sometimes associated with Carnap only goes through on a more radical interpretation of ‘framework’: On a second, relativist, understanding, Carnap’s ‘frameworks’ are not mere language-fragments; instead, frameworks are like perspectives or outlooks. On the relativist view, framework-relativity is not the trivial dependence of meaning upon language, but the sentences ontologists are concerned with are framework-dependent in some more radical sense: the propositions the sentences express are not true or false absolutely but only relative to frameworks. An internal question is a question of what is true relative to some framework in this demanding sense of ‘framework.’ An external question is a question of what framework is the correct one.34

Eklund’s view is that Carnap is not a relativist in this sense, but merely a language pluralist, which is the view that emerges from the interpretation of

frameworks as linguistic systems or language-fragments. We’ll return to a contemporary discussion of language pluralism in the fifth section. Leaving Carnap exegesis aside, the internal/external distinction is not just of historical relevance. The distinction has been applied and adapted 33 34

For discussion, see Eklund, ‘Carnap and Ontological Pluralism,’ p. 132. Eklund, ‘Carnap’s Metaontology,’ p. 233.

29

30

Quine vs. Carnap: on what there is and what there isn’t

in various contemporary accounts, especially by those who consider themselves (or whom others consider) broadly ‘neo-Carnapian’. Many contemporary views in metametaphysics have been motivated by the internal/external distinction; one such is Thomas Hofweber’s. 35 Although Hofweber’s reading of Carnap’s distinction may be controversial, the view that he develops has an interesting contrast with Carnap’s view. Hofweber distinguishes an internal and an external reading of quantification . The external reading is the standard domain-related understanding of quantification, so when quantifiers are read in this sense, our focus is on the domain of objects being quantified over, whatever it may be. The resulting truth-conditions on this reading follow the standard semantics for quantifiers – this is how quantifiers are normally used. But Hofweber thinks that quantifiers are polysemous, semantically underspecified. That’s because sometimes quantifiers have, in addition to the domain-related reading, an internal reading, which concerns the inferential role of the quantifier. The idea is that sometimes, in ordinary English, we mean to use quantifiers in a different sense than the external reading suggests. Hofweber doesn’t claim that this is a particularly common phenomenon, but there are, he thinks, clear special cases where quantifiers must be read in another sense. One such case is when we attempt to communicate under partial ignorance – when we have some idea about what we are talking about, but perhaps lack detail. In such a case, the domain-related understanding of quantifiers may be inappropriate because we may ignorant about the domain. To illustrate Hofweber’s external/internal distinction, consider how we might understand the statement ‘Everything exists’ – something that Quine might utter. On the face of it, this statement is trivially true, if ‘everything’ includes all the things that exist. But we need to keep in mind the problems concerning non-existents; there may be some things, such as Sherlock Holmes, that do not exist, and hence not everything exists. So if ‘everything’ includes Sherlock Holmes, then the statement is false. These two senses are supposed to correspond with the external and the internal reading of quantification, respectively. The internal reading is more appropriate when we consider the possibility of non-existents – it’s not clear 35

See for instance Thomas Hofweber, ‘Ambitious, Yet Modest, Metaphysics,’ in Chalmers, Manley, and Wasserman (eds.),Metametaphysics, pp. 260–289.

2.4 External and internal questions

whether they are in the domain of the quantifier or not. Hofweber illustrates the distinction further with statements such as ‘Someone kicked me’, where the reading of the quantifier is clearly external – there is a specific someone who kicked me and it doesn’t necessarily even matter if I don’t know who it is. But consider: ‘There is someone we both admire,’ but I have forgotten who it is that we both admire. In this case the idea is that the quantifier will not range over the whole world, even though it could be anyone that we both admire. Why is that not enough? Well, if it happens to be, say, Sherlock Holmes that we both admire, then it appears that the quantifier would range over something that is not included in the world at all. So while communicating under partial ignorance about who it is that we both admire, the quantifier cannot be understood in the external sense. Hofweber goes on to apply the internal/external distinction to the familiar question: ‘Are there numbers?’ On the internal reading, the answer turns out to be a trivial ‘yes’. But Hofweber is not an advocate of the trivialization of metaphysics. Indeed, he suggests that on the external reading the question is not trivial, and furthermore, he thinks that mathematics does not provide an answer to the external reading. Accordingly, perhaps it is the external reading of existence questions that metaphysics is concerned with: they are not (entirely) trivial and the special sciences do not answer them. On the face of it, this of course contrasts with Carnap’s view – Hofweber, after all, considers himself to be defending the possibility of metaphysics, albeit in a ‘modest’ sense. But since we have seen that it is rather difficult to pin down Carnap’s view, it is an open question how much of this is truly in contrast with Carnap. For Hofweber, the external and internal readings have the same status, but his account does agree with Carnap’s to the extent that internal readings of quantified statements are very often trivial. Indeed, when it comes to the question ‘Are there numbers?’ Hofweber considers the question to be trivial on the internal reading, but that’s not what metaphysicians are usually thought to be after when they ask this question. Carnap’ s judgement about this is that the metaphysician asks a meaningless question, but Hofweber has a little more to say. In particular, he considers number words to be non-referential. Hofweber posits that number-talk does not even aim to refer. If this is correct, then there is nothing in the world that numerals pick out and hence no such things as ‘numbers’:

31

32

Quine vs. Carnap: on what there is and what there isn’t

Number words, just like any other words, can be used by particular speakers with the intention to refer, and these speakers can succeed in referring to something. I can use ‘two’ to refer to my biggest tomato plant, and succeed. But I can’t use it or any other word to refer to the number two (as this phrase is commonly used).36

More generally, Hofweber concludes that all that is left for metaphysics to do is to determine whether, in regard to a given existence question, internalism or externalism is true. If internalism is true, then there are no entities of the type in question. If externalism is true, then metaphysics has nothing to do with the answer. According to this view, the role of metaphysics is modest indeed, even if it is supposed to be more ambitious than according to Carnap.37 Admittedly, Hofweber’s use of the internal/external distinction has only a remote connection to Carnap’s, but the distinction has been taken up by many others as well. For instance, David Chalmers adapts the distinction and develops a new, related distinction: ordinary existence assertions vs. ontological existence assertions.38 The former, as the name suggests, are ordinary assertions such as ‘There are three chairs in this room’. The latter are typical philosophical assertions, such as ‘There is a universal “redness” shared by this red chair and that red table’. Chalmers illustrates the distinction with the idea of an ‘ontology room’ – a special place where it is acceptable to make ontological existence assertions and possible to have debates about matters ontological. Outside the ontology room, only ordinary existence assertions can be properly evaluated. The ontology room has become a standard tool in metametaphysics, although it may have only a loose connection to Carnap’s srcinal view. In any case, we ought to look into this tool in some more detail, since we will encounter it later on as well. Before Chalmers, Cian Dorr proposed that the questions of ontology could be conceived as being discussed in the special language of ontology, call it Ontologese, which is more suitable for ontological debates than 36 37

38

Hofweber, ‘Ambitious, Yet Modest, Metaphysics,’ p.268. For further discussion, see T. E. Tahko, ‘In Defence of Aristotelian Metaphysics,’ in T. E. Tahko (ed.), Contemporary Aristotelian metaphysics (Cambridge University Press, 2012), pp. 26–43. See David Chalmers, ‘Ontological Anti-Realism’ (p. 81), in Chalmers, Manley, and Wasserman (Eds.), Metametaphysics, pp. 77–129.

2.4 External and internal questions

ordinary English.39 The scenario that Dorr entertains is that of several (isolated) language communities, each one of them speaking English, but differing in one crucial respect. Let us focus on just two of these communities, the universalists and the nihilists. These labels reflect the attitude of each community regarding the ‘Special Composition Question’ – something of a default test case in the metametaphysics literature.40 Special Composition Question (SCQ) Under what conditions do a number of objects compose a further object?

The universalists answer this question by saying that a number of objects always compose a further object, no matter how far apart or how different

they are; your nose and the Eiffel Tower compose a further object, so do the Milky Way and Andromeda galaxies, and indeed all those four things. The nihilists are at the other extreme, as they would say that a group of objects never composes a further object: all there is are simple objects, so-called

‘mereological atoms’. SCQ is a question concerning mereology, the study of parts and wholes. Between the mentioned extremes lies the view that we can give a set of conditions under which a number of objects compose a further object – one famous answer on these lines is van Inwagen’s, according to whom only living things – lives – constitute composite objects in the required sense.41 Our current interest, however, is not so much the SCQ itself, but rather our understanding of the debate in general. Since these different language communities differ in their attitude regarding the question, we must ask: is their disagreement merely linguistic or is there a substantial difference in their views about the structure of reality? In other words, is there any genuine metaphysics involved? We can already see how this discussion is connected to Carnapian themes: the different language communities in Dorr’s example could be interpreted in terms of Carnap’s ‘frameworks’. So what would Carnap say about the example? He would presumably insist that each of the language communities truly has its own language and the claims that the 39

40

41

See Cian Dorr, ‘What we Disagree about when we Disagree about Ontology,’ in M. E. Kalderon (ed.), Fictionalism in Metaphysics (Oxford University Press, 2005), pp. 234–286. The label ‘Ontologese’ is from Ted Sider, see for instance his ‘Ontological Realism,’ in Chalmers, Manley, and Wasserman (eds.), Metametaphysics, pp. 384–423. The modern formulation of the Special Composition Question originates from Peter van Inwagen’s Material Beings (Ithaca, NY: Cornell University Press, 1990). Ibid.

33

34

Quine vs. Carnap: on what there is and what there isn’t

universalists and the nihilists make in their respective languages hold true in that language. Dorr calls this type of view ‘conciliatory’, because we can

conciliate the debate between the communities simply by translating their respective claims about composition. How exactly this could be done is of course another question, and one which Dorr has attempted to answer. Others, such as Ted Sider, would disagree with the Carnapians about the possibility of such a translation.42 We need not go into all the details of the possible translation scheme here, but there is one aspect that we will have to consider in some detail later, namely the possible variation of the meaning of quantifiers between languages.43 Before we move on, let us use this debate to further illustrate the idea of the ontology room. To understand this tool in the present context, we might think that it is us, as metaphysicians, who constitute one language community, and ‘folk’ – the woman and man of the street – who constitute another language community. In fact, the previous example, SCQ, is something that ‘folk’ may not have a very strong opinion about. In the ‘folk language’, both universalism and nihilism seem to make rather strange claims. Folk will likely think that a number of objects compose a further object whenever they seem to do so. There are tables and chairs, cars and houses. It would seem utterly silly to deny that these composite objects exist. Similarly, ‘objects’ like the combination of your nose and the Eiffel Tower seem blatantly not to exist. But many metaphysicians take these types of debates quite seriously. So we may ask: is the disagreement between ordinary folk and metaphysicians a substantial one? Here is the charitable reading that Dorr proposes: What we debate in the ontology room is the question what there is strictly speaking – what there really, ultimately is – what there is in the most fundamental sense. Of all the many meanings a quantifier like ‘something’ might

have, one is special. This is the one in terms of which all the rest are to be this sense would analysed; it is the one such that to find out what there is in

be to fulfil the traditional metaphysical goal of comprehending reality as it is 44 in itself. When we do ontology, our quantifiers bear these special meanings.

42 43

44

See Ted Sider, ‘Hirsch’s Attack on Ontologese,’Noûs 48.3 (2014), pp. 565–572. For a classic account on this issue, see Eli Hirsch, ‘Quantifier Variance and Realism,’ Philosophical Issues 12 (2002), pp. 51–73. Dorr, ‘What we Disagree about when we Disagree about Ontology,’ p. 250.

2.5 Language pluralism

The italicized words in this

passage – really, ultimately, fundamental

–

may not seem very helpful. Indeed, unless they are combined with an account about the hierarchical structure of reality and the import of the fundamental – the most ‘real’ – things, then they do not help us make sense of the metaphysician’s task. The idea behind such talk wil l be revisi ted in due course, primarily in Chapters 5 and 6, but we wil l kee p coming across it repeatedly in thi s book. We can already note one easy reply to this type of suggestion, which Dorr mentions as well: one can question the idea that there is any ‘special’ sense of quantification and insist that reality can be described equally well in a number of languages. Such a resp onse can be motivated by Carnap’s approach, since we’ve seen that Carnapians would consider the idea of the ontology room to be a doomed attempt at aski

ng external ques-

tions that are factive. Instead, the Carnapian would claim that the meaning we ascribe to quantifiers is a pragmatic choice. Hence, if we are debating the SCQ, we should choose the answer most appropriate to our pragmatic goals. It seems likely that the upshot would then be some kind of modest position, between the extremes of

universalism

and nihilism . For an ontological realist, whose approach the passage quoted from Dorr reflects, this result is unsatisfactory. We will discuss this position in more detail later, but in the remainder of this chapter we will focus on the Carnapian approach. Proponents of the Carnapian approach would embrace the pragmatic result rather than be disheartened by it. This has resulted in metaontological theories that purport to be able to deal with the type of language pluralism that the Carnapian approach would seem to entail.

2.5 Language pluralism In contemporary terms, Carnap could be described as a language pluralist . This is, perhaps, the most charitable interpretation of his view: there

are a variety of different frameworks – interpreted as languages or language fragments – that we can adopt, and different existential assertions come out as true in these frameworks because each language has its own sense of ‘existence’. This claim is blatantly in conflict with the Quinean

35

36

Quine vs. Carnap: on what there is and what there isn’t

approach according to which existence is univocal. Indeed, one way to understand the debate between Carnap and Quine – or, at any rate, between neo-Carnapians and neo-Quineans – is that they disagree about the interpretation of existence claims. More accurately, the disagreement is sometimes taken to concern the interpretation of the existential quantifier: whether it is univocal in the Quinean sense or whether it varies from context to context. This debate falls under the rubric of quantifier variance , which is a topic that we will discuss further in the

ter.

45

next chap-

For the time being, let us focus on the historical background and

the basic idea of language pluralism , without all the complications of the contemporary debate. We can illustrate the idea of language pluralism with the debate about the ontological status of numbers, which has already made its appearance in earlier examples. Platonists would answer affirmatively when asked ‘Do numbers exist?’ They would consider numbers to be abstract objects. Many mathematicians seem at least implicitly to favour this type of view. Nominalists would disagree, arguing that numbers do not exist (they might

think there are no abstract objects at all). Now, enter the language pluralist; she thinks that the Platonists and nominalists are both correct, relative to their respective frameworks. In the Platonist language the assertion of existence regarding numbers comes out as true, but in the nominalist language it comes out as false. So from the point of view of the language pluralist, there is no real disagreement here. We saw a similar attitude emerging in the last section with regard to the Special Composition Question. It seems that the position holds quite generally. The view that disagreements about ontology tend to be merely linguistic disagreements rather than substantial ontological disagreements naturally follows – this could be labelled as a type of deflationism about ontology. But it is not really clear how we should understand the view, for there is a sense in which it is perfectly trivial: the utterance ‘Numbers exist’ seems to come out as true in some languages (the Platonist’s language) and false in some other languages (the nominalist’s language). This is something we already saw in Hofweber’s analysis. Yet, unless we have an understanding about whether

45

Eli Hirsch, a kind of neo-Carnapian, is perhaps the best known proponent of quantifier variance, e.g., his ‘Quantifier Variance and Realism’.

2.5 Language pluralism

‘exist’ (or ‘numbers’, for that matter) means the same thing in these different languages, we are not saying anything very worthwhile. Surely, the language pluralist wants to say something more than this. But what does she want to say? First, regarding the historical case, we should note that language pluralism does not fall straight out of Carnap’s distinction between internal and external questions. We can accept this distinction without committing ourselves to language pluralism. Still, the tool provided by the internal/ external distinction is certainly attractive for anyone who is inclined to accept language pluralism. If we agree that all the external questions – which ontology mostly deals with – are pragmatic, then it might seem that our ontological disagreements quickly begin to dissipate. However, this inference does not go through as it is. As Matti Eklund puts it, the internal/external distinction needs to be complemented by the assumption that there are ‘equally good’ languages to choose from. 46 In other words, language pluralism follows only if one thinks that it genuinely doesn’t matter – other than for pragmatic purposes – which language we choose. But this makes language pluralism a much less attractive view, for there are surely many questions that we would regard less sensitive to the Carnapian analysis. For instance, scientific questions such as ‘Are there tigers in Australia?’ are evidently not subject to the same type of variation between languages as, perhaps, the question about the existence of numbers. It would be rather strange indeed to insist that it is a question of which language we adopt whether there turn out to be tigers in Australia. So it is difficult to see how the language where tigers exist in Australia and the language where they do not could be ‘equally good’ in the required sense. The upshot is that even if in ontology we ask questions that often seem to turn on a choice of language, this does not appear to be true of sci-

ence or ordinary discourse in general. Nevertheless, language pluralism, at least in some of its forms, requires the additional assumption regarding ‘equally good’ languages. It turns out that getting to the bottom of the language pluralist’s position is more difficult than it may first have seemed. On the one hand, the position is threatened by trivialization. On the other, the initial force and

46

Matti Eklund, ‘Carnap and Ontological Pluralism,’ p. 140.

37

38

Quine vs. Carnap: on what there is and what there isn’t

appeal of the position may be lost unless it generalizes. There is more to be said about language pluralism, but we will return to some of these issues in later chapters, especially in Chapter 4, where different metametaphysical positions are discussed in more detail. In the next chapter we will go deeper into the metaontological issues in the background of the Quine vs. Carnap debate, quantification and ontological commitment in particular.

3

Quantification and ontological commitment

In the previous chapter, we saw that quantification and ontological commitment are central themes in the history of metaontology. The central question is: Does quantifying over some entity entail ontological commitment to the existence of that entity? The Quinean story is now familiar to us: to be (an entity) is to be the value of a bound variable, that is, a variable in the scope of the existential quantifier. In this chapter we will, among other things, clarify this Quinean idea and examine some of the implications of the different views we have already outlined. We begin with an analysis of the meaning – or interpretation – of the existential quantifier. We’ve seen that W.V. Quine and Alexius Meinong would disagree about whether there is a difference between ‘there is’ and ‘exists’. For Quine they mean the same thing:existential quantification is univocal, and does not allow for variation in meaning. But could the disa-

greement between Quine and Meinong be explained simply by a difference in their chosen interpretation of the existential quantifier? This might seem to be a tempting explanation and it is not untypical to hear mention of ‘Meinongian quantification’ as opposed to the ‘Quinean’ sort. To be sure, there are different views about quantification and existential quantification in particular. According to one such view, even the name of the quantifier is misleading: it should be called the ‘particular quantifier’ instead. On this view, the particular quantifier is not existentially loaded (and hence should not be called ‘existential’). This type of view is defended by some noneists (about whom we learned in Chapter 2). Recall, however, that for

contemporary Quineans and Meinongians the important difference is whether existence may be regarded as a property – not all Meinongians would resist the Quinean univocality thesis. Furthermore, some would consider it a mistake to regard ‘exist’ as any kind of a quantifier at all (recall the example: ‘Some characters in the Bible existed and some did not’). On 39

40

Quantification and ontological commitment

this sort of view, we should perhaps be focusing on ‘there is’ and its ilk instead. We might then suggest that the debate about univocalism concerns ‘there is’, which means something different from ‘exists’. 1 Some of these issues are closely related to the problem of whether ‘exists’ should be considered a predicate as opposed to a quantifier. We have discussed related matters in some detail in Chapter 2 and seen that there are arguments on both sides. Below we will get a little deeper into this issue, but despite a recently active debate, it should be noted that the received view still takes existential quantification quite seriously, leaving the view taking existence as a predicate to the minority.2 We will also consider a more neutral way to understand ontological commitment, making use of a distinction between the ontological commitments of a sentence and the semantic machinery assigning truth conditions to that sentence. Later on, in the third section, we will return to the topic of quantifier variance in its contemporary form. A brief overview of the ongoing debate between Eli Hirsch and Ted Sider on this topic will be given, although these two are by no means the only philosophers who have contributed to the discussion. Finally, after we have considered quantification and ontological commitment from several angles, the reader may be pleased – or frustrated! – to discover that the emphasis on quantification in metaontology may in fact be misguided, as argued by Kit Fine. Fine’s view will be considered in the fourth section. The discussion so far has assumed that the manner in which Quine understands the task of ontology is at least broadly correct, even if we have heard from many dissenting voices already. This understanding is that ontological questions are (primarily) quantificational questions. But this assumption has been called into question; we will consider some reasons to think that the most interesting ontological questions are in fact not quantificational. So what else could these interesting ontological questions be? That will be a recurring topic in later chapters, but one popular suggestion is 1

2

For a related discussion, see Hanoch Ben-Yami,Logic and Natural Language (Surrey: Ashgate, 2004); Hanoch Ben-Yami, ‘Plural Quantification Logic: A Critical Appraisal,’ Review of Symbolic Logic 2.1 (2009), pp. 208–232. For further discussion on ‘existence’ and especially on the issue of whether it should be considered a predicate, see Michael Nelson, ‘Existence,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2012 edn); see http://plato.stanford.edu/archives/ win2012/entries/existence/ and the references therein.

3.1 The meaning of the existential quantifier

that the interesting questions concern what is

fundamental – what are

the most basic building blocks of reality. The Quinean account does not address this type of question, but others claim to be able to do so. Of course, the debate goes on, since there are those who would take a (neo-) Carnapian attitude towards these supposedly interesting ontological questions. After all, the question about what is fundamental looks very much like an external question in Carnap’s sense, as discussed in the previous chapter : if the fundamental is assumed to concern the objective, framework-independent structure of reality, then it would seem to violate Carnap’s framework-dependence. Perhaps this isn’t necessarily the case, though, since a Carnapian could understand questions regarding fundamentality in a different manner, internal to a particular (philosophical) framework. Be that as it may, one goal of the present chapter – and this book in general – is to provide the reader with some further tools to assess this convoluted debate.

3.1 The meaning of the existential quantifier On the basis of the discussion in the previous chapter, we can summarize two competing views about the meaning of the existential quantifier – or simply the meaning of ‘exists’. On the Quinean view, ‘exists’ has only one meaning, it is univocal. When we say that numbers exist or that cabbages exist, we are using the very same existential quantifier, and its meaning is the same in both cases. On one version of the Meinongian view, statements like ‘There are things that do not exist’ express something different: we might call it ‘Meinongian quantification’. So we can distinguish between normal cases of ‘exists’ and cases where the entity under investigation lacks the property of existence, but nevertheless has some sort of being. We have examined the problem of non-existence in the previous chapter, so it is not our primary concern in this chapter. But are there any other reasons to doubt the Quinean doctrine regarding univocalism about ‘exists’? Note that the debate could be understood on two different levels. On the one hand, ‘exists’ may refer to the use of the term in our current, actual language. On the other, we may consider which language is the ‘best’ for our theoretical or pragmatic purposes. So, whether ‘exists’ is univocal in the first sense and whether it is univocal in the second sense could be regarded as separate questions. We will primarily focus on the first sense of

41

42

Quantification and ontological commitment

the question, but when we get to discussing quantifier variance, the second issue becomes more pressing. On the face of it, it may seem preferable to ascribe different meanings to ‘exist’ in different contexts. For instance, when we say that numbers exist, we do not seem to mean it in quite the same sense as when we say that some sort of material object exists. Accordingly, perhaps the meaning of the existential quantifier could be thought to depend on the type of entity that we are quantifying over? After all, numbers are abstract objects, so their ‘existence’, whatever it amounts to, would not seem to be quite like that of material objects. The natural conclusion to draw from this is that ‘existence’ simply means something else in these cases; ‘exists’ is equivocal rather than univocal. Attractive as this line of thought may be, Quineans would not find it persuasive. The mistake, they would insist, is thinking that existence is some kind of activity – something that a thing does. Peter van Inwagen, commenting on Gilbert Ryle’s argument to the effect that existence is equivocal, takes just this line and specifies further what the univocacy of existence amounts to: The thesis of the univocacy of existence […] does not imply that existence comes in species or that ‘existent’ is a ‘generic’ word like ‘colored’ or ‘sexed.’ This thesis does not imply that there are or could be ‘species’ words that stand to the generic ‘existent’ as ‘red and ‘green’ stand to ‘color’ and as ‘male’ and ‘female’ stand to ‘sexed.’3

It turns out that the issue is rather more complicated than it might have seemed at first. Indeed, to clarify what the Quinean position amounts to, we should take aboard the thesis that quantification commits us ontologically in addition to the thesis that it is univocal. One way to understand the idea of ontological commitment is to think about the truth-conditions of sentences.4 To cash out the ontological commitments of a sentence, we need to determine what kinds of constraints the 3

4

Peter van Inwagen, ‘Being, Existence, and Ontological Commitment,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.),Metametaphysics (Oxford University Press, 2009), p. 486. For further discussion on this and other ways to understand ontological commitment, see Augustín Rayo, ‘Ontological Commitment,’ Philosophy Compass 2/3 (2007), pp. 428–444; see also Philip Bricker, ‘Ontological Commitment,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2014 edn); see http://plato.stanford.edu/ archives/win2014/entries/ontological-commitment/.

3.1 The meaning of the existential quantifier

sentence’s truth imposes on the world. However, an important caveat that should be noted here is that the ontological commitments of a sentence can be distinguished from the assignment of truth-conditions to that sentence. This is an idea taken up by Augustín Rayo, who argues that the machinery a semantic theory uses in specifying truth-conditions (e.g., sets) is not itself a constraint on the world. In other words, even if sets are used in specifying truth-conditions, this does not itself constitute a demand on the world. Moreover, on Rayo’s view, when we take a familiar quantificational claim such as ‘Numbers exist’, that is, ‘ ∃x(NUMBER(x))’, it will depend on our semantic theory as to exactly what the sentence is committed to. Rayo himself has recently developed an interesting metaontological position, but here we will focus on his critical remarks.5 The target of Rayo’s critique is the metaphysicalist, who may be making the type of error we have just discussed: the metaphysicalist holds that a type of correspondence must exist between the logical form of a true sentence and the metaphysical structure of reality. This can be illustrated with an example. In order for the sentence ‘Elizabeth reads’ to be true, there must be a certain ‘metaphysically privileged’ structuring of reality into the constituents of the sentence. First, let us suppose that what we are referring to with the sentence is the fact that Elizabeth reads. In this case, the constituents of the fact could be Elizabeth and the property of reading (although this may not be the only way to do it!). The important thing is that there is a correct way of structuring reality, which will take into account the metaphysically privileged deconstruction of the fact into its constituents. If there are several ways of deconstructing the fact, the metaphysicalist typically adds that only one of them is fundamental.6 These are all issues to which we shall return, but note that

the metaphysicalist seems to be committed to something that we saw Rayo resist earlier, namely that the truth-conditions of the sentence ‘Elizabeth reads’ would now appear to impose direct constraints on the world, via the metaphysically privileged structuring. So Rayo’s point is that the metaphysicalist may be employing a problematic philosophy of language, since the truth of a sentence does not in fact need to impose the restriction on the world that the metaphysicalist assumes. 5

6

See Augustín Rayo, Construction of Logical Space (Oxford University Press, 2013). For discussion, see Matti Eklund, ‘Rayo’s Metametaphysics,’ Inquiry: An Interdisciplinary Journal of Philosophy 57.4 (2014), pp. 483–497. This simplifies matters a bit; compare with Rayo, Construction of Logical Space, pp. 6–7.

43

44

Quantification and ontological commitment

Let us return to the quantificational claim, ‘Numbers exist’. In contrast with Rayo, Quine’s view is that such first-order quantificational sentences always carry a commitment to the objects being quantified over. But it is

important to see that this does not necessarily follow automatically. For further illustration, consider the following two sentences:7 (1) There is a pig on the sofa. (2) A pig is on the sofa. Sentences (1) and (2) seem to express the same fact. But (1) seems to carry with it more existential ‘weight’ than (2), since it is natural to interpret it as expressing the existential quantifier, whereas (2) does not immediately strike us as containing an existence claim. As Tim Crane notes, (2) does of course entail the existence of a pig and would normally only be uttered by someone who believes that a particular pig – the one on the sofa – exists.8 But reflecting Rayo’s line of thought, Crane points out that we ought to distinguish between the content of a sentence and what the content of the sentence entails. We can interpret both (1)and (2) as involving quantification, since ‘a pig’ may perhaps be considered to function as a quantifier. But this does not mean that (1) makes an existential claim unless a semantic link exists between existence and quantification. (2) of course entails that a pig exists and normally a person uttering (2) would surely not deny a commitment to the existence of pigs. However, the point of saying that (1) makes an existential claim relies on the thought that (1) and (2) express the same proposition. Crane accepts that (1) and (2) say the same thing and even that

they both involve quantification, but the point he is pressing is that just the fact that a claim is quantificational does not make it an existence claim. Not unlike Rayo’s suggestion, semantic theory would be needed to establish such a link and both Crane and Rayo seem to be sceptical of the existence of such a link. Furthermore, even if Quine is correct about the existential commitment of first-order quantification, we must keep in mind that his claim does not extend to all quantificational sentences, such as those involving modality. We might say that ‘It is possible that there are such things as dragons’, and

7

8

These examples are discussed in Tim Crane, The Objects of Thought (Oxford University Press, 2013), p. 43 ff. Ibid., p. 44.

3.2 The existential quantifier and ontological commitment

a standard way of specifying the truth-conditions of such sentences is by including those possible dragons as values of variables. But it does not yet follow from this that the world must contain possible dragons. Rather, as Rayo insists, we can only conclude that the semantic theory used is a commitment to such possibilia. So Rayo’s lesson is that we can talk about the ontological commitments of a sentence and the ontological commitments of semantic theory, but these should be distinguished from one another. We do not need to work out all the implications of these remarks for the Quinean criterion of ontological commitment, but since there clearly are alternatives to the Quinean criterion, then how did it come to be the default view in the first place?

3.2 The existential quantifier and ontological commitment It is not difficult to see why something called the existential quantifier would be considered to commit us ontologically to the existence of whatever it is that we quantify over. But it would be a mistake to think that this is just due to the name of the quantifier. To be sure, the Quinean analysis does suggest that the first-order existential quantifier is ontologically committing, but the history of this existentially loaded reading of the quantifier reveals that before Quine and Russell, such a reading was by no means the default.9 To see that existential quantification is not obviously existentially loaded, consider the following example, due to Graham Priest: ‘I thought of something that I would like to give you as a Christmas present, but I couldn’t get it for you because it doesn’t exist’. 10 This seems to be a perfectly legitimate instance of quantification, but it does not appear to be existentially loaded. On the contrary, it is explicitly said that what is being quantified over does not exist. Moreover, this is not a particularly odd statement by any means. I might be thinking of Escher’s impossible staircases, for instance, and be playing with the idea of getting a miniature version of one as a Christmas present for you. In this case, not 9

10

For a historical overview, see Graham Priest, ‘The Closing of the Mind: How the Particular Quantifier Became Existentially Loaded Behind Our Backs,’ The Review of Symbolic Logic 1.1 (2008), pp. 42–55. Graham Priest, Towards Non-Being: The Logic and Metaphysics of Intentionality (Oxford University Press, 2005), p. 148.

45

46

Quantification and ontological commitment

only does the entity in question not exist, but it could not exist. So it seems that there is a natural use for existentially non-loaded quantification. But perhaps there are some theoretical reasons to prefer existentially loaded quantification? Before modern logic emerged, there is little reason to think that quantification was used in an existentially loaded sense. Of course, Aristotle and the mediaeval logicians did not talk about ‘quantification’, but the principle was already in use. The idea of quantifying over something featured in the practice of determining the domain of discourse. According to this approach, ‘existential’ quantification does not involve any special connection with existence. Rather, ‘existence’ becomes a technical term which should not be considered as ontologically loaded. Indeed, some nineteenth-century logicians, such as Keynes and Venn, were very careful to point out that in the context of logic ‘existence’ should be understood simply as a sort of placeholder for whatever it is that we are currently investigating.11 It was only at the beginning of the twentieth century that Bertrand Russell knowingly started using existentially loaded quantification. In the previous chapter it was mentioned that Russell was interested in Meinong’s work, but ended up reacting to it quite critically. In fact, if Priest’s interpretation is to be believed, Russell seems to have decided to use existentially loaded quantification and only started developing arguments to that effect as an afterthought.12 The result was that Quine had a natural ally in Russell against Meinong. In any event, Quine’s rather rhetorical case against Meinong in ‘On What There Is’ has been hugely influential and, perhaps partly because of Russell’s earlier influence, it was accepted without much quarrel. But as we have seen, there are those who remain sceptical about the ontological weight of the existential quantifier. What is the alternative? If existential quantification were not considered existentially loaded, then perhaps we would do better if we called it the ‘particular quantifier’ instead, as is sometimes suggested. In that case, the quantifier should be read as ‘for some’, where it is left open whether the entity being quantified over exists or not. This is the type of view that Crane ends up considering as well:

11 12

For details, see Priest, ‘The Closing of the Mind,’ p. 46. Ibid., p. 51.

3.2 The existential quantifier and ontological commitment

[H]ow should we represent ‘some’ and ‘exists’ in a formal or logical representation of natural language? If we want to account for our initial data (e.g. sentences like [Some characters in the Bible existed and some did not]), we have a choice. We could translate ‘some’ as ‘∃’ in the usual way. But in this case, we should not understand ‘∃’ as ‘there exists’; we should express existence in another way. Or we could translate ‘there exists’ using ‘∃’, in accordance with Quine’s claim that ‘existence is what the existential quantifier expresses.’13 But in this case, we should not understand ‘∃’ as ‘some’; we need another quantifier symbol for ‘some.’14

It is up to us to decide which way to go. Some, like Priest, have in fact introduced a novel symbol for ‘some’ to make it clear that it is not being used in the Quinean existentially committing sense. But we face this decision only if we reject Quine’s criterion of ontological commitment and its close association with quantification. Note also that if the Quinean criterion, which associates ontological commitment with (scientific) theories, is rejected, then it remains unclear what ontological commitment amounts to. Rayo’s analysis gives us some clues as to what the options are, but there are plenty of other ways to go. Crane himself is partly sympathetic to a proposal due to Zoltán Gendler Szabó, according to which ‘ontological commitment’ is to be explained with reference to the attitudes of thinkers.15 More precisely, Szabó suggests that for a thinker to be ontologically committed is for her to ‘believe in’. This corresponds partly with the ordinary use of ‘believing in’ – not in the sense that one may, say, not believe in violence, but in the sense that the object of belief is part of one’s ontology. Someone might say, ‘I don’t believe in violence’, but that doesn’t mean that this person believes there to be no violence! That is, violence can be a part of one’s ontology regardless of a more abstract non-belief in violence. For believing-in to correctly represent the world, one must have a true conception of whatever it is that one does believe in. If one believes in Martians in Szabó’s sense, then one stands in a certain relation to the term [Martians], where the square brackets

13

14 15

W. V. Quine, ‘Existence and Quantification,’ in hisOntological Relativity and Other Essays (New York, Columbia University Press, 1969). Crane, The Objects of Thought, p. 47. Zoltán Gendler Szabó, ‘Believing in Things,’ Philosophy and Phenom enological Research 66 (2003), pp. 584–611.

47

48

Quantification and ontological commitment

indicate what the relevant plural term refers to. To put this a bit more clearly: to think that Martians are green is to stand in a certain relation to the proposition . Now, for a thinker’s representation of [Martians] to be correct, Martians must exist and at least part of one’s conception of them should be true (e.g., Martians being green). Moreover, we can think of Martians even if they don’t exist, because thinking of them is just to stand in a certain relation to a proposition about them. But it is a further question whether the proposition, say, , is representationally correct, or in other words, true. Szabó’s summary of the view is as follows: [T]he content of thinking of Fs is [Fs], and that [Fs] is representationally correct just in case Fs exist and the conception of Fs is true. Putting these things together yields the following result: if Fs exist but the conception of [Fs] is false, then someone who believes that Fs exist would be right while someone who believes in Fs would be wrong.16

In this way, one can say that there are true propositions about Fs (where Fs are, e.g., Martians), while denying belief in Fs. This is possible if the term is representationally incorrect: if there’s nothing that fits the ordinary conception of Fs. Yet, once it’s agreed that there aretrue propositions about Fs, that would suggest that one should believe that there are Fs. This seemingly contradictory (and admittedly a little confusing!) situation is resolved in Szabó’s theory because one can assume that the ordinary conception of Fs does not apply to Fs. For instance, one could believe that there are Martians, but deny that they are anything like what the stereotypical story about little green men suggests. The upshot that Szabó draws is that ontological commitment cannot be as simple as the Quinean idea of theory acceptance, since one may think that even a true theory misrepresents some of its crucial terms. On Szabó’s terminology this would mean that one should not believe in some of the entities that the theory is – on Quine’s criteria – existentially committed to. Of course, Quine does not talk about ‘believing in’, at least not in this fashion, so strictly speaking Szabó’s theory is just an alternative way of understanding ontological commitment, one that a Quinean could readily dismiss. Alternatively, if a Quinean states: ‘I believe that Pegasus flies’, this would not entail that 16

Ibid., pp. 606–607.

3.3 Quantifier variance and verbal debates

she ‘believes in’ Pegasus; we can instead give a paraphrase of the sentence following the Russelian account of definite descriptions. But at any rate, it appears that there are a variety of alternative views regarding quantification and its links or lack thereof to ontological commitment, as well as some reasons to think that even the term ‘existential quantifier’ is itself misleading. Regardless, we will continue to talk about the ‘existential quantifier’ in this book, for problematic as the term and its connotations may be, this is certainly the standard usage. However, the reader is advised to remain wary about these matters in what follows.

3.3 Quantifier variance and verbal debates There is a popular strand of the quantification debate that we have not yet properly discussed, even though it focuses exactly on the issue of whether (existential) quantification is univocal or not: quantifier variance. Since quantifier variance is one of the central topics of metaontology, this is also an important reason for considering the debate between Carnap and Quine to be central in metaontology. But we do not need to dwell on Carnap and Quine any longer, especially because we saw that the contemporary reconstruction of their debate is problematic at best. Nevertheless, the issue concerning the univocalism vs. equivocalism of quantification is, by many, considered to be at the heart of metaontology. For instance, here is Sider’s view on the matter: ‘the central question of metaontology is that of whether there are many equally good quantifier meanings, or whether there is a single best quantifier meaning’.17 Sider himself is strongly of the opinion that there is indeed a single best quantifier meaning. Note that, to be precise, we must keep in mind that the thesis that there is a single best quantifier meaning is in fact different from the thesis that quantification is univocal: all that Sider’s position requires is that in the context of ontology – when we speak Ontologese (see previous chapter) – quantification is univocal. So there could be cases where the meaning of quantification is context-dependent and hence equivocal, but those cases do not undermine the fact that quantification is univocal with regard to ontological questions.

17

Ted Sider, ‘Ontological Realism,’ in Chalmers, Manley, and Wasserman (eds.), Metametaphysics, p. 397.

49

50

Quantification and ontological commitment

In Sider’s terminology, quantification ‘carves at the joints’, which is to say that it latches on to the structure of reality in a reliable manner – quantification tells us something about how things really are in reality. 18 For Sider, this is what ontological realism amounts to, so he effectively identifies univocalism about quantification with ontological realism. But this usage of ‘ontological realism’ is somewhat problematic, for one could deny that quantification is univocal or that it occupies such a central position in metaontology, but still hold that reality has an objective, mind-independent structure. The latter is a more traditional, broader sense of realism, sometimes called metaphysical realism. Moreover, ontological realism is sometimes taken to be the view that ontological debates are, by and large, substantial rather than merely verbal or linguistic, but again, this view is independent of the question regarding quantification.19 We will return to these options in much detail in Chapter 4, where some terminological clarification regarding different metametaphysical positions is provided. Accordingly, for now it will be better to abstain from using the label ‘realism’ in any but the broadest possible sense, namely in the sense that ontological facts are, in some sense, objective. If Sider and van Inwagen are some of the loudest contemporary proponents of univocalism about quantification (although Sider prefers to talk about ‘single best quantifier meaning’ and hence there is an important difference between him and van Inwagen), then Eli Hirsch is no doubt the most vocal defender of equivocalism. One might even get a sense that the metaontology literature has been dominated by a dialogue on this very issue between Sider and Hirsch. But by now the reader should be aware that we are treading on difficult terrain. As our survey of the Quine vs. Carnap debate in the previous chapter showed, it is not always easy to determine the source of the disagreement – and if this is so in the case of metaphysics, the problem is even more acute in the case of metametaphysics.20 However, let’s give the stage to Hirsch before we pass judgement about this debate. 18

19

20

For a defence of this view, see Ted Sider,Writing the Book of the World (Oxford University Press, 2011). For a survey of all three options, see C. S. Jenkins, ‘What Is Ontological Realism?’, Philosophy Compass 5/10 (2010), pp. 880–890. A case in point due to Gerald Marsh, ‘Is the Hirsch–Sider Dispute Merely Verbal?’, Australasian Journal of Philosophy 88.3 (2010), pp. 459–469.

3.3 Quantifier variance and verbal debates

As we might expect from a proponent of equivocalism, Hirsch calls his position ‘roughly Carnapian’, although he immediately distances himself from Carnap with regard to a few issues. 21 One of these issues is that Hirsch fancies himself a realist, whereas he takes Carnap to subscribe to anti-realism. The sense of ‘realism’ that Hirsch has in mind is realism in the broadest sense described above, namely he does not think that our subjective, linguistic choices determine what exists in the world. Hirsch claims that our linguistic choices do, however, determine what we mean by ‘existence’. If we then take the meaning of ‘existence’ to correspond to existential quantification, it seems that the meaning of the existential quantifier can vary from language to language. This is the phenomenon of quantifier variance that we have already discussed in passing. As a point of interest, note that on the basis of quantifier variance, Hirsch would also disagree with Rayo’s metaphysicalist, because it is central to his view that none of the candidate quantifier meanings is metaphysically privileged in the sense that Rayo’s metaphysicalist might suggest. Quantifier variance is the first important feature of Hirsch’s view, but there is a second, albeit related feature of his view, namely the thesis that some ontological debates are ‘merely verbal’. More precisely, Hirsch argues that debates concerning the ontology of physical objects are merely verbal. One example is the debate about composition, which we discussed briefly in Chapter 2 – we will return to this debate below. What makes a debate merely verbal?22 One way that Hirsch puts this is that both disputants ought to agree that the other disputant speaks the truth in their respective language. Of course, when we are dealing with typical philosophical debates, it is quite unlikely that the disputants do in fact think that their opponent is speaking the truth, hence the ‘ought’. So, ‘merely verbal’ is used in a sense which suggests that the debate is not genuine, contrary to what the disputants think – it is not substantial. In its simplest form, this is easy to understand: a debate between two people is merely verbal if

21

22

Eli Hirsch, ‘Ontology and Alternative Languages,’ in Chalmers, Manley, and Wasserman (eds.), Metametaphysics, p. 231. We will here focus on Hirsch’s account. The literature on this topic is expanding rapidly though, see for instance Kathrin Koslicki, ‘On the Substantive Nature of Disagreements in Ontology,’ Philosophy and Phenomenological Research 71 (2005), pp. 85–105; David Chalmers, ‘Verbal Disputes,’ Philosophical Review 120.4 (2011), pp. 515–566; C. S. I. Jenkins, ‘Merely Verbal Disputes,’Erkenntnis 79.1 (2014), pp. 11–30.

51

52

Quantification and ontological commitment

they agree on objective facts, but use a different language to speak about them. If the same word means different things in different languages, then this may give rise to a very simple merely verbal dispute. If apples are ‘apples’ in your language and ‘oranges’ in someone else’s, then this is likely to cause some confusion. But once you’ve successfully translated their ‘orange’ into your ‘apple’, then the disagreement dissipates. Hirsch’s claim is stronger, though, as he wants to say that some debates are merely verbal quite generally. It’s not just that two people may have a merely verbal dispute; two groups of philosophers holding competing views can also have such a dispute. It is plausible that these two features of Hirsch’s view, quantifier variance and verbal disputes, are logically independent.23 Someone may defend quantifier variance without thinking that many ontological debates are merely verbal. Equally, someone may think that many ontological debates are verbal, but resist quantifier variance. Given that we have already discussed the complications related to the interpretation of quantifiers in some detail, it is the second aspect of Hirsch’s view that deserves more attention than the first here. We will return to quantifier variance and related issues in Chapter 4. Let’s get a little deeper into the verbal dispute issue. Hirsch makes some important qualifications to this view, the most important of which is the following: In my view, an issue in ontology (or elsewhere) is “merely verbal” in the sense of reducing to a linguistic choice only if the following condition is satisfied: Each side can plausibly interpret the other side as speaking a language in which the latter’s asserted sentences are true.24

What Hirsch insists is that a debate can fail to be merely verbal only if each side in a debate is genuinely charitable and still fails to interpret the assertions of the other side as being true in their respective language. This is, roughly, Hirsch’s principle of charity. But note that Hirsch is open to there being debates that are substantial – that is, debates that do not reduce to a linguistic choice. This is another aspect in which Hirsch considers himself 23

24

As proposed by Matti Eklund, ‘Review of Quantifier Variance and Realism: Essays in Metaontology,’ Notre Dame Philosophical Reviews (2011); see https://ndpr.nd.edu/news/24 764-quantifier-variance-and-realism-essays-in-metaontology/ Hirsch, ‘Ontology and Alternative Languages,’ p. 231.

3.3 Quantifier variance and verbal debates

to part ways with Carnap, whom Hirsch takes to make the sweeping claim that all ontological debates are merely verbal. Interestingly, Hirsch mentions the debate between Platonists and nominalists as an example of a debate that he considers not to be verbal. Recall that in Chapter 2 we used the debate between Platonism and nominalism as an example of a debate which a language pluralist might consider to be merely linguistic. In view of this, Hirsch’s position would seem to amount to a type of restricted or moderate language pluralism. Let’s try to understand this view a little better. The first issue to note is that a merely verbal dispute cannot just be one in which each side makes true assertions in their respective language. For, as Hirsch notes, there could be a merely verbal debate even in cases where the sides do not make true assertions: ‘There may be a verbal dispute about whether Harry is running around the squirrel or merely running around the tree containing the squirrel, when in fact Harry is home in bed and the runner is someone who looks like Harry’.25 Accordingly, it should be sufficient to consider a debate to be merely verbal if each side agrees that the other party speaks the truth in their respective language. It then becomes a matter of linguistic interpretation to determine whether a given dispute is indeed merely verbal. This would leads us towards the view that issues in metaontology can sometimes be settled in terms of linguistic interpretation. Hirsch does not put it quite this way, but this is close to a view that some philosophers working on the topic hold. One of Hirsch’s primary examples of a merely verbal debate is the one between perdurantists and endurantists. This is a classic debate about the spatiotemporal parts of material objects: everyone would agree that material objects have spatial parts, but do they also have temporal parts? In other words, how do objects persist through time? Perdurantists hold that material objects have temporal parts in addition to spatial parts and hence that objects ‘perdure’. Endurantists hold that material objects do not have temporal parts, which leads them to think that objects are wholly present at any time that they exist; they ‘endure’. Due to related issues concerning space and time, these views are versions of four-dimensionalism and three-dimensionalism, respectively. On the perdurantist view, material objects can be thought of as a type of ‘spacetime worm’, since all the 25

Ibid., p. 238.

53

54

Quantification and ontological commitment

different temporal parts are also, in some sense, parts of the object. In contrast, the endurantist view attaches no special significance to past or future temporal parts of an object – the object is there at a given time and that’s it. This may look like a perfectly normal ontological debate, presumably with some connection to physics, but it has become something of a test case in metaontology – primarily because some philosophers, like Sider, take it to be a substantial debate whereas some, like Hirsch, consider it merely verbal. Much ink has been spilled trying to settle the matter, but we will consider a slightly different approach. The reason for this is that the perdurantism/endurantism debate may not be a particularly good example of a substantial ontological debate in the first place. Why? Because there are plausible arguments to the effect that the theories are in fact metaphysically equivalent, as defined by Kristie Miller: [T]wo theories will be metaphysically equivalent if they are empirically equivalent theories for which there is a non-arbitrary mapping of the sentences of one theory onto the sentences of the other theory that preserves the truth values of those sentences. 26

Miller goes on to argue that perdurantism and endurantism, at least when combined with similar assumptions about composition and other ancillary issues, are equivalent exactly in this sense. We will skip her case for the empirical equivalency of the views, since this is more of an issue for (philosophy of) science. More importantly for our present purposes, Miller argues that there is no genuine difference in the ontological elements that the two competing theories postulate (even if there may be a difference in how they talk about these elements). We can understand the idea with the help of a simple illustration.27 When we watch TV, we see movement and think that we are perceiving motion. But in fact it is only an illusion of motion, since the movement on the TV just consists of a rapid sequence of static images. The upshot is that we may think of four-dimensional temporal parts succeeding one another as creating a similar illusion of motion rather than revealing anything about the true ontology of spacetime. Likewise, 26

27

Kristie Miller, ‘The Metaphysical Equivalence of Three and Four Dimensionalism,’ Erkenntnis (2005) 62.1, p. 92. For further details, see E. J. Lowe and Storrs McCall, ‘3D/4D Controversy: A Storm in a Teacup,’ Noûs (2006) 40, pp. 570–578.

3.3 Quantifier variance and verbal debates

the three-dimensional picture, which suggests that an object endures throughout the period of time it moves, can be understood simply as a different perspective regarding the same ontological elements. Entrenched proponents of perdurantism/endurantism may not be satisfied with this story, but we need not concern ourselves with this – it suffices that we have established how controversial the example is. But if we were to accept the metaphysical equivalency of perdurantism and endurantism, then would that not support Hirsch’s view? Not all the way: most if not all ontological realists would accept that some ontological debates are merely linguistic, but Hirsch extends the claim to many other debates concerning ordinary physical objects. For instance, Hirsch thinks that the debate about when a collection of material objects composes a further object is a prime example of a merely verbal debate. This debate about mereology and composition is one that we briefly touched upon in the previous chapter, in connection with van Inwagen’s ‘Special Composition Question’ (SCQ). The two alternatives that we discussed were universalism and nihilism – the first view takes it that a collection of material objects always composes a further object whereas the second view holds that composition never occurs.28 Hirsch specifies that ‘The dispute over nihilism that I view as verbal occurs when both sides agree that there are simples and disagree about whether there are composites’.29 In this case, there is perhaps more at stake than in the perdurantism/endurantism debate, as the proponents of nihilism will end up saying, for instance, that there are no such things as tables (there are just ‘simples arranged tablewise’), whereas the universalists must say that even something like the combination of your nose and the Eiffel Tower composes a further object. Hirsch himself thinks that on a charitable interpretation, the nihilists and universalists can agree, especially because they already agree that in ordinary language we talk as if there were things like tables and chairs but we don’t typically consider that any odd combination of material objects should be considered a further object. Of course, the proponents of these seemingly odd views can resort to the methodological move that we also discussed in the previous chapter, namely to insist that in the ‘ontology room’, the debate between nihilists 28

29

The SCQ originates in Peter van Inwagen’s Material Beings (Ithaca, NY: Cornell University Press, 1990). Eli Hirsch, ‘Physical-Object Ontology, Verbal Disputes, and Common Sense,’ Philosophy and Phenomenological Research (2005) 70.1, p. 68, fn. 2

55

56

Quantification and ontological commitment

and universalists can be taken seriously. But there may be a better move available, as Hirsch’s opponent only needs to provide some sort of plausible criterion for when composition occurs and then the debate turns on whether everyone accepts that criterion or whether there are other, competing criteria available. For example, one might suggest that composition occurs only when the combination of the various candidate objects possesses novel causal powers over and above of the causal powers of the candidate objects on their own.30 This type of criterion is testable: we can study each of the candidate objects to determine their individual causal powers and then study the presumed composite object that these candidate objects compose. If the causal powers of the presumed composite object are not merely a conjunction of the causal powers of the candidate objects, then the criterion under consideration would suggest that the composite object does indeed exist. Here we have a genuine difference between the ontological elements of two competing theories, namely the set of causal powers with which composite objects are endowed. Moreover, Hirsch’s condition for a merely verbal debate is not obviously satisfied, because the role of causal powers in the suggested theory is not something that can easily be considered simply as a different language or a different way of speaking of the same ontological elements. However, in fairness it should be noted that the theory regarding the causal powers of composite objects is certainly a controversial one – for one thing, it’s not clear which causal powers we should focus on (some causal powers could be considered trivial when it comes to composition). Unfortunately, the task of settling the Special Composition Question – or even whether it really is an ontologically substantial question – is far too great for us to undertake here, so we will have to just take it as a lesson in the convolutions of metametaphysics. The debate between Hirsch and Sider is ongoing. In his most recent rejoinder to Hirsch, Sider argues that ‘Ontologese is supposed to be just like English except for the semantics of quantifiers’.31 The idea is that one could suspend Hirsch’s requirement for charity, at least in those contexts where quantification is involved. Here it is worth noting again that we have been focusing on the status of quantification in our actual language. If we instead consider the best possible 30

31

For one version of such a view, although restricted to persons, see Trenton Merricks, Objects and Persons (Oxford University Press, 2001). Ted Sider, ‘Hirsch’s Attack onOntologese,’ Noûs 48.3 (2014), p. 567.

3.4 Beyond existence questions

language for the purposes of philosophy/science, then one could in principle agree with Sider – in the best language, ontological disputes are plausible, not merely verbal. But Hirsch (or someone else) might still insist that our actual language just isn’t like this: we are not in fact using the best possible language, Ontologese, but rather a much inferior one. In any case, Sider is primarily putting forward a conditional claim: if quantification ‘carves at the joints’, argues Sider, then it is not subject to the type of variation in meaning that some other terms, like ‘chair’ – and perhaps even composite objects in general – are. Sider thinks that the same goes for certain scientific terms, such as ‘electron’. However, we cannot really do justice to Sider’s positive project before we have a clearer idea about the metaontological machinery he uses, such as the appeal to ‘joint-carving’. But there is one reaction to this debate and its ilk that we still ought to take into consideration in this chapter, namely Fine’s influential claim that metaontology should not have existence questions as its primary focus.

3.4 Beyond existence questions We have seen that many metaontological debates concern questions about the existence of different types of objects. The debate about when a number of objects compose a further object is of this type as well, since we are looking for existence-conditions for composite objects. But the most commonly cited example is surely the existence of numbers, which we have already encountered repeatedly. The question ‘Do numbers exist?’ is a typically philosophical one, as it would rarely be asked outside the ‘ontology room’. The metaontological issue is whether such questions, which may seem trivial, are substantial questions at all. The number of papers written on the subject is certainly substantial and some philosophers think that we ought to reconsider the debate altogether. In the preceding discussion we have assumed that the manner in which Quine understands the task of ontology is at least broadly correct, even if there are several issues with the details. This understanding suggests that ontological questions are quantificational questions. So when philosophers ask ‘Do numbers exist?’, they mean to ask: ‘∃x(x is a number)?’ In what might be labelled a ‘neo-Aristotelian’ sprit, Fine has argued that rather than focusing on the quantificational question about the existence of numbers, the initial ontological question should be replaced with one

57

58

Quantification and ontological commitment

focusing on the nature of numbers. The core of this alternative approach is that the question about the existence of numbers is indeed trivial, at least in a sense, whereas there are still questions to settle regarding the ontological status of numbers. The way Fine puts this is that the question is about whether numbers are real: The realist and anti-realist about natural numbers, for example, will most likely take themselves to be disagreeing on the reality of each of the natural numbers – 0, 1, 2, … ; and this would not be possible unless each of them supposed that there were the numbers 0, 1, 2, … It is only if the existence of these objects is already acknowledged that there can be debate as to whether they are real (Quine’s error, we might say to continue the joke, arose from his being unwilling to grasp Plato by the beard).32

Fine suggests, controversially, that existence should be considered as a predicate: if we wish to ask whether integers exist, we should not formulate the question like this: ‘∃xIx?’, where ‘I’ refers to integers, but rather like ∀

⊃

this: ‘ x(Ix Ex)?’, where ‘E’ is the predicate for existence – the question is not whether some integer exists, but whether every integer does. Consider Fine’s example concerning realism about integers vs. realism about natural numbers. Fine takes it that realism about integers is intuitively the stronger position, because it is committed to the existence of both the integers and the natural numbers: if all the integers exist, then all the natural numbers do as well, because integers contain the natural numbers. Realism about natural numbers is clearly the weaker position, as it implies only a commitment to natural numbers. But on the existence-quantificational construal it turns out that realism about natural numbers is the stronger position: realism about integers claims that there is at least one integer (negative or not), whereas realism about natural numbers claims that there must be at least one non-negative integer, specifically a natural number. So Fine claims that the existence-quantificational account of ontological commitment gets the logic of ontological commitment wrong. This would seem to be an artefact of using the existential quantifier instead of expressing the thought that all the integers or all the natural numbers exist. As a possible rejoinder, Fine considers an appeal to a theory associated with a kind of object F, call the theory ‘TF’. TF states F’s existence conditions. 32

Kit Fine, ‘The Question of Ontology,’ in Chalmers, Manley, and Wasserman (eds.), Metametaphysics, p. 169.

3.4 Beyond existence questions

Now, if Fs are integers, then TF will state their existence-conditions, and it might be thought that Fine’s analysis was simply too detailed – just a commitment to TF will do if we want to express realism about integers, we need not even know the exact content of the theory. But it’s not at all clear how this would help, since both realists and anti-realists will surely believe in the truth of TF, if it is a true theory to begin with. The upshot of Fine’s analysis is that the existence-quantificational approach really can’t tell realism and anti-realism apart. Instead, Fine takes the realism/anti-realism debate to be concerned with which objects are real, and in order to have this debate in the first place it must be assumed

that the objects under investigation exist. The joke about Plato’s beard in the earlier quotation highlights just how different this approach is from Quine’s, since Fine takes it that the great puzzle that Quine was preoccupied with has a rather simple solution: realism in the usual sense must be assumed at the outset. By way of motivating Fine’s view, consider what follows if we agree that answers to quantificational questions are often trivial. Surely, answers to ontological questions had better not be trivial. Of course, there are non-trivial questions that fit this pattern, such as ‘Does the Higgs particle exist?’ This is a highly non-trivial and clearly important question. But since this question has nothing to do with metaphysics, the worry arises that all non-trivial existence question are answered by the special sciences, so what is there for philosophy to contribute? If Fine is correct, there are plenty of non-quantificational questions to address after we have settled the quantificational ones. Fine’s judgement is that Quine’s approach to ontology is based on ‘a double error’. On the one hand, Quine asks a scientific question (numbers are indispensable for science) rather than a philosophical one; on the other, Quine uses philosophical machinery (existence-quantificational tools) to answer that question. This analysis applies also to the other central questions of the metaontology literature that we have discussed, as Fine observes: I believe that the case of mereological sums and temporal parts has been especially misleading in this regard. For the question of their existence has often been taken to be a paradigm of ontological enquiry and, indeed, it is this case more than any other that has given rise to the recent resurgence of interest in meta-ontology. But the case is, in fact, quite atypical since it is

59

60

Quantification and ontological commitment

one in which the quantificational question is also philosophical and hence is much more liable to be confused with the ontological question[.]33

If Fine is correct, we may have done metaontology something of a disservice by having discussed the case of mereological sums and temporal parts even to the modest extent that we here have. However, given the extensive literature on these topics and the fact that these debates are by no means at its end, it will be useful for anyone interested in metametaphysics to have a basic grasp of the debates. Fine’s own analysis of the cases of mereological sums and temporal parts is as follows. At one point in history, equations such as ‘x2 = –1’ did not have a solution. But after the introduction of complex numbers, we have been able to solve such equations. In other words, we can extend the ‘domain of discourse’ to include such things like complex numbers. Fine suggests that the upshot of this is that there really is no substantive question as to whether there are complex numbers – we could have decided not to introduce them, after all. Similarly, Fine thinks that the existence of arbitrary mereological sums or temporal parts are subject to a decision about domain extension. If we wish, we can extend the domain of discourse in such a way that the mereological sum consisting of your nose and the Eiffel Tower is included. But it makes little sense to ask whether such strange objects really exist – the only substantive question that remains is whether the domain can be extended in such a manner (and it appears that the answer is ‘yes’). Hence, on Fine’s view, these questions do not appear to be particularly interesting or important questions in metaphysics.34 What remains to be done in this section is to outline Fine’s positive view. It was mentioned in passing that Fine thinks ontology should be primarily interested in what is real rather than what exists. This type of move is required, since when existence is taken as a predicate – as Fine proposes – realism in the usual sense becomes trivial. So if we continue to use the existential quantifier, we must introduce a ‘thick’ sense of existence to distinguish realism and anti-realism. But instead of talking about ‘existence’, which would admittedly be rather confusing given Fine’s approach, he proposes to use the term ‘real’ for this ‘thick’ sense 33 34

Fine, ‘The Question of Ontology,’ p. 159. But for a recent attempt to defend the importance of existence questions in metaphysics, see Chris Daly and David Liggins, ‘In Defence of Existence Questions,’ Monist 97.7 (2014), pp. 460–478.

3.4 Beyond existence questions

of existence. Accordingly, instead of being committed to the existence of numbers, a realist about numbers in Fine’s sense would be committed to numbers being ‘real’. A range of positions about the reality of ‘Fs’ can thus be formulated, where ‘R’ is the reality operator: 

Thoroughgoing realism:



Thoroughgoing anti-realism:



An intermediate position where some Fs are real and some are not; G being the dividing line:

∀x(F x ⊃

Rx).

∀x(F x ⊃ ∼Rx ).

∀x(F x ⊃

(R x

≡

Gx)).

In contrast, the quantificational account presents realism and anti-realism as contradictories with no room for an intermediate position: ∼∃xFx.

∃xFx

vs.

Fine concludes that the realist and anti-realist about natural num-

bers disagree about the reality of each natural number, but this would not be possible if natural numbers didn’t exist! Tempting as Fine’s account may be, the ‘thick’ reality operator is obviously in need of further explication. Fine does make an attempt to clarify the idea. First, he suggests that the reality operator R[…], where ‘…’ stands in for a sentence, expresses what is ‘constitutive of reality’. Metaphysics is concerned with how things stand in reality and thus this account helps explain how ontology is part of metaphysics. The objects of our ontology will be those that figure in ‘…’ – and this, Fine suggests, gives us a complete ontology. We can’t ‘beef up’ ordinary quantificational claims into ontological claims with R though, for instance, ‘R[∃xFx]’ – the prefixed claim is neither necessary nor sufficient for a commitment to Fs: It is not necessary since affirming the reality of an F is compatible with denying the reality of an existential fact of the form ∃xFx on the grounds that it is only the underlying particular facts which are real; and it is not sufficient since affirming the reality of there being an F would be compatible under a ‘bundle theory’ which only recognizes the reality of general facts with denying the reality of any particular fact of the form F x. And, similarly, it would seem, for any other prefixed claim.35

Fine’s suggestion is ingenious, but it may not be convincing to someone who isn’t sympathetic to realism to begin with. At the very least, we need to further clarify what realism itself amounts to, and we also need to hear more about the interpretation of Fine’s reality operator. Fine’s work on 35

Fine, ‘The question of ontology,’ p. 173.

61

62

Quantification and ontological commitment

grounding may be of some help in this regard, but we will postpone discus-

sion of this notion until Chapter 5.36 Fine, we should note, is not alone in holding the view that metaphysics should not focus on existence questions. This is also the view of Tim Crane, who puts it as follows: Recent philosophy has tended to express ontological or metaontological debates in terms of quantification. […] If I am right in what I have argued here [see sections 2.3 and 3.2 for Crane’s arguments], then this can’t strictly speaking be the central question, if what is at issue is the meaning of natural language quantifiers like ‘some’. Natural language quantifiers are not ontologically committing, so metaontology cannot be about the meanings of natural language quantifiers.37

In this regard, Crane is strongly opposed Sider’s suggestion, according to which the central question of metaontology is whether quantification is univocal or equivocal. This, as we saw, is indeed the central issue in Sider’s debate with Hirsch. What, then, is Crane’s positive view about metaontology? He is sympathetic to the line that Fine (and others, such as Jonathan Schaffer) takes, namely that ontology should be concerned with what is fundamental, what really exists. But as Crane notes, Sider’s view as well can

be translated into something similar: ‘the central question of metaontology is whether there are many different kinds of being (or existence), or whether there is one kind which is fundamental’. So Crane, Sider, Schaffer, and Fine, as well as many others, would all be in broad agreement about there being some kind of a privileged type of being – the fundamental type as opposed to the derivative type – which is of special interest in metaphysics. But this is a topic that we will postpone until Chapters 5 and 6. Note though that someone like Hirsch would be unlikely to think that any significant progress has been made here: why should we think that the debate about what really exists is any less verbal than the debate about what exists? Indeed, it’s not obvious that Fine and others have successfully undermined this type of deflationary reaction, which suggests that many or even all ontological debates are not substantial. This type of reaction comes in many

36

37

For further discussion, see also Kit Fine, ‘What Is Metaphysics?’, in T. E. Tahko (ed.), Contemporary Aristotelian Metaphysics (Cambridge University Press, 2012), pp. 8–25. Crane, The Objects of Thought, p. 51.

3.4 Beyond existence questions

varieties though and not all of these are compatible with Hirsch’s own picture – this topic is central in the next chapter. Now that we have covered the topic of quantification and ontological commitment from both a historical and a contemporary angle, it is time to move into a more general topic, namely that of identifying the different metaontological positions, such as realism, deflationism, and conventionalism.

63

4

Identifying the alternatives: ontological realism, deflationism, and conventionalism

This chapter examines various metametaphysical positions in a broader sense than the two previous chapters. At this point in the book, we do not yet have the full ‘toolbox’ of metametaphysics in use; but with a sense of the Quine vs. Carnap debate, the role of quantification, ontological commitment, and existence questions, we can put forward a rough classification of views. In the first section we will discuss ontological realism and ontological anti-realism. Ontological realism, in its various forms, is no doubt the dominant position. However, although the majority of metaphysicians call their view some kind of realism, we will see that the notion of ‘ontological realism’ can be specified in a number of ways, not all of which accurately reflect the views of those who claim to defend some form of realism. Typically, ontological realism is traced back to Quine whereas ontological anti-realism is thought to have Carnapian roots, but given the discussion in the last two chapters, we’d better not make such associations lightly – especially as it turns out that Quine and Carnap may not have disagreed as fundamentally as it is sometimes suggested. In this chapter, the historical context will be set aside, as the primary goal is terminological clarification. The second group of views, discussed in the second section, is the range of deflationary approaches to metaphysics. The notion of ‘deflation’ contrasts with ‘inflation’ – one way to understand this contrast is to think of the ‘seriousness’ that is attached to ontological questions. Accordingly, an inflationary view will ‘inflate’ the seriousness of ontological disputes, claiming that there is truly something of substance at issue when we have ontological disputes. This can go all the way up to the ‘full-blown’ – to extend the analogy – ontological realism typically defended for instance by ‘neo-Aristotelians’. Adeflationary view, in contrast, will ‘deflate’ the seriousness of ontological debates. However, as we will see, characterizing 64

4.1 Ontological realism and anti-realism

these views in such terms is not so straightforward. Some deflationist positions consider many of the typical ontological questions to be uninteresting or misguided, since there is generally an easy answer available. These answers may draw on simple conceptual and/or empirical work. This type of deflationism is sometimes labelled the ‘easy’ approach to ontology.1 But note that deflationism as such is not directly opposed to (all forms of) realism – as we saw in the previous chapter, Eli Hirsch is one example of a philosopher holding a type of deflationary position while also defending a (modest) form of realism. The third family of views to be discussed is opposed to realism: extreme conventionalism, to be discussed in the third section. Extreme conventionalism is explicitly defended only by very few philosophers, but it is not untypical to see various philosophers going some way towards such a view, or putting forward views which could be subjected to ‘slippery-slope’ arguments that lead to extreme conventionalism. In the fourth section, we will discuss one version of ontological realism in more detail, primarily as an exercise in assessing the relative merits of different metametaphysical positions. The focus here will be Ted Sider’s version of ontological realism. The chapter concludes with a brief assessment of the various options discussed.

4.1 Ontological realism and anti-realism One might think that defining ‘ontological realism’ should be relatively easy, at least if the notion is on a par with the notion of ‘realism’ without the qualification ‘ontological’, or perhaps replaced with ‘metaphysical realism’. Indeed, ‘metaphysical realism’, even though its precise formulation is certainly much debated, can at least be defined at such a high level of generality that most philosophers would be able to agree on its definition. At the highest level of generality, ‘metaphysical realism’ could be understood as the view that there is a mind-independent, objective reality. But disagreement enters the picture immediately when the epistemic dimension is opened, that is, when we start to discuss how we can know anything about that objective reality. When we move to metametaphysics, this epistemic dimension starts to cause disagreement even earlier. For instance, 1

See especially Amie Thomasson, Ontology Made Easy (Oxford University Press, 2015).

65

66

Identifying the alternatives: ontological realism, deflationism

we could attempt to understand ontological realism by comparing it with ontological deflationism of the ‘easy’ type mentioned in the opening of this chapter. The easy approach to ontology suggests that many or all ontological questions are epistemically easy to answer, whereas ontological realism is often, though not always, combined with the view (or rather, an attitude) that many or all ontological questions are difficult to answer. This appears to be primarily an epistemic issue. Alternatively, since some ontological deflationists hold that many or all ontological disputes are non-substantial or misguided, then ontological realism could be understood as the view that many or all ontological disputes are substantial or (synonymously) serious. But again, there is clearly an epistemic dimension to the question of whether a dispute is substantial or non-substantial. Moreover, it should be noted that there is no clear reason why one would be compelled to think that all ontological disputes must fall into just one group. One might, as an example, think that the dispute about numbers is not substantial, yet consider the dispute about three-dimensionalism vs. four-dimensionalism to be substantial. So if the notion of ontological realism is attached to the question of substantiveness, then one may be an ontological realist about some things and an ontological deflationist about others. It turns out, then, that ontological realism is not so easily defined. But let us set aside the contrast with ontological deflationism for a moment, since we will discuss it further in the next section. In any case, it would seem to provide only a negative definition of ontological realism. Can we define ontological realism positively? One way to attempt this would be to focus on the question of quantifier meanings, which was discussed in theprevious chapter. On Sider’s view, ontological realism amounts to the vi ew that there 2 is a ‘single best quantifier meaning’, that ‘carves at the joints’. In other

words, the meaning of the existential quantifier is not context-dependent, or even if it is, there is a way to pick out a meaning which is the best for the purposes of ontology. We discussed the debate between Sider and Hirsch on this issue of quantifier variance in theprevious chapter, but there are reasons to think that attaching the notion of ontological realism to the view that there is a single best quantifier meaning is not the most natural usage of the notion, as it is rather strict. For instance, those, such as the

2

Ted Sider, ‘Ontological Realism,’ in D. Chalmers, D. Manley, andR. Wasserman (eds.), Metametaphysics (Oxford University Press, 2009), pp. 384–423.

4.1 Ontological realism and anti-realism

‘neo-Aristotelians’, who do not regard existence questions to be at the heart of metaphysics would likely object to defining ontological realism strictly in terms of there being a single best quantifier meaning. But it does not seem unreasonable to think that there is something important that those influenced by Sider’s idea that quantification carves at the joints and those of the ‘neo-Aristotelian’ ilk are inagreement about. Presumably, they would agree about there being substantive ontological questions, so they are both opposed to (some forms of) ontological deflationism. But we’ve already set aside the attempt to define ontological realism merely as the view that opposes ontological deflationism. Perhaps there is a third sense of ontological realism that captures the relevant agreement? Unsurprisingly, the sense of ontological realism that we are moving towards must be very broad, but this is only to be expected of a label which ought to encompass a variety of views. Why ‘ought to’? Because it would be unreasonable for a relatively small group of philosophers to ‘steal’ a notion which, on the face of it, is quite general indeed. At the very least, we should attempt to find a sense of the notion that the mentioned groups might be able to agree about, since there is no doubt that they agree about something that certain other groups of philosophers would resist. Now, one

issue is that the area of agreement might be so basic that there is little to discuss here. This may be the result if we characterize ontological realism as a type of ‘knee-jerk realism’, which Sider describes as the view that the point of human inquiry is to conform itself to the world rather than to make the world.3 For Sider, knee-jerk realism is the starting point of inquiry – it is not separately argued for. But we don’t want ontological realism to come out as a primitive assumption that can’t be argued for. Rather, it seems that Sider’s ‘knee-jerk realism’ is metaphysical realism of a more traditional sort, to be separated from a more specific usage of ontological realism. In that case, the primitive assumption of metaphysical realism is something that we are already familiar with from the previous chapter, via Kit Fine’s argument to the effect that realism in ‘the usual sense’ has to be assumed.4 Hence, it seems that knee-jerk realism is too broad to capture the relevant sense of ontological realism.

3 4

Ted Sider, Writing the Book of the World (Oxford University Press, 2011), p. 18. Kit Fine, ‘The question of ontology,’ in Chalmers, Manley, and Wasserman (eds.), Metametaphysics, pp. 157–177.

67

68

Identifying the alternatives: ontological realism, deflationism

The three candidate senses for ontological realism that have been outlined so far correspond to a large extent with those discussed by Carrie Jenkins under the labels ‘ontological disputes are serious’, ‘one best quanti5 fier meaning’, and ‘ontological facts are objective’, respectively. To clarify,

the third view suggests that there are some ontological facts and they are objective (so one could reject this view either by claiming that there are no ontological facts or by claiming that they are not objective). As Jenkins argues, it is possible to hold various combinations of these three views, even though it sometimes seems that they are assumed to stand or fall together. Someone could very well have quarrels with the idea of there being one best quantifier meaning while accepting that there are objective ontological facts and that ontological disputes are generally serious (i.e., substantial). This goes some way against Sider, who seems to think that the seriousness of ontological disputes and the rejection of quantifier variance go hand in hand, so that to reject seriousness is to accept quantifier variance.6 As we saw in the previous chapter, Hirsch would likely disagree, but we do not need to dwell on the quantifier variance debate here. What is the upshot? Jenkins suggests, roughly in agreement with Karen Bennett and David Chalmers, that the best terminological practicewould be to use the label ‘realism’ for the most general of the three positions, namely ‘ontological facts are objective’.7 ‘Anti-realism’ will then be the rejection of this view, either by claiming that there are no ontological facts orby claiming that ontological facts are not objective. One motivation whichJenkins states in favour of this terminological practice is that we already have labels for the two other possible senses that could be attached to ‘ontological realism’. The first case we discussed – having to do with the seriousness or substantiveness of ontological debates– can be captured with the notion of ontological non-deflationism (or ‘inflationism’), to be discussed further in a moment. The second case, ‘one best quantifier meaning’, can be captured in terms 5

C. S. Jenkins, ‘What Is Ontological Realism?’, Philosophy Compass 5/10 (2010), pp. 880–890. 6 In fact, it’s slightly more complicated than this, as Sider’s case has to do with the importance of structure and ‘joint-carving,’ but we will return to this in more detail in the fourth section of the chapter. 7 Jenkins, ‘What Is Ontological Realism?’, p. 888; David Chalmers, ‘Ontological Anti-Realism,’ in Chalmers, Manley, and Wasserman (eds.), Metametaphysics, pp. 77–129; Karen Bennett, ‘Composition, Colocation and Metaontology.’ in Chalmers, Manley, and Wasserman (eds.),Metametaphysics, pp. 38–76.

4.1 Ontological realism and anti-realism

of the debate on quantifier variance; that is, Sider sides with the ‘quantifier invariantist’, who subscribes to the view that there is one best quantifier meaning, whereas Hirsch sides with the ‘quantifier variantist’. Given these labels, which are already in use, the notion of ‘ontological realism’ remains distinct from the two other issues. Jenkins’s suggestion seems reasonable and is consistent with the initial, broad usage of the notion that we adopted in the previous section. Hirsch likewise wishes to defend ontological realism in this broader sense, so the disagreement between Sider and Hirsch is not about ontological realism understood broadly. It seems that we have some good reasons to follow this terminological practice. But one issue still looms large: as noted earlier, there is the risk that, on this usage, ontological realism becomes an issue that we cannot easily argue for – a knee-jerk view that must be accepted at the outset, if at all. Related to this, consider Bennett’s reaction to anti-realism (Bennett’s anti-realist claims that there is no fact of the matter about whether or not there are Fs, for a given F): I do not know how exactly to argue against it, and I am not entirely sure what it means. ‘There are Fs’ might be vague or ambiguous in some way, in which case the unprecisified sentence might not have a determinate truth-value.8

This reaction leaves open the possibility that one could accept there to be a fact of the matter about whether or not there are Fs given an appropriate precisification; that is, given that we are able to make it precise, in what manner the relevant existence question is to be understood. Yet, one might think that this is such a difficult task that we will never be able to precisify the question in a satisfactory manner. Strictly speaking, this view would not appear to be anti-realism, however, as it is effectively being said that there at least could be a fact of the matter about whether or not there are Fs – we just don’t know. Bennett defines a closely related view called epistemicism: there is a fact of the matter about whether or not there are Fs, and

disputes about it are not merely verbal, but we have very little to justify our belief in either of the options.9 When laid out like this, epistemicism 8 9

Bennett, ‘Composition, Colocation and Metaontology,’ p. 40. Compare this with epistemicism about vagueness, the view that there is a sharp cut off for the applicability of vague predicates such as ‘is bald,’ but we cannot possibly know what that cut off is.

69

70

Identifying the alternatives: ontological realism, deflationism

is committed to ontological realism – the idea that ontological facts are objective – but claims that settling ontological questions may be very difficult, at least in some cases. This is not a very radical position, of course, and would likely be accepted by most. But Bennett herself is attracted to an even weaker version of this position, which is not committed to ontological realism. This weaker version of epistemicism suggests only that there are claims of the form ‘there are Fs’ that are not merely verbal, and further, that there is very little justification for believing either that these claims are true or that they are false. So this version of epistemicism is fully compatible both with ontological realism and anti-realism – it is a purely epistemic position.10 Bennett’s epistemicism (in either one of its versions) is certainly not our first glimpse of the intimate connection between metametaphysics and the epistemology of metaphysics, but here we see it in a very clear manner. It turns out that many metametaphysical positions make implicit epistemological assumptions, for example regarding our ability to know whether a given dispute is substantial and when it has been settled if it is substantial. Bennett has something further of interest to say in this connection. She focuses on two popular test cases in metaontology, one which we have already discussed: the problem of material constitution (can two objects, e.g., a lump of bronze and a bronze statue, spatiotemporally coincide?) and Peter van Inwagen’s ‘Special Composition Question’ (when do a number of objects compose a further object?). Bennett’s claim is that epistemicism is a reasonable position regarding these popular problems. Here’s why: we have already done more or less all the work that we can on these topics! 11 So Bennett’s idea is that since so many clever minds have worked on these topics for so long and we still don’t seem to have a determinate answer, it’s unlikely that an answer is forthcoming. Hence, we should adopt epistemicism regarding these problems. She does however admit that someone might still come up with a new, persuasive line of reasoning, which would change our epistemic situation. But this does admittedly seem fairly

10

11

For a view which takes there to be no fact of the matter about whether or not there are Fs, hence taking existence questions to lack truth-value, see Stephen Yablo, ‘Must Existence-Questions have Answers?’, in Chalmers, Manley, and Wasserman (eds.), Metametaphysics, pp. 507–525. Ibid., p. 73.

4.2 Ontological deflationism

unlikely with regard to these two problems which have, after all, received the attention of our leading metaphysicians for quite some time. Note that all this is neutral with respect to ontological realism vs. anti-realism, on Bennett’s intended reading. Since we have little reason to go one way or the other regarding these problems, the debates are simply underdetermined, they take no stand on whether or not there really are objective facts. Having said that, the anti-realist could get some mileage from this situation: isn’t it grist to the anti-realist’s mill that we have not been able to settle these debates? What other evidence could there even be for anti-realism? Is it not unreasonable to require that the anti-realist is able to demonstrate that there could be no objective fact of the matter? Perhaps so, but this line of argument is unlikely to persuade the ontological realist, since the realist does think that she is able to make at least some progress regarding these problems – they are at the very least tractable. Of course, if Bennett is correct, then these debates are considerably less tractable than the ontological realist may have led us to believe, but this does not automatically vindicate the anti-realist position. So if we follow Bennett, we should perhaps remain agnostic about these debates, at least until further evidence emerges.

4.2 Ontological deflationism Bennett’s epistemicism – although it is non-committal about ontological realism – rules out by definition the view that the ontological debates concerning composition and spatiotemporal co-location are merely verbal disputes. But if we cannot make any epistemic progress regarding these debates, isn’t it more likely because these debates are merely verbal rather than it just being exceedingly difficult to settle them? This, at any rate, is a line of thought that an ontological deflationist could entertain. A quantifier variantist like Hirsch could argue to this effect. However, since we have already discussed the debate between Sider and Hirsch as well as Fine’s reaction to this and similar debates in the previous chapter, we can now focus on another branch of deflationism, due to Amie Thomasson. But first, let us briefly consider ontological deflationism more generally: can we easily define it? One attempt at a general definition of the core content of ontological deflationism would be in terms of relative triviality: where most scientific

71

72

Identifying the alternatives: ontological realism, deflationism

discoveries involve finding out something new about the world, something empirically tractable, ontological discoveries often seem to lack this type of significance – indeed, it is not even appropriate to talk about ontological discoveries. If we know all the properties of a lump of bronze, which hap-

pens to be in the shape of a statue, there seems to be little at stake when we ask whether the lump and statue are two different things spatiotemporally co-located or just one thing. The answer, one might think, reflects only our semantic, conceptual, or linguistic preferences. While the preceding is not exactly a definition of ontological deflationism, it should at least make clear what motivates the view: deflationism is the type of reaction that most of us engaged in metaphysics have heard when trying to explain what it is that we metaphysicians actually do to those not trained in the discipline (and indeed sometimes even to those who are trained in it!). The issue is not, of course, quite as simple as this. For one thing, we should note that the somewhat sceptical reaction of the deflationist does not automatically lead to ontological anti-realism. Even if the answers to some ontological questions are trivial or merely a matter of semantic preference, it does not mean that there aren’t objective facts concerning them. This is easy to see in the case of trivial answers, for we can supposedly easily agree on what the fact is in these cases. The case of semantic preference might seem to lead towards anti-realism, though – at least the lack of any objective fact would perhaps explain why there is a disagreement, say, about the co-location of the statue and the lump. But while this type of semantic relativism is certainly compatible with anti-realism, it does not entail it. This should already be clear from our discussion of epistemicism above, as we’ve seen that it is possible to remain neutral about the ontological realism vs. anti-realism issue on epistemic grounds while developing, say, a view about one’s semantic preference towards a given ontological question. However, note again that Bennett’s epistemicism rules out this type of semantic deflationism by definition; that is, her epistemicism assumes that (some or many) ontological disputes are not purely verbal or relative to semantic preference. So it is possible to separate ontological deflationism from the ontological realism vs. anti-realism issue, but as a matter of fact, many ontological deflationists, such as Hirsch and Thomasson, take themselves to be committed to some version of ontological realism. One way to distinguish the type of ontological realism that deflationists may adopt and the

4.2 Ontological deflationism

type of ontological realism that we’ve seen Fine and Sider defend would be to talk about lightweight and heavyweight realism, respectively, following Chalmers.12 But since we’ve seen that it might be best to define ontological realism simply as the view that there are objective ontological facts, this distinction is perhaps not entirely helpful. Chalmers suggests that heavyweight realism comes with the additional commitment that ontological questions are highly non-trivial, but this seems an unnecessary complication, given that we already have the notion of ontological deflationism (and non-deflationism or inflationism), which corresponds with the lightweight vs. heavyweight distinction. Accordingly, we will follow this more established usage in what follows and leave the lightweight vs. heavyweight distinction aside. Despite its common-sense appeal, ontological deflationism faces an initial challenge: if we are to believe that many ontological questions – existence questions in particular – have a trivial answer, then how does the ontological deflationist explain the fact that we often disagree about what that answer is? A partial explanation might be that different people have different semantic preferences, perhaps due to cultural norms and so forth. Yet, philosophers typically claim to have arguments in favour of answering certain existence questions either positively or negatively. Indeed, one might think that existence claims could never be trivial, unless they turn out to be analytically true or false. But it is difficult to see how existence claims could be analytically true or false. These types of considerations partly motivate Chalmers to favour a type of anti-realism, but the same considerations could of course motivate a version of non-deflationism as well. Ontological deflationism is not quite so easily refuted though. As we’ll see in a moment, Amie Thomasson provides a version of deflationism whereby existence claims are not analytically true or false, but nevertheless turn out to have ‘easy’ answers. 13 Thomasson’s version of ontological deflationism does not rely on quantifier variance, in contrast to Hirsch. This is one reason why we should distinguish between ontological deflationism andquantifier variance:one can be a deflationist without being a quantifier variantist (whether the opposite is possible, i.e., to be a quantifier variantist without being a deflationist, is more 12 13

Chalmers, ‘Ontological Anti-Realism,’ p. 78. For the complete version, see Amie Thomasson, Ontology Made Easy. For a shorter overview, see her ‘The Easy Approach to Ontology,’ Axiomathes 19 (2009), pp. 1–15.

73

74

Identifying the alternatives: ontological realism, deflationism

controversial). On Thomasson’s view, ontological questions do have answers (they are not moot), but they are not worthy of ‘serious’ debate because the questions are typically not at all difficult to answer. So how do we answer them? Usually the answer is reached with the help of a combination of conceptual and empirical tools. Importantly, Thomasson regards these answering strategies to be so straightforward that they do not warrant philosophical attention. Of course, this is not to say that allconceptual and empirical enquiries are easy, but according to Thomasson there are reasons to think that the typical philosophical existence questions that we have been discussing in this book certainly fall into this group of ‘easy’ questions. One of these reasons, or what at least motivates the reason, is metasemantic: Thomasson holds that there is a connection between ‘exists’ and ‘refers’ in such a way that the truth-values of existence claims are a semantic matter (concerning reference), whereas existence claims themselves are ‘about’ the world. The real work, then, is done at the level of semantics, which demands a theory about the conditions under which our terms refer. A crucial part of Thomasson’s approach is thus the project of detailing the conditions under which a term may be properly applied, calledapplication conditions. We need not go into all the intricacies of Thomasson’s metasemantic approach, but it might be helpful to consider an example of how her approach delivers the easy answers. Thomasson is mainly concerned with sortals, so let us take a sortal such as ‘kangaroo’ as our example. In answer-

ing the question ‘Do kangaroos exist?’ we ought to start by considering the application conditions for the term ‘kangaroo’. It seems clear, for instance, that the conditions we normally associate with ‘kangaroo’ will be fulfilled in a situation where we observe a large group of moderately sized brownish mammals hopping on two well-developed legs in the Australian outback. We might further engage in some empirical research to ensure that what we are observing are indeed kangaroos rather than, say, wallabies, but this can be quite straightforwardly determined, ultimately by a DNA analysis if not otherwise. So it seems as if we can conclude, without any great difficulty, that kangaroos exist. A similar strategy can be applied to most sortals, such as ‘chair’ or ‘table’, and indeed to abstract objects and social constructions such as ‘symphony’ and ‘corporation’.14 Ontological debates

14

For further discussion, see Amie Thomasson, Ordinary Objects (Oxford University Press, 2007).

4.2 Ontological deflationism

concerning the existence of ordinary objects would thus seem to dissipate, but surely there are cases where such easy answers are not forthcoming? There are indeed cases in which Thomasson’s strategy does not provide an easy answer. But she doesn’t regard this to be a flaw in the approach, but rather an indication of an unanswerable question. On this view, any properly formulated existence question can be answered, but the ‘proper formulation’ requires that we are able to determine the relevant application conditions. On Thomasson’s view, some of the problematic ontological disputes involve terms for which application conditions have not been determined and in fact cannot be determined, often because they involve a usage of the term ‘object’ in such a way that application conditions for ‘object’ cannot be determined (but note that it is controversial whether ‘object’ can be understood as a sortal). For instance, when the nihilist about composition – whom we encountered in Chapter 2 – says that there is no object composed of the various parts of a table, but rather just particles arranged tablewise, she might have different application conditions for ‘object’ in mind than the universalist. More generally, an ontological dispute may turn out to be merely verbal if the disputants have different application conditions in mind (cf. quantifier variance). The upshot of Thomasson’s easy approach is that things which exist according to common sense, such as kangaroos, tables, chairs, and symph onies, can easily be determined to exist, whereas dragons and unicorns do not exist, because they lack sufficiently well-defined application conditions. For instance, we could never complete the empirical part of the test required to determine whether something fulfilling the application conditions for the term ‘dragon’ exists, for there is no DNA test for dragons! Note however that there may be application conditions for various fictional objects as well, including fictional ‘dragons’. But for the purposes of the example we may assume that on one, non-fictional understanding of ‘dragon’, a part of the application conditions for the term would involve that dragons are some sort of an animal. Yet, according to Thomasson there is nothing particularly interesting, philosophically speaking, about these existence claims. Thomasson suggests that the deflationism resulting from the ‘easy’ approach, whereby most existence questions can be straightforwardly answered with the combination of conceptual and empirical tools, is ‘Carnapian’. This is partly because the focus is on language: existence claims must be made using a language and once the relevant application

75

76

Identifying the alternatives: ontological realism, deflationism

conditions for the terms in that language have been settled, the existence claims are ‘internal’, more or less in the Carnapian sense. We do not need to return to the complications involved with the interpretation of the Carnapian position and the notion of ‘framework’, which were discussed extensively in the preceding chapters, but it is not difficult to see that Thomasson’s approach could be appealing to those with Carnapian sympathies. The view does have a clear point of connection with language pluralism, as discussed in Chapters 2 and 3. Be that as it may, we now have a fairly good idea about what ontological deflationism along Thomasson’s line amounts to, if true: many or most ontological disputes, at least the ones concerning existence claims of the usual type, are relatively easy to settle. In some cases it may turn out that the dispute is not even substantial, if the disputants have been using different application conditions for the same term. But one advantage that Thomasson’s approach seems to have is that even in these cases it is not all that difficult to come to an agreement once the relevant application conditions are explicitly stated. Of course, this still leaves the cases where it just isn’t possible to specify application conditions in a satisfactory manner – Thomasson’s judgement about these cases is that they are ‘unanswerable’, questions that on a closer look turn out to be pseudo-questions. 15 But what if most of ontology turns out to be confused in this manner? What if there really are no genuine ontological questions, just pseudo-questions? What if we can carve up the world any which way we like? Then we would end up with a more extreme type of ontological deflationism: full-blooded conventionalism. This is the view that we will examine in the next section.

4.3 Towards extreme conventionalism Extreme conventionalism is not usually listed as one of the main contenders amongst metametaphysical positions, nor is it typically discussed in much detail separately from ontological anti-realism. But there are some good reasons to discuss the view separately. For one thing, even though anti-realism may naturally go with extreme conventionalism, it is implied by it only if combined with some further assumptions. To see this, we will

15

See Amie Thomasson, ‘Answerable and Unanswerable Questions,’ in Chalmers, Manley, and Wasserman (eds.),Metametaphysics, pp. 444–471.

4.3 Towards extreme conventionalism

have to lay out the position in more detail. First, let’s focus on a version of conventionalism that starts from the apparent subjectivity of our manner of classifying reality. A realist would think that even though some of our classificatory efforts are indeed guided by mind-dependent reasons, such as different pragmatic considerations, the majority of them are nevertheless mind-independent. In other words, there really are such things as apples, cats, mountains, and stars, and they have what we might call natural boundaries, something that separates them from other things. The realist would

consider the boundaries of these objects and kinds to ‘carve nature at the joints’. It is not always easy to state the exact identity conditions of various things that we believe to have natural boundaries, but the extreme conventionalist thesis is that there are no such mind-independent identity conditions and that all our efforts to determine the boundaries of things are subjective.16 On this basis, it may seem that anti-realism immediately follows, but this is not quite right, for it would be possible to maintain that even if we are fundamentally unable to overcome our conventions and hence state mind-independent identity conditions for anything whatsoever, this could simply be due to our psychological limitations. Thus, one might adopt something along the lines of Bennett’s epistemicism, withholding judgement about realism vs. anti-realism. Perhaps this would be a pyrrhic victory, though; honest ontological realism surely requires that we can, at least sometimes, reach correct judgements about the structure of reality, to carve at the true joints of reality. What makes extreme conventionalism a metametaphysical position rather than just a first-order metaphysical position? Well, one upshot of extreme conventionalism is that many of our ontological debates turn out to be non-substantial. For instance, the debate about composition to which we keep returning would, on the conventionalist analysis, be merely due to a clash between differentconventions about composition. Moreover, there is nothing objective in the world that settles this clash between conventions, for they have been adopted for some mind-dependent reason, such as a cultural norm or a pragmatic choice. So the view at hand has metametaphysical implications that are very similar to quantifier variance, but the motive 16

For an entertaining construal of this type of view, see Achille C. Varzi, ‘Boundaries, Conventions, and Realism,’ in J. K. Campbell, M. O’Rourke, and M. H. Slater (eds.), Carving Nature at Its Joints: Natural Kinds in Metaphysics and Science (Cambridge, MA: MIT Press, 2011), pp. 129–153.

77

78

Identifying the alternatives: ontological realism, deflationism

is very different. Indeed, we could discuss conventionalism under the rubric of deflationism as well, distinct from both Hirsch’s and Thomasson’s versions, but it is perhaps better to further distinguish it from these deflationist views, as the roots of the views are somewhat different. What could possibly motivate extreme conventionalism? One argumentative strategy relies on the fact that some boundaries are clearly not natural, but artificial: borders of countries are one obvious example. We may also express the artificial/natural boundary distinction in terms of fiat and bona fide boundaries, following Barry Smith and Achille Varzi.17 It is import-

ant to recognize here that the fiat/bona fide distinction applies equally to the physical boundaries of objects and to the objects themselves: according to the realist, the physical boundary of an apple, for instance, is a bona fide boundary, and the individual apple is a bona fide entity. However, this is exactly what the extreme conventionalist questions: when we look at the apple closely enough, it is clear that, far from the smooth boundary that it appears to have, we are in fact dealing with a very loose arrangement of molecules, and further, with a swarm of subatomic particles. Moreover, the debate concerning composition further supports the conventionalist’s line of thought: when Tibbles the cat eats some fish, at what point does the fish become a part of Tibbles? Which criteria we apply, the extreme conventionalist will argue, is ultimately a matter of fiat; Tibbles may continue to exist, but his identity conditions are not mind-independent. Varzi takes this line of thought to its logical conclusion: Consider the debate on unrestricted composition. There is no question that we feel more at ease with certain mereological composites than with others. We feel at ease, for instance, with regard to such things as the fusion of Tibbles’ parts (whatever they are), or even a platypus’s parts; but when it comes to such unlovely and gerrymandered mixtures as trout-turkeys, consisting of the front half of a trout and the back half of a turkey, we feel uncomfortable. Such feelings may exhibit surprising regularities across contexts and cultures. Yet, arguably they rest on psychological biases and Gestalt factors that needn’t have any bearing on how the world is actually structured. 18

17

Barry Smith and Achille C. Varzi, Fiat ‘ and Bona Fide Boundaries,’ Philosophy and Phenomenological Research 60.2 (2000), pp. 401–420.

18

Varzi, ‘Boundaries, Conventions, and Realism,’ pp. 144–145.

4.3 Towards extreme conventionalism

Varzi’s analysis may help us to get our feet back on the ground: is there really any connection between how the world is structured and our evaluation of things such as trout-turkeys? Varzi argues that there might not be, since even though we initially feel uncomfortable about strange hybrids, we have nevertheless welcomed, for instance, a variety of genetically manipulated plant-hybrids, such as orange-mandarins and peppermint. Indeed, our intuitions and feelings of discomfort should not be relied on if we hope to determine the actual structure of reality, for it is true that we are biased in our evaluations. Once again, we would appear to be on our way towards a position not unlike the deflationism emerging from language pluralism. But this time the view arises from the simple realization that our judgements are thoroughly influenced by our psychological biases. However, even though Varzi makes a commendable effort to entertain the extreme conventionalist line of thought, it does become less plausible when we move from everyday examples and cases such as the debate over composition to scientific examples. Consider the Standard Model of particle physics and its spectacular success in producing accurate predictions. The Standard Model gives us a list of fundamental particles that do not seem to be subject to the type of argument that Varzi presents; that is, we can ‘look’ at electrons as closely as we want, but if electrons are truly fundamental particles, then we will not discover any substructure, anything that might cause us to think that there are various candidates as to how the boundaries of the electron are to be determined. Of course, if it turns out that electrons are not fundamental, then we might have to look closer into their substructure to find truly fundamental particles – if there are any (we will return to the question of fundamentality inChapter 6). Moreover, there are various ways to interpret the relevant physics, but none of these interpretations would seem to support extreme conventionalism. Even views according to which there is strictly speaking no particles, but merely some structure or relations, as certain versions of structuralism hold, would fall short of extreme conventionalism. Although the world might turn out to be merely a lump of dough on these views, it would not be an amorphous lump, as the lump would have internal structure, some ‘joints’ according to which we may structureit.19

19

For further discussion, see Tuomas E.Tahko, ‘Boundaries in Reality,’Ratio 25.4 (2012), pp. 405–424. For a brief discussion of the notion of ‘amorphous lump,’ see Michael

79

80

Identifying the alternatives: ontological realism, deflationism

Accordingly, extreme conventionalism turns out to be a difficult to view to hold consistently. But perhaps a slightly less extreme version of conventionalism is more plausible? One version is due to Alan Sidelle, whose conventionalism starts from modality, but ultimately extends also to objects: Sidelle thinks that if modal properties are conventional, then so are objects.20 Sidelle’s conventionalism may not be as extreme as the one that Varzi entertains, but its metametaphysical implications are severe. We cannot do justice to Sidelle’s detailed arguments here, but what we can do is consider a test case and determine its metametaphysical implications. According to conventionalism about modality, what is possible or necessary is so because of the way we think about the world; our modal judgements are based on conventions. So how does a conventionalist about modality treat typical cases of metaphysically interesting modal truths, such as the case of ‘Water is H2O’, which is commonly considered a metaphysical a posteriori necessity? Sidelle puts it as follows: [E]ach necessary a posteriori truth should be seen as derived from a combination of an analytic principle of individuation that has empty spaces to be filled in by empirical findings and a particular empirical finding that of itself carries no modal weight. For example, in the case of water’s being necessarily H2O, the analytic principle might be ‘Nothing counts as water in any situation unless it has the same deep explanatory features (if any) as the stuff we call ‘‘water,’’’ and the empirical fact, which makes the result a posteriori, is that the deep explanatory feature of the stuff we call ‘water’ is being composed of H2O.21

On Sidelle’s view, each example of the necessary a posteriori can be divided into an a priori principle concerning the type of empirical fact required to generate a necessary truth – the deep explanatory features – and the empirical discovery of the relevant fact. Therefore, ‘the modal force’ of a posteriori necessities comes from a priori principles, which Sidelle

20

21

Dummett, Frege. P hilosophy of Language, 2nd edn (Cambridge, MA: Harvard University Press, 1981), p. 577. For the connection between conventionalism about modality and objects, see Alan Sidelle, ‘Modality and Objects,’ The Philosophical Quarterly 60.238 (2010), pp. 109–125. Alan Sidelle, ‘On the Metaphysical Contingency of Laws of Nature,’ in T. S. Gendler and J. Hawthorne (eds.), Conceivability and Possibility (Oxford University Press, 2002), p. 319.

4.3 Towards extreme conventionalism

claims to be analytic principles that are true in virtue of our linguistic conventions. There is an interesting complication to this story: what if the conventions themselves had been different, for instance, what if the convention that guarantees the necessity of ‘Water is H2O’ had been different? Then it would seem that either water might have failed to be H2O or water might not have necessarily been H 2O.22 This issue has one especially interesting metametaphysical implication concerning the analysis of changing conventions. If conventionalism (about modality) is true, then metaphysical debates concerning modality will presumably have different outcomes when associated with different conventions. That is, if something like Sidelle’s story is correct and ‘the modal force’ of necessary truths comes from conventions, then a change in those conventions ought to have an influence on the modal truths. Yet, as Sidelle argues, this is not quite so simple.23 To illustrate the problem, let us stipulate that the term ‘water’ refers to the macrophysical, superficial qualities of water, such as being a clear, tasteless liquid, etc. (let us ignore the various complications with such phenomenal qualities for the moment). We can now imagine a counterfactual scenario where another liquid with a different microstructure, say, XYZ, as in Hilary Putnam’s traditional ‘Twin Earth’ scenario, falls under the extension of ‘water’ due to having the same superficial qualities. Now it seems that in the counterfactual scenario it would be true to say that ‘There is water which lacks hydrogen’, but Sidelle insists that we could not translate this sentence directly to our normal usage of ‘water’, as there must be different rules governing the use of ‘water’ in effect here. However, this failure of translation is not as radical as using the sound ‘water’ (or its equivalent in other languages) for something quite different, like overcooked spaghetti (Sidelle’s example!), since the usage of ‘water’ in the counterfactual scenario is, after all, very much like our use of ‘water’, given that we are referring to the superficial qualities that XYZ also has. What is the upshot? Well, it appears that there are different conventions governing the use of ‘water’ for us and for the Twin Earth folk. So we cannot 22

23

This type of complication concerning conventionalism, about modality and possible strategies to address it, is discussed in Alan Sidelle, ‘Conventionalism and the Contingency of Conventions,’ Noûs 43.2 (2009), pp. 224–241. Ibid., pp. 233 ff.

81

82

Identifying the alternatives: ontological realism, deflationism

simply translate their true statement ‘There iswater which lacks hydrogen’ to our equivalent statement, which is false.Thus, the difference in conventions does not have any implications towards the essence of water after all. But according to Sidelle, this doesnot undermine conventionalism, for it demonstrates that we could have used thesame term in very similar circumstances to generate different truth values for the same counterfactual and modal sentences. What is crucial in this account is that Sidelle thinks we would not have been making a metaphysical mistake despite the different truth-values. Thus, the conventionalist can argue that regardless of whether or not there are mind-independent modal facts, it is the conventions themselves that generate ‘the modal force’; either there simply are no modal (or essentialist) facts or if there are, then we can use ‘water’ just as well to refer to one such fact or the other, without making any mistake (of a metaphysical sort). In the latter case, we would end up with what Sidelle calls ‘metaphysical universalism’: The Conventionalist notes that sincewe could have chosen– metaphysically speaking – more or less any feature toguide our counterfactual usage, and this then would have generated other, but parallel, true (and false) modal sentences, then this story is going tohave to generate a sort of metaphysical universalism, according to which whatever conventions we adopted, there would have been a ‘mirror’ so-individuated entity or property to which these rules determined reference.24

According to Sidelle, the fact that varying conventions fail to have modal implications is due to a failure of translation, thus at the level of language rather than ontological fact. Although, as Sidelle admits, if ‘metaphysical universalism’ is true, then there is still a sense in which the realist would be correct in saying that the source of modality is in the world – since there are modal facts after all. But Sidelle thinks that this would still be compatible with a type of conventionalism about modality, since any debate about modal facts would nevertheless be answered with reference to our conventions rather than the objective facts.

This compact presentation of Sidelle’s conventionalism about modality is too brief to enable us to fully articulate the upshot of the position, but for the purposes of our overview of various metametaphysical positions in this chapter, we can already clearly see what kind of implications this line 24

Ibid., p. 235.

4.4 A case study: Sider’s ontological realism

of thought would have. Interestingly, one of these is, once again, a type of epistemicism: Sidelle ends up accepting, although reluctantly, the possibility of a position according which there could be objective modal facts, even though he insists that they would be inert when it comes to actual debates concerning modality.

4.4 A case study: Sider’s ontological realism As we saw earlier in this chapter, Sider understands ontological realism somewhat differently (that is, much more strictly) from the broad sense that we have adopted. But looking into his view in somewhat more detail than we have done so far might be useful, for Sider’s work in this regard has been very influential and anyone interested in metametaphysics is likely to encounter his ideas repeatedly. Moreover, now that we have a clearer idea about the various metametaphysical positions, we can apply our understanding to Sider’s theory and assess it in the light of some of the competing views. In his recent book, Sider sets up an ambitious project in favour of realist metaphysics. 25 Sider’s central task is to defend realism about structure , which he connects with the notion of fundamentality : ‘a fact is fundamental when it is stated in joint-carving terms’.26 We will postpone an analysis of fundamentality until Chapter 6, but all we need to know at the moment is that, for Sider, fundamentality is a metaphysical notion, which he uses ‘more or less interchangeably’ with ‘joint-carving’ and ‘part of reality’s structure’. He wants to defend the view that some notions are fundamental in the sense that there is a perfectly fundamental description of reality expressible by terms that carve perfectly at the joints. Note also that Sider is committed to a type of ‘levels’ of reality view, as he thinks that there are degrees of fundamentality and hence degrees of joint-carving – perfect joint-carving being at the fundamental level. All this talk about levels and fundamentality will be clarified in Chapter 6, but let us try to make some progress with regard to the notion of ‘joint-carving’, which is often used in these contexts and which we have already come across in this book.

25

Sider, Writing the Book of the World.

26

Ibid., p. vii.

83

84

Identifying the alternatives: ontological realism, deflationism

The srcin of the notion of ‘joint-carving’ is Plato’s dialogue Phaedrus, which talks about dividing things by classes and natural joints. 27 The idea could be described as follows. If one thinks that there is some structure in reality, as Sider clearly does, then one way to describe this structure is in terms of the ‘joints’ of reality. If we hope to talk about this structure, to pick out these joints, then we must have some terms or concepts that ‘latch on’ to – that is, accurately describe – the joints. In other words, there are some concepts that ‘carve’ nature or reality at its joints. Moreover, as noted already, some concepts supposedly do this carving ‘perfectly’. What this amounts to is a more difficult question, since we have not yet laid out the full metametaphysical ‘toolbox’, but the driving idea is that some concepts capture something about the most basic structure of reality. The role of perfectly joint-carving terms in Sider’s theory is complex, but here we are primarily interested in their role in securing the substantivity of metaphysical questions. Sider regards it as a ‘near enough’ sufficient condition for substantivity that a debate is cast in perfectly joint-carving terms.28 Why only ‘near enough’? Because there may be multiple joints in the vicinity (more on this later), or the joint-carving candidate may be of the ‘wrong sort’: The line through the Urals, though physically distinguished, isn’t geographically or politically distinguished, and therefore seems irrelevant to the boundaries of Europe. The physically significant line is indeed a joint in nature, but it’s the wrong sort of joint in nature, relative to a dispute over ‘Europe.’29

Sider goes on to offer a suggested answer as to what qualifies as the ‘right sort’ of joint-carving candidate – and we will look into this answer in what follows. But he does not think that this is a major problem, since he decides to mostly ignore this complication. Sider’s central concern is to build a case in favour of (what he calls) ontological realism, and against deflationism. Sider’s definition of ontological realism, which we will call ‘Siderian ontological realism’ to distinguish it from the use we adopted earlier, is roughly as follows: 27

28 29

Plato, Phaedrus, translated by R. Hackforth (Cambridge University Press, 1972), 265d–266a. Sider, Writing the Book of the World, p. 71. Ibid., p. 48.

4.4 A case study: Sider’s ontological realism

Siderian ontological realism (S-OR) Ontological realism says that the crucial terms in ontological questions carve perfectly at the joints.

What are the crucial terms? Reflecting our earlier discussion, Sider takes them to be quantificational . In fact, he states that ‘the thesis that quantifiers carve at the joints is the best way to defend the substantivity of ontological questions’.

30

Ontological assertions are often of the form

‘∃xFx ’, or, ‘There are Fs’. Thus, the first-order existential quantifier is a crucial term, and carves at the joints. If F also carves at the joints, then the question of whether there are Fs is ‘guaranteed’ to be substantive; so there is a tight connection between S-OR and quantification. In particular, Sider hopes to defend S-OR against deflationism by arguing that the first-order existential quantifier carves (perfectly) at the joints. However, Sider’s project is primarily methodological, as he is not concerned about what the Fs are. Hence, we could attempt to reconstruct Sider’s key claim as the following conditional claim about a fundamental language: Fundamental language (FL) If L is a fundamental ( joint-carving) language, then there are some perfectly joint-carving Fs that we can successfully quantify over in L.

Accordingly, Sider is fallibilist about the details of fundamental ontology, but he does insist that the language of our best science is a likely candidate for a fundamental language – this assertion is evidence of Sider’s naturalism; he is interested in laying out the ontological foundations of scientific inquiry. But a point of interest, then, is what we can say about the perfectly joint-carving Fs. If science is truly a fundamental language, then it should be able to reliably quantify over Fs. We should at the very least have some idea about how this happens. This could be done by pointing out some Fs that carve perfectly at the joints. If no F carves perfectly at the joints, then we have no cases where the debate over the existence of Fs is guaranteed to be substantive, and indeed no clear sense as to what the supposed fundamentality of the existential quantifier amounts to. Fortunately, Sider gives us an example of a perfectly joint-carving candidate term, namely ‘electron’ (or ‘being an electron’). To establish this, Sider ought to show us that ‘electron’ latches on the ‘right sort’ of joint and hence can be

30

Ibid., p. 96.

85

86

Identifying the alternatives: ontological realism, deflationism

shown to capture a perfectly fundamental aspect of reality rather than a mere semantic (i.e., conventional) bias. Sider acknowledges that in ordinary language, for example when quantifying over things such as tables and chairs, we are often not dealing with perfect joint-carving. But he claims that perfectly joint-carving terms are available when dealing with fundamental ontology – perhaps when quantifying over subatomic particles like electrons. He does however admit that there’s a caveat even with examples from fundamental physics: When a term like ‘mass’ is introduced in physics, it’s intended to stand for a fundamental physical magnitude, and so if there’s a joint-carving property in the vicinity then that property is meant by ‘mass’, even if it doesn’t quite fit the physicists’ theory of ‘mass.’31

This comes from a chapter of Sider’s book that deals with reference magnetism. The idea is that highly joint-carving candidate terms like ‘mass’ or

‘electron’ latch on to the joints of reality even if their usage, namely the physicists’ theory of ‘mass’, does not quite reflect the actual joint-carving terms ‘mass’ and ‘electron’. Recall that there can be multiple joints in the vicinity of a perfectly joint-carving candidate, but any candidate sufficiently close to a joint-carving one will do as long as there is a genuine joint-carving property or object in the vicinity. So Sider’s view seems to be that any characteristic feature that we actually use to identify electrons can turn out to be mistaken. All that is required is that there is a joint-carving category that is an approximate fit for actual usage. Let us reproduce Sider’s key example, which focuses on the property of being an electron.32 We have a theoretical term,E, which is intended to carve

at the joints. We also have a single element of structure, e (understood as a feature of reality rather than a concept), which carves at the joints and is associated with E by the relevant disputants (in an ontological or a scientific debate). However,e only needs to satisfy enough of the ‘core theory’ typically associated with E. By ‘core theory’ Sider means the defining characteristics of E. In the case of ‘electron’, Sider lists mass, negative charge, and ‘being a subatomic particle that orbits nuclei’ as part of the core theory. Note that although Sider connects e with meaning, the issue is not conceptual; here e would reflect the feature of reality expressed by the property of being

31

Ibid., p. 32.

32

Ibid., pp. 48–49.

4.4 A case study: Sider’s ontological realism

an electron. Sider is interested in genuine, mind-independent joints in real-

ity, and e is supposed to pick out one such joint, whereas the core theory ensures that E captures a sufficient number of the defining characteristics of e to reliably latch onto it, in accordance with reference magnetism. This model can be applied, for instance, to a debate about whether there are any electrons in a certain space-time region, R. Since enough of the core theory associated with ‘electron’ is satisfied by the property of ‘being an electron’, we seem to have a substantive dispute in Sider’s sense. Sider concludes that ‘To have a substantive dispute about electrons, there must remain enough common ground about what electrons are so that all disputants can be regarded as talking about the same thing’.33 We could attempt to revise the core theory about electrons to produce a more charitable reading of Sider’s example. Plausibly, the core theory involves properties like having half-integer spin, and being governed by the Pauli Exclusion Principle and Fermi–Dirac statistics. 34 But one might still ask: what guarantees that there is a joint-carving category that is even an approximate fit for the actual usage of ‘electron’ if we cannot point to a single feature of electrons that is necessarily included in the core theory? This is not a question that Sider considers, as he is committed to full-blown externalism. For Sider, it does not matterwhich features of electrons are part

of the core theory, as long as all the disputants can be regarded as talking about the same thing. This is an external constraint, rather than one based on the intrinsic features of electrons. But others might worry that unless we have some previous information about what electrons are – that is, which features are necessarily included in the core theory – then we have no means to determine whether the disputants are indeed talking about the same thing.

In what remains of this section, we will discuss certain examples from physics in order to put Sider’s theory in action. Readers not comfortable with this level of scientific detail may safely move on to the next section, although the essential content should be understandable without any background in physics.

33 34

Ibid., p. 49. The Pauli Exclusion Principle states that no two fermions can occupy the same quantum state (and so have the same four quantum numbers) at the same time. Fermi–Dirac statistics apply to (systems of) particles with half integer spin, i.e., fermions.

87

88

Identifying the alternatives: ontological realism, deflationism

The purpose of bringing in some actual science is to demonstrate how difficult it may be, at least sometimes, to determine whether the disputants in a scientific debate are really talking about the same thing. Recall that Sider’s example of a substantive question is whether there are any electrons located in a certain space-time region, R. But suppose R consists of a so-called ‘Bose–Einstein condensate’ (BEC).35 Strangeness ensues: since all the participating particles are in the same quantum state, they cannot be described by Fermi–Dirac statistics (we must use Bose–Einstein statistics instead) and (rather strikingly) they are not subject to the Pauli Exclusion Principle, which we included in the charitable interpretation of the ‘core theory’ of electrons. So we are dealing with particles that have integer instead of half-integer spin; that is, bosons as opposed to fermions. One of the many peculiar aspects of these BECs is that if electrons, which individually are fermions, participate in them, they must come in tightly bound electron pairs, thus producing the required integer spin and behaving like bosons instead of fermions. Interestingly, physicists themselves are debating about whether such tightly bound electron pairs – so-called Cooper pairs36 – should be regarded as bosons or not: [W]hether or not the tightly bound electron pairs can really undergo the BEC directly, is an old but fundamental (and nontrivial) problem, particularly as we know that the bound electron pairs are not exactly bosons.37

We do not have to concern ourselves with all the relevant scientific details, but what we can already take from this is that the question about 35

36

37

A Bose–Einstein condensate is a phase of matter which emerges at extremely low temperatures. In a Bose–Einstein condensate, an unlimited number of bosons can be grouped together in a single (lowest) quantum state. The theoretical possibility of such a condensate was predicted by Bose and Einstein in the 1920s; the first condensate was produced in 1995. This can be achieved for instance by cooling down rubidium atoms. For more details, see the popular article by E. A. Cornell and C. E. Wieman, ‘The Bose-Einstein Condensate,’ Scientific American 278.3 (1998), pp. 40–45. Named after the Bardeen–Cooper–Schrieffer (BCS) theory of superconductivity. In fact, what we are talking about here are strictly speaking fermionic condensates , in which a pair of fermions are bound together, as in a Cooper pair, and then, acting like a boson, can form a Bose–Einstein condensate. Gang Su and Masuo Suzuki, ‘Towards Bose–Einstein Condensation of Electron Pairs: Role of Schwinger Bosons,’International Journal of Modern Physics B 13.8 (1999), p. 926.

4.4 A case study: Sider’s ontological realism

whether there are any electrons located in R is getting rather complicated! But is the question substantial in Sider’s sense? Even this is not clear: whether we should consider a tightly bound electron pair to be a boson might very well be a conventional matter, perhaps based on a merely semantic decision by physicists. Or it might not. The question is whether we are dealing with genuine, joint-carving natural kinds to begin with – such as bosons and fermions – and what constitutes a member of these kinds. If ‘boson’ and ‘fermion’ carve at the joints, then this question would seem to be substantial. However, if the distinction between bosons and fermions is itself conventional – if it doesn’t pick out any genuine joints in reality – then the question about whether Cooper pairs are bosons is similarly a matter of convention. Physicists are unlikely to be very alarmed about any this, since there clearly is determinate fermion-like behaviour and boson-like behaviour, which we can describe statistically in accordance with the above-mentioned Fermi–Dirac and Bose–Einstein statistics. So fermions and bosons can generally be distinguished in scientific contexts. But cases like the Cooper pair instigate some doubts regarding this distinction. Physicists are happy to take advantage of the peculiar behaviour of Cooper pairs (in superconductors, for instance), but it’s at best a curiosity to them whether the fermion–boson distinction carves at the joints in Sider’s sense. At any rate, there may be no clear core theory or common ground that would settle the matter. One interpretation of this situation is that according to Siderian ontological realism (S-OR), we are dealing with a non-substantive question. Let’s put the physics aside for now and return to the philosophical analysis. The case of BECs does not show that the more general question, ‘Are there electrons?’ is non-substantive in Sider’s terms – that would be a worrying result. But given the lack of a determinate core theory about electrons, it does become more difficult to test Sider’s theory. The relevant test is supposed to be that every candidate meaning for ‘electron’ that approximately fits the actual usage of the term, mutatis mutandis, provides the same answer to the question ‘Are there electrons?’ But whether the previous scientific example passes this test is unclear. We might compare this with another popular example that Sider discusses, srcinally from Bennett.38 We may ask: ‘Is that a martini?’ of a particular appletini. This 38

Bennett, ‘Composition, Colocation and Metaontology.’

89

90

Identifying the alternatives: ontological realism, deflationism

question, as Sider as well would admit, is non-substantive. But it could be maintained (as Sider does) that this question is not about the concept of a martini. Presumably, the question ‘Are there martinis?’ is nevertheless substantive, because every candidate meaning for ‘martini’ that approximately fits the actual usage of the term, mutatis mutandis, provides an answer in the affirmative – appletinis notwithstanding. Accordingly, one might think that Bose–Einstein condensates are like appletinis, a non-standard case that physicists may argue about at bars, but not a substantive ontological question. While this analogy is initially appealing, it has a clear problem: ‘martini’ is surely not a candidate for a perfectly joint-carving term in the first place! In the case of martinis, it is of little consequence whether all the disputants can be regarded as talking about the same thing – they may have very different conceptions about what constitutes a martini, even to the extent that they cannot be regarded as talking about the same thing. But this is hardly a concern for S-OR. The situation is different with ‘electron’, which is supposedly a good candidate for a perfectly joint-carving term. The point is that the debate about BECs and Cooper pairs presumably should count as a substantive question, whereas the debate about appletinis should not. If ‘electron’ really is a perfectly joint-carving term, then we should strive to establish a determinate core theory about electrons, part of which will involve their behaviour in any number of odd cases, such as BECs. This is no mean task.

4.5 Taking stock What would a ‘serious’ ontological realist make of all the above? It is unlikely that she would be too discouraged, as ontology was never supposed to be easy. However, the ontological realist (in the broad sense) is now in a rather peculiar position: on the one hand common sense is on the realist’s side, but on the other the burden of proof may nevertheless be on the realist. Why is that? Because despite our best efforts, we do not seem to have much in terms of secure or established knowledge of matters metaphysical. In other words, it is relatively easy for various deflationists, conventionalists, and even anti-realists to point to ontological debates where it is not at all clear that anything of substance is at issue. Moreover, in some cases where it might appear to be possible to present easy arguments that

4.5 Taking stock

might settle ontological debates, there is also the threat that it is too easy (cf. Thomasson’s ‘easy’ approach) to do so – too easy for the serious ontological realist at any rate. Even if it isn’t reasonable to conclude that the burden of proof is now on the ontological realist, it does seem clear that the realist owes us a more fine-grained story about how, exactly, we come to know metaphysical truths. For the most part, metaphysicians are quite happy to pursue the questions they are interested in without an in-depth analysis of the tools they use to do so, but this arguably leaves them too vulnerable to the deflationist’s critique. Indeed, since the rise of metametaphysics, metaphysicians have been much more aware about the pressing need to thoroughly survey the epistemological foundations of the discipline, including its ‘toolbox’. At least partly because of this, certain new areas of research have in recent years gained a substantial boost. Among them are the topics of the next two chapters: grounding, ontological dependence, and fundamentality. As we’ll see, the metaphysician’s toolbox has been supplemented with some very promising tools due to these areas of research. One central issue has been to shift our interest towards what is fundamental rather than what is derivative: if the status of composite objects, for instance, is subject to such an intense debate without there being much hope of settling the Special Composition Question (see Chapter 2), then perhaps we should focus on what it is that grounds the existence of composite objects – what they depend on for their existence. The next two chapters will address such questions. Besides driving efforts to expand the metaphysician’s toolbox, problems emerging from metametaphysics have, to a certain extent, shifted attention towards purely epistemic issues. At least, this would seem to be the case once we have made sure that a given debate is not merely linguistic. In other words, we also need conceptual clarification to ensure that the language of metaphysics is accurate. Assuming that we can do this, research concerning epistemic issues in metaphysics becomes central. Of course, epistemology lives a life of its own, but there has recently been a growing interest in such research – also by metaphysicians themselves and not just epistemologists. It may not be appropriate to speak of ‘the epistemology of metaphysics’ as a subdiscipline of its own quite yet, but the various topics to be discussed later in this book – modal epistemology, the role of intuitions in philosophy, and the relationship

91

92

Identifying the alternatives: ontological realism, deflationism

between science and metaphysics – are linked precisely by the pressing need to establish a more rigorous epistemology of metaphysics. Naturally, these efforts are of interest especially to the ontological realist, for many deflationists and anti-realists would consider the

project of

the epistemology of metaphysics to be hopeless, unnecessary, or simply impossible. But since this is a project that is now receiving ample attention, it appears as if the ontological realists have taken the challenge seriously and are indeed focusing on the (epistemic) foundations of their discipline with an unforeseen enthusiasm. The rest of this book is dedicated to this very project.

5

Grounding and ontological dependence

The notion of ‘ground’ stormed into contemporary analytic metaphysics at the beginning of the twenty-first century, 1 but the roots of the notion go all the way back to Aristotle. At its simplest, grounding may be understood as ‘metaphysical explanation’. To be more precise, when somex is grounded in some y, it is usually thought that y explains x. Moreover, the status of y is generally thought to be somehow prior to that of x – grounding is typically understood to express priority between things. For instance, we might say that the members of a set are prior to the set itself; the existence of the set is grounded in its members. Or to take a more concrete example, the existence of any given composite object is grounded in the existence of its parts. For instance, we might suggest that the existence of any given water molecule is grounded in the existence of hydrogen and oxygen atoms. Somewhat more controversially, we might also say that mental states are grounded in physical states. One thing that each of these explanations has in common is that they arenon-causal. The existence of hydrogen and oxygen atoms, for instance, does not cause the existence of water molecules. Accordingly, grounding is often called ‘metaphysical explanation’ exactly to distinguish it from ‘causal explanation’. In each of the mentioned examples there appears to be some sort of dependence relation between the grounded thing and the grounding thing or things. Quite generally, this would seem tobe ontological dependence, which is the other primary topic of this chapter. In fact, one question that we will 1

The definitive work is Kit Fine, ‘The Question of Realism,’ Philosophers Imprint 1 (2001), pp. 1–30, but for more recent discussion, see especially F. Correia and B. Schnieder (eds.), Metaphysical Grounding: Understanding the Structure of Reality (Cambridge University Press, 2012); see also R. L. Bliss and K. Trogdon, ‘Metaphysical Grounding,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2014 edn); see http://plato. stanford.edu/archives/win2014/entries/grounding/.

93

94

Grounding and ontological dependence

consider is whether grounding is just (a variety of) ontological dependence. The latter comes in a number of different forms and one goal of thischapter 2 is to familiarize the reader with different senses of ontological dependence.

This chapter is the most technical in the book, but formal notation will only be used for clarification: all definitions appear first in natural language. A basic introduction to typical formal features of grounding and ontological dependence will be presented, but the reader is not expected to master all of this material; the main emphasis is on gaining a basic understanding of the notions in order to be able to apply to them to various philosophical problems. We will start from a basic introduction to the notion of ontological dependence and some of its primary uses; this will occupy the first two sections. We will then move on to consider the relationship of grounding and ontological dependence in the third section. A somewhat more technical presentation of the various formal features of ground will follow in the fourth section, although formal notation will once again be avoided. The last two sections of the chapter concern applications of grounding to various other topics as well as related, more established notions, namely causation, reduction, modality, and truthmaking.

5.1 Ontological dependence: a fine-grained notion It is not uncommon to see the notion of ontological dependence used in a rather coarse-grained manner, given that it encompasses a family of relations. For instance, we often see claims such as: (1) ‘Sets ontologically depend on their members’. (2) ‘Electricity ontologically depends on electrons’. (3) ‘God doesn’t ontologically depend on anything’. While all of the above no doubt express some important type of dependence relationship, they are also clearly quite different from each other. In (1),

2

Additional recommended resources on ontological dependence include Fabrice Correia, ‘Ontological Dependence,’ Philosophy Compass 3 (2008), pp. 1013–1032, Tuomas E. Tahko and E. J. Lowe, ‘Ontological Dependence,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2015 edn); see http://plato.stanford.edu/archives/spr2015/entries/ dependence-ontological/; and Kathrin Koslicki, ‘Varieties of Ontological Dependence,’ in Correia and Schnieder (eds.), Metaphysical grounding, pp. 186–213.

5.1 Ontological dependence: a fine-grained notion

we mean that a set {x, y, z} could not exist if its members, namely x, y, z, did not exist. The type of dependence in question rigid is existential dependence, to be clarified in a moment. (In fact, there is another sense of dependence at work in (1)as well, namely identity-dependence, but we will return to this example later on, in the next section.) In (2), we seem to have in mind a more general kind of dependence:there could not be electricity if there were no electrons. So existence of electricity depends on the existence of particles of a specific kind, electrons. This second typeof dependence is also existential, but to separate it from the rigid dependence in (1), we may callgeneric it existential dependence. In (3)we are instead referring to the ontological independence

of God. Presumably, God does not depend for her existence on anything, by her very nature. In other words, it is part of the essence of God that she is ontologically self-sufficient. We might call this essential independence, in contrast to essential dependence. A family of notions is beginning to emerge. However, we should formulate each notion somewhat more precisely. The first thing to note in defining ontological dependence is the modal-existential element in dependence claims. For instance, we’ve said that a set cannot exist unless its members do. So there is a sense in which the existence of a set necessitates the existence of its members. Indeed, it is common to talk, for example, about rigid existential necessitation as synonymous with rigid existential dependence

and generic necessitation as synonymous with generic existential dependence. Typically, statements of ontological dependence are thought to refer to metaphysical modality (rather than, say, conceptual or logical modality). 3 This is primarily because they concern matters that are broader than just conceptual or logical; the ontological independence of God being a case in point. Besides God, substances and perhaps fundamental particles are often considered to be entities that do not depend for their existence upon anything else. We can now give an initial definition of rigid existential dependence:4

3

4

The distinction between different kinds of modality is a matter of some debate, but metaphysical necessity is typically understood as a broader kind of necessity than conceptual or logical necessity, while logical necessity is strictest. In other words, a metaphysically necessary truth would be something that is necessary in virtue of something other than the definitions of concepts and the laws of logic. This and the following non-formal definitions are taken from Tahko and Lowe, ‘Ontological Dependence’.

95

96

Grounding and ontological dependence

Rigid existential dependence (RXD) x depends for its existence upon y =df Necessarily, x exists only if y exists.

Why ‘rigid’? Because there is no flexibility here: the existence of a given x requires the existence of that very y.5 It could not be something a little bit

like y, something falling roughly in the same category, for instance; it must be y. The definiens in (RXD) is equivalent to ‘Necessarily, if x exists, then y exists’, so that according to (RXD) the existential dependence of x upon y amounts to the strict implication of y’s existence by x’s existence. We have mentioned one example of rigid existential dependence, namely that sets ontologically depend on their members (more precisely, a set depends rigidly on the very members it has, i.e., any change in a set’s members will change the set itself). Another, although more controversial, example is a particular person depending for his existence on his parents, or, more precisely, on the particular sperm and egg that he srcinates from. This example is of course related to the essentiality of srcin .6 However, (RXD) clearly fails to capture the intuitive sense of dependence in some cases. Consider a living organism. A living organism would appear to depend for its existence upon its parts, such as cells. But we also know that a living organism may survive a change of any of its cells, provided that the change is not disruptive. Such an organism must certainly have parts such as cells if it is to exist, but which objects those parts are is

inessential – and consequently it is not the case that it depends for its existence, in the sense defined by (RXD), upon any one of those parts. But it is possible to define another sense of existential dependence in which it is true to say that a composite object depends for its existence upon its proper parts; a generic notion of existential dependence, defined as follows:7 Generic existential dependence (GXD) x depends for its existence upon Fs =df Necessarily, x exists only if some Fs exist.

Composite objects are existentially dependent objects in the sense of (GXD), since they require the existence of proper parts (set F as ‘proper

5

6 7

The srcin of the term ‘rigid’ is Saul Kripke, Naming and Necessity (Harvard University Press, 1980). As discussed in Kripke, Naming and Necessity. The labels (RXD) and (GXD) are to separate eXistential dependence from eSsential dependence – see below.

5.1 Ontological dependence: a fine-grained notion

part of x’ in (GXD)). The important difference between the rigid and the generic cases is that (RXD) refers to a specific object whereas (GXD) only requires that at least some Fs exist. Another example, mentioned earlier, where (GXD) would seem to capture the correct sense of dependence is (2), ‘Electricity ontologically depends on electrons’. Let us now turn to a formal presentation of these notions. We use the sentential operator ‘ ☐’ for metaphysical necessity, the one-place predicate ‘ E’ for existence, and the two-place sentential operator ‘

→’

for

material implication. Following this notation, we can formalize rigid existential dependence (RXD) and generic existential dependence (GXD) as follows: (RXD) ☐ (Ex → Ey) (GXD) ☐ (Ex → ∃yFy)

(RXD) can be read as ‘ x rigidly depends for its existence on y’, or alternatively ‘x rigidly necessitates y’. In (GXD), we have added the existential quantifier ‘∃’ as well as the general term F to express the thought that ‘x generically depends for its existence on something being an F’, or alternatively ‘x generically necessitates F’. The important difference between the rigid and the generic cases is that (RXD) refers to a specific object whereas (GXD) only requires that at least some Fs exist. We would now have the tools to formalize (1) and (2), but note that there are cases where further tools are required. Consider: (4) ‘Children ontologically depend on their parents’. On the face of it, what we mean in (4) is that if parents x and y had not existed, then their child z could not have come into existence. This looks like a case of rigid existential dependence, but it is clear that once z has been conceived/born, her father/mother can go out of existence without any effect on her own existence. At that point, there is only past rigid existential dependence. For cases such as this, we would require temporally relativized versions of (RXD) and (GXD), but we will omit these complications here.8 In any case, we now have an initial understanding of the modal-existential analysis of ontological dependence.

8

The specific formalization is a voluntary exercise. See Correia, ‘Ontological Dependence,’ p. 1016 for further details.

97

98

Grounding and ontological dependence

But what about (3)? It does not seem to fit this analysis neatly. Moreover, there are other famous examples where the modal-existential analysis would seem to break down, or not to be sufficiently fine-grained. Consider the relationship between Socrates and the temporally extended event or process that was his life.9 If we attempt to formulate this relationship in terms of (RXD), it will turn out that Socrates’s life depends rigidly for its existence upon Socrates. But surely Socrates’s existence also depends upon his life! At the same time, we may state numerous things that are evidently true of the life of Socrates, yet are not true of him, and vice versa (for example, his life was long, but came to an abrupt ending, while Socrates himself was snub-nosed). The upshot is that Socrates and his life surely cannot be identical. It seems thus that there are certain difficult questions which the modal-existential analysis may not fully address, at least not in the simple form that we have presented it. This warrants us to consider more fine-grained notions of dependence, such as identity-dependence and essential dependence.

5.2 Identity-dependence and essential dependence Until quite recently, it was common to think that ontological dependence can be fully characterized in modal-existential terms, as we have done above. One obvious reason for this is that if one adopts the usual ‘modalist’ analysis of essence, essential dependence will collapse into a form of modal-existential dependence.10 But there is an alternative way to formulate (some varieties of) ontological dependence, which is needed if essence is not analysed in modal terms. 11 One motivation for developing a non-modal conception of ontological dependence is that the modal-existential analysis appears to be too coarse-grained for some cases. We have already mentioned one example, from E. J. Lowe, but the most well-known examples have been made famous by Kit Fine.12 Consider, for instance, what the modal-existential account entails in the case of necessary existents. Take 9 10

11 12

The example srcinates from Tahko and Lowe, ‘Ontological Dependence.’ For a classic account, see Ruth Barcan Marcus, ‘Essentialism in Modal Logic,’ Noûs 1.1 (1967), pp. 91–96. See Kit Fine, ‘Essence and Modality,’ Philosophical Perspectives 8 (1994), pp. 1–16. Kit Fine, ‘Ontological Dependence,’ Proceedings of the Aristotelian Society 95 (1994), pp. 269–290.

5.2 Identity-dependence and essential dependence

Socrates and the number 2, for example. Given that numbers necessarily exist, it is necessarily the case that the number 2 exists if Socrates does. But presumably we do not want to say that Socrates depends upon the number 2, or indeed on most necessary existents that you might put in the place of the number 2. Yet, the modal-existential account makes everything depend upon every necessary existent, which seems like the wrong result. Admittedly, those who defend a modal-existential analysis of ontological dependence could insist that it applies only to contingent objects.13 Simons makes this type of qualification by focusing on concrete entities, hence excluding necessary existents by definition; he also excludes self-dependence. Simons calls the resulting notion of dependence weak rigid dependence, but a stronger notion,strong rigid dependence, is also defined – the

latter is a special case of the former. 14 One example of weak rigid dependence as defined by Simons would be a particular water molecule depending for its existence on a particular oxygen atom. In the case of strong rigid dependence, the dependent object cannot be a proper part of the object it depends upon. So object x is strongly rigidly dependent on object y if x depends for its existence on y and y is not a proper part of x. One example of strong rigid dependence defined thus would be a particularized property (these are often known as ‘tropes’ or ‘modes’) depending for its existence on a substance, e.g., the particular redness of an apple depending for its existence on the apple. In addition to these rigid notions, Simons defines corresponding notions of (weak and strong) generic dependence. But while it is possible to avoid some of the challenges raised for the modal-existential account with these qualifications, they do nevertheless leave room for an alternative account of ontological dependence that could also be applied to necessary existents. Alternatively, a proponent of the modal-existential account could simply bite the bullet and insist that every contingent entity does rigidly depend for its existence on necessary existents. One reason to

do so would be the ability to get by with a sparser battery of ontological tools – a consideration motivated by parsimony. The modal-existential analysis of ontological dependence can thus be developed further and it can perhaps overcome some of the problems that are typically associated with it. But there are areas where a more

13 14

See Peter Simons, Parts: A Study in Ontology (Oxford: Clarendon Press, 1987), p. 295. Ibid., p. 303.

99

100

Grounding and ontological dependence

fine-grained notion would seem to be of use. It should be noted though that much of the contemporary literature in defence of a non-modal, fine-grained analysis operates in a ‘neo-Aristotelian’ framework which typically assumes some ‘non-modalist’ version of essentialism. 15 Accordingly, there is an obvious rift between the modal-existential analysis and the essentialist analysis – one that we cannot fully bridge here. In any case, it is good to keep in mind that the notion of ontological dependence itself does not immediately force one to make a commitment in this regard, even though some of its applications may entail such a commitment. As a point of entry to the idea that there could be notions of dependence not easily analysable in terms of the modal-existential account, consider the fact that not all forms of dependence seem to involve a requirement for existence at all. Indeed, as we already saw in the case of (3), ‘God doesn’t ontologically depend on anything’, it seems that something beyond mere existential independence is being exp ressed. Instead, one might say that God would not be the being that she is if she were not ontologically independent by her very nature. The notion of

essen-

tial dependence , which involves requirements for identity or essence , may

better express what God’s supposed ontological independence is about. In other words, it is an essential property of God that she is ontologically independent. It is not quite straightforward to define essential dependence, although a formal definition will be given below. But before that, we ought to get a better picture of what it means to say that an object depends upon something for its identity ; that is, we should clarify the relation of identity-dependence . Note that the notion of ‘identity’ at play here is not the one symbolized with the ‘equals’ sign, namely ‘=’. Rather, we mean ‘identity’ in the sense of what a thing is, or which thing of a certain kind a thing is. To say that the identity of x depends on the identity of y is to say that which thing of its kind y is metaphysically determines which thing of its

kind x is. For instance, then, the identity of a set is metaphysically determined by the identities of its members. Identity-dependence expresses the determination of the individuality of objects in terms of the individuality of other objects. Thus, the identity-dependence of a set upon its members is a consequence of the fact that the Axiom of Extensionality 15

E.g., Koslicki, ‘Varieties of Ontological Dependence.’

5.2 Identity-dependence and essential dependence

functions as the criterion of identity for sets.

16

We are now in a position

to give a definition of identity-dependence: Identity dependence (ID) x depends for its identity upon y =df There is a two-place predicate ‘F’ such that it is part of the essence of x that x is related by F to y.

We can exemplify (ID) by letting x be {z} and y be z, in which case we have that {z} depends for its identity upon z, because there is a two-place predicate – namely ‘being a member of the singleton set’ (also known as the unit set function) – such that it is part of the essence of { z} that it is the singleton set of z. But, of course, for (ID) to express a usable notion, a full account of the notion of ‘essence’ would be required – and this is something that we will not attempt to provide here. However, various attempts to construct such an account have been made, for instance by Fine and Lowe.17 A central idea behind both their accounts is the Aristotelian notion of real definition. The real definition of an entity in this context could be interpreted simply as a proposition which states the essence of the entity. In this sense, a thing’s essence may be said to constitute its identity, when one uses the word ‘identity’ in this distinctive manner to speak of a thing’s identity, rather than using it to speak of the identity relation. Seen in this light, identity-dependence as defined by (ID) is simply a species of essential dependence, that is, a way in which the essence of a certain thing is determined by a relation in which it stands to another thing. So let us now move to essential dependence. As with existential dependence, we can define a rigid and a generic notion of essential dependence (sometimes called essential existential dependence), as follows. First, rigid essential dependence:18 Rigid essential dependence (RSD) x depends for its existence upon y =df It is part of the essence of x that x exists only if y exists.

Note that, whereas two distinct entities plausible cannot be identitydependent upon each other, it plausibly is possible for each of two distinct

16

17

18

The Axiom of Extensonality is part of the standard Zermelo–Frankel formulation of set theory; it states that any two sets with exactly the same members are equal. Kit Fine, ‘Logic of Essence,’ Journal of Philosophical Logic 24 (1995), pp. 241–273; E. J. Lowe, ‘Two Notions of Being: Entity and Essence,’ Royal Institute of Philosophy Supplement 83.62 (2008), pp. 23–48. These definitions follow Tahko and Lowe, ‘Ontological Dependence.’

101

102

Grounding and ontological dependence

entities to depend essentially for its existence on the other. Moreover, if x depends essentially for its existence upon y, then x also rigidly depends for its existence upon y; that is, essential dependence entails rigid existential dependence ((RSD)

→

(RXD)) – but not, of course, vice versa. Similarly, we

can define generic essential dependence, as follows: Generic essential dependence (GSD) x depends for its existence upon y =df It is part of the essence of x that x exists only if some F exists.

(GSD), unsurprisingly, operates in a similar fashion with regard to (RSD) as (GXD) does to (RXD). We can also express both of these notions formally, using Fine’s notation. First, we need to introduce a new operator ‘☐x’, which behaves like the more familiar ‘☐’, but should be read as ‘it is part of the essence/nature of x that …’.19 With the help of this new operator, we can introduce formal definitions of rigid essential dependence and generic essential dependence, which correspond to (RXD) and (GXD): (RSD) ☐x(Ex → Ey) (GSD) ☐x(Ex →∃yFy)

(RSD) can be read as ‘it is part of the essence of x that it exists only if y does’, whereas the natural reading of (GSD) is ‘it is part of the essence of x that it exists only if something is an F’. It is important to note that while it is uncontroversial that all essential truths about x are necessary truths about x, the converse – the claim that all necessary truths about x are essential truths about x – has been challenged by Fine and Lowe.20 The challenge is important in this connection because if all necessary truths about x were also essential truths about x, then (RSD) and (GSD) would simply reduce to (RXD) and (GXD) (recall that they can be understood in terms of metaphysical necessitation). In other words, essential dependence is of genuine interest only to those who are non-reductionist about essence, that is to those who adopt a non-modal conception of essence as defended by Fine and Lowe. We can find slightly different formulations of essential dependence in the literature. For instance, on Kathrin Koslicki’s re-construal of Fine’s essentialist account, we get: 21 19 20

21

Fine, ‘Logic of Essence.’ See Fine, ‘Essence and Modality’; E. J. Lowe, ‘What is the Source of our Knowledge of Modal Truths?’,Mind 121 (2012), pp. 919–950. Koslicki, ‘Varieties of Ontological Dependence,’ p. 190.

5.2 Identity-dependence and essential dependence

Constitutive essential dependence (EDC) An entity x ontologically depends on an entity (or entities), y, only in the case that y is a constituent (or are constituents) in x’s essence.

This account relies on the notion of constitutive essence, which is developed in Fine’s work; as Fine puts it: ‘we may take x to depend upon y if y is a constituent of a proposition that is true in virtue of the identity of x or, alternatively, if y is a constituent of an essential property of x’.22 It should be noted that even though the notions of essential dependence defined above are no doubt more fine-grained than the modal-existential notions, there may be reasons to think that even they are not sufficiently fine-grained for all purposes (as argued by Koslicki). To illustrate, consider Fine’s well-known discussion of Socrates and the singleton set that has Socrates as its sole member. To use Fine’s terminology, we could say that it is part of the constitutive essence of Socrates’s singleton set that it has Socrates as its sole member, whereas it is not part of the constitutive essence of Socrates to be the sole member of Socrates’s singleton set. But as Koslicki points out, this is really just to say that Socrates’s singleton set ontologically depends on Socrates whereas Socrates does not ontologically depend on Socrates’s singleton set.23 The relevant notion of dependence appears to be built into Fine’s notion of essence and, similarly, the notion of essence assumed by Fine (and many other ‘neo-Aristotelians’) is already built into (EDC). Before we conclude our discussion of essential dependence, a more general issue regarding the essentialist vs. modal-existential account should be mentioned. It appears that those who are willing to buy into a suitably fine-grained notion of essence will find the modal-existential account of ontological dependence far too coarse-grained. Yet, those who are not sympathetic to the relevant primitive notion of essence and would rather analyse essence in terms of modality would insist that the modal-existential analysis is quite sufficient, and indeed that essential dependence collapses into modal-existential dependence. A further issue is that there is some disagreement amongst those who think that a more fine-grained analysis than the modal-existential account is needed about how the relevant notion of

22 23

Fine, ‘Ontological Dependence,’ p. 275. Koslicki, ‘Varieties of Ontological Dependence,’ p. 195.

103

104

Grounding and ontological dependence

essence is to be constrained. For instance, Koslicki regards Fine’s propositional notion of essence according to which there is little or no distinction

between essence and real definition as overly restrictive. Koslicki identifies the source of this restrictive conception of essence, which is also present in Lowe’s work, to be the focus on essences as individuating – this is the type of view regarding essential dependence, namely identity-dependence, which we have been focusing on here. In Koslicki’s alternative picture, essences ‘must do more than individuate the entities whose essences they are; and real definitions must do more than state conditions which uniquely identify and delineate the entities under consideration at every time and in every world in which they exist’.24 However, this is not the place to settle the controversy regarding the essentialist vs. the modal-existential account; we will return to the relationship between essence and modality in Chapter 7.

5.3 Is grounding ontological dependence? We have already made use of the notion of ‘ground’ and it seems clear that it has something to do with ontological dependence. But is grounding just a variety of ontological dependence? At the outset, we can assume that if grounding were to be understood as a type of ontological dependence, it would be some sort of explanatory dependence. The idea that whatever does the grounding also somehowexplains what is being grounded is a crucial part of the notion’s appeal. Relations of ontological dependence often seem to have a similar type of explanatory role, but the link toexplanation is weaker:even though the existence of water depends on the existence of hydrogen and oxygen, it does not seem to be the case that the existence of hydrogen and oxygen explain the existence of water. Rather, what explains the existence of water is the ability of hydrogen and oxygen atoms to form molecules (even though this is rather simplified). So itseems that not all relations of ontological dependence can be grounding relations in the usual sense. We need something stricter than just ‘an explanatory role’ to identify grounding – otherwise we would end up with a much too liberal notion, for we may regard a number of loosely connected things explanatory in some very loose sense. For instance, we might say that the fact that Smith 24

Ibid., p. 200, fn 13.

5.3 Is grounding ontological dependence?

murdered Jones is explained by certain events in Smith’s childhood, but a more direct explanation might be Smith’s desire to rob Jones. Certain events in Smith’s childhood may help us understand why Smith has murderous desires, but it’s not clear that they serve to ground the fact that Smith murdered Jones. One suggestion that may help to make grounding more precise would be to focus on priority. At the beginning of this chapter, the example was given that if the members of a set are prior to the set itself, then the existence of the set is grounded in its members. So the grounding entities are prior to – or more fundamental than – the grounded entities. This is another sense in which grounding would seem to detach from certain types of ontological dependence, as a purely modal understanding of dependence (without any claim to priority) is also possible. This is of course the sense which we defined above in terms of necessitation: even if x rigidly necessitates y, this does not entail that y must be ontologically prior to x. At any rate, it would be odd to say that parents are ontologically prior to their children! An essentialist notion of dependence, as familiar from Fine and Lowe, seems a more promising tool if we hope to express the relevant sense of priority. 25 But we do not need to decide the matter here, as pluralism is a perfectly viable option as well: there is certainly a use for all of these different notions of dependence. Nevertheless, there may be an independent motivation for finding a systematic link between ontological dependence and grounding, as defining one in terms of the other would be more parsimonious than having two primitive notions.26 In any case, even if we could define one notion in terms of the other, it seems that there are aspects of ontological dependence that are not captured by all accounts of grounding as well as aspects of grounding that are not captured by all accounts of ontological dependence. Accordingly, for the time being it is advisable to keep the notions apart, especially since there are some further, formal differences that need to be taken into account. To these we now turn.

25

See Koslicki, ‘Varieties of Ontological Dependence,’ for a thorough discussion of the essentialist notion of dependence. 26See Fabrice Correia and Benjamin Schnieder, ‘Grounding: An Opinionated Introduction,’ in Correia and Schnieder (eds.), Metaphysical Grounding, pp. 1–36.

105

106

Grounding and ontological dependence

5.4 Formal features of ground A more precise treatment of ground requires taking into account a number of general specifications as well as certain formal features. The general specifications fall into four groups: the first concerns the distinction between full and partial grounds, the second concerns the distinction between mediate and immediate grounds, the third concerns the distinction between weak and strict grounds, and the fourth concerns the so-called ‘operational’ vs. ‘relational’ approaches to grounding.27 The full/partial ground distinction is fairly straightforward: if P and Q together fully ground the conjunction P & Q, then P and Q are each a partial ground for the conjunction P & Q. One reason to adopt this distinction is

that it is often difficult to determine what the full ground of something is. For instance, we might say (to simplify) that the phenomenon of turbulence is partially grounded in the rapid variation of pressure and velocity in a region, but the complexity of the phenomenon and our limited understanding of it make it difficult to determine what would constitute the full ground of turbulence. In fact, partly for the same reasons, turbulence is typically treated statistically. There is nothing particularly controversial about the full/partial ground distinction, except perhaps the question of whether one of them is prior to the other. Fine is of the opinion that full ground is certainly prior, and partial ground should be defined in terms of full ground.28 The simple reason that Fine states for this preference is that the conjunction and disjunction of P and Q both have the same partial grounds, namely P and Q, but in the case of the disjunction, P

∨

Q, both P

and Q are also full grounds. To appreciate this difference, we must make reference to full grounds. In general, when we talk about grounding, we mean full grounding, but as we will see below, sometimes it is important to specify that we mean only partial grounding (and not only for epistemic reasons, as in the case of turbulence). 27

28

For discussion of the first three distinctions as well as a number of other distinctions, see Kit Fine, ‘Guide to Ground,’ in Correia and Schnieder (eds.), Metaphysical Grounding, pp. 37–80. We will not cover all of the relevant distinctions here. For discussion of the operational/relational approaches (the latter is sometimes also called ‘predicational’), see Kelly Trogon, ‘An Introduction to Grounding,’ in M. Hoeltje, B. Schnieder, and A. Steinberg (eds.),Varieties of Dependence (Munich: Philosophia Verlag, 2013), pp. 97–122. Fine, ‘Guide to Ground,’ p. 50.

5.4 Formal features of ground

The distinction between mediate and immediate grounds is also straightforward. Immediate grounds are such that they do not need to be ‘mediated’ by any other grounding relations. For instance, the immediate grounds of the conjunction P & Q are simply P and Q. But if we were to add another conjunct, such as in (P & Q) & R, we would say that the entire conjunction is only mediately grounded in P, Q, and R, since we first have to consider the immediate ground of (P & Q) and the immediate ground of R before moving on to the ground of (P & Q) & R. In this way, the full ground of the more complex conjunction is mediated through other grounding relations. Mediate ground can be defined in terms of immediate ground, as all the mediate grounds can be arrived at by analysing immediate grounds, like we just did in the case of (P & Q) & R. Fine finds this distinction important, because it gives us a natural way to obtain a ground-theoretic hierarchy. The immediate ground of a given truth always follows immediately at the next (lower) level of the hierarchy, and via the chaining of mediate grounds, we can arrive at all the partial grounds for that truth at lower levels of the hierarchy. Considerations such as this will prove to be crucial for certain interesting applications of grounding. The distinction between weak and strict ground is somewhat more complicated – and there is some disagreement about it – but the idea can be summarized as follows: a strict ground for a given fact always occurs at a lower level in the explanatory hierarchy than the fact itself, whereas a weak ground occurs at the same level.29 This means that some facts could even be weak grounds for themselves! However, it is somewhat more difficult to find illustrative examples of weak grounding. One potential example of weak grounding might be Jack’s being Jill’s sibling, which explains Jill’s being Jack’s sibling, and also the fact that Jack and Jill are a pair of siblings. Here Jack’s being Jill’s sibling and Jill’s being Jack’s sibling occur at the same level of explanatory hierarchy. We do not need to dwell on this case, though, as it turns out that the notion of strict ground is generally of much more interest to us. In fact, given all these distinctions, we may identify the focus of much of the grounding literature: typically when we talk about grounding we have in mind a strict partial ordering.30 This bit of 29

30

For an analysis of weak ground, see Louis deRosset, ‘What is Weak Ground?’, Essays in Philosophy 14.1 (2013), Article 2. This has been labelled the ‘orthodox’ approach by Michael Raven, ‘Is Ground a Strict Partial Order?’, American Philosophical Quarterly 50.2 (2013), pp. 191–199.

107

108

Grounding and ontological dependence

set-theoretic jargon captures three important formal features of ground – irreflexivity, transitivity, and asymmetry– which we will discuss in more detail below. Grounding understood as a strict partial order produces the type of hierarchical chain of explanation that we have already been referring to implicitly in this book. Let us now move to a slightly different issue, namely the distinction between operational and relational approaches to grounding. Some proponents of grounding take it to be a relation, whereas some prefer the operational approach. The relational approach is perhaps easier to understand as it considers grounding expressions to be simply predicates that we use to state certain relations that obtain between entities. On this view, an important question is what kind of entities the relata are, that is, an account of the grounded and the grounding entities is required. Candidates include, for instance, facts and states of affairs– in this chapter we primarily refer to facts as the relata. The operational approach, which is preferred for instance by Fabrice Correia and Kit Fine, is ontologically neutral concerning the relata.31 As Correia puts it, ‘it should be possible to make claims of grounding and fail to believe in facts’.32 On this view, the notion of ground is to be expressed by means of a sentential operator, ‘because’ being a natural candidate. We often say that p‘ because q’, and the idea of the operational approach is that grounding is expressed by the ‘because’ operator in such cases. One motivation for this view is that the grammatical form of many grounding claims seems to point towards an operational approach. When I say, ‘I took an umbrella because it was raining’, we see thatp and q can be taken as sentences, and the ‘because’ operator takes these sentences to make a further sentence.33 There are no obvious constraints as to the ontological category denoted byp and q in these sentences, which is why the operational approach can remain neutral about the existence of facts or other typical candidates for the relata in grounding statements. The ontological neutrality of the operational approach may be considered an advantage, but it also has its costs:the approach loses a natural point 31

32 33

For further details, see Fabrice Correia, ‘Grounding and Truth-Functions,’ Logique & Analyse 211 (2010), pp. 251–279. Ibid., p. 254. The example ‘I took an umbrella because it was raining’ may look like a case of causation rather than (or in addition to) grounding. This may indeed be the case. We will return to the relationship between grounding and causation below.

5.4 Formal features of ground

of connection between grounding and ontological dependence, as it does not seem to account for the idea ofexplanatory dependence. If explanation is thought to be a dependence relation (between facts), then grounding is a natural choice for describing this relation. This would perhaps strengthen the case towards the claim that grounding is a specific variety of ontological dependence, namely explanatory dependence between facts. However, it should be noted that the distinction between the operational and relational approaches should be understood as concerning the mostbasic notion of ground, for it is possible to ‘translate’ statements of ground formulated in 34 terms of one approach to the other, given certain assumptions.

As mentioned above, grounding is commonly considered to be constrained by three formal features: asymmetry, irreflexivity, and transitivity. These features are widely but not universally agreed upon. 35 The relevant definitions are as follows, although note that the first two define positive features rather than the negative ones which are taken to hold for grounding. Symmetry A relation R is symmetric if and only if: if x is related by R to y, then y is related by R to x. E.g., ‘being siblings’. Reflexivity A relation R is reflexive if and only if: R is self-relating, i.e., everything bears R to itself. E.g., ‘being self-identical’. Transitivity A relation R is transitive if and only if: if x is related by R to y, and y is related by R to z, then x is related by R to z. E.g., ‘being taller than’.

Since grounding is understood to be a priority relation, it seems clear that it cannot be symmetric or reflexive. Ifx were to ground y and y to ground x, then we would clearly have a loop– a vicious one if we are hoping to explain one thing in terms of another. Similarly, ifx were to ground itself, then it

34 35

For further details, see Correia, ‘Grounding and Truth-Functions.’ For some challenges, see C. S. Jenkins, ‘Is Metaphysical Dependence Irreflexive?’, The Monist 94.2 (2011), pp. 267–276; Jonathan Schaffer, ‘Grounding, Transitivity, and Contrastivity,’ in Correia and Schnieder (eds.), Metaphysical Grounding, pp. 122–38; Tuomas E. Tahko, ‘Truth-Grounding and Transitivity,’ Thought: A Journal of Philosophy 2.4 (2013), pp. 332–340.

109

110

Grounding and ontological dependence

would appear to be ontologically self-sufficient. Perhaps there are such ontologically self-sufficient entities, but we do not typically talk about them as if they were self-grounding, but rather as being fundamental:that is, primitive in the sense of not needing to be grounded at all. Accordingly, grounding as it is usually understood is asymmetric and irreflexive. But what about transitivity? To give an example in terms of grounding, consider sets once again. We might say that the existence of a set is grounded in the existence of its members, and if its members include other sets, its existence is grounded in the existence of their members. In this way, a natural chaining of ground emerges, all the way back to the fundamental, ungrounded entities. In the next chapter we will consider the implications of such grounding chains in more detail, but the possibility of this type of chaining, and hence transitivity, is typically considered central for grounding. Note that these formal features apply only to a subset of the distinctions that were introduced above. For instance, because immediate grounds are always direct grounds and mediate grounds indirect, it follows that only the notion of mediate ground is transitive. Also, since a given fact may be a weak ground for itself, only the notion of strict ground is irreflexive. Accordingly, only grounding understood as a strict partial ordering ‘automatically’ satisfies transitivity, irreflexivity, and asymmetry. These formal features are not only relevant for grounding, of course. They also provide us with a further method for distinguishing between different types of ontological dependence. For instance, rigid existential dependence is reflexive and (therefore) not asymmetric. But the type of ‘non-self-sufficiency’ that we are interested in when it comes to grounded entities arguably requires irreflexivity and asymmetry. However, we can also define an asymmetric relation of rigid existential dependence, which holds when x rigidly necessitates y but not vice versa. This type of one-way rigid necessitation is certainly of some interest, but it may not express the

type of non-self-sufficiency that we have in mind in the case of grounding. To see this, consider the following. If Socrates is a contingent existent and the empty set is a necessary existent, then Socrates one-way rigidly necessitates the empty set. But surely, the existence of Socrates is not derivative of the existence of the empty set!36

36

On this, see Fine, ‘Ontological Dependence.’

5.4 Formal features of ground

Even though asymmetry, irreflexivity, and transitivity are widely agreed to be among the core formal features of ground, they have not gone unchallenged. Considering some of these challenges may help us to understand grounding better, so it is worthwhile to think about them. Here we will focus on transitivity. Consider the proposition ‘There exists a particular bottle of beer, b’.37 We might state the following grounding claims regarding this proposition: (i)

The fact that a particular bottle of beer, b, exists is partially grounded in the fact that b has a stable macrophysical structure.

(ii) The fact that b has a stable macrophysical structure is partially grounded in the fact that certain fundamental laws of physics hold. Both of these grounding claims may be considered plausible. Regarding (i), we wouldn’t think that the bottle of beer exists if it didn’t have roughly the sort of properties – all of them macrophysical – that we expect from a bottle of beer, such as being rigid (assuming that it is a glass bottle), and containing a tasty, usually alcoholic and thirst-quenching beverage. All of the macrophysical properties of b are grounded in certain microphysical properties, the causal features of the constituent microphysical particles in particular, which together produce a stable macrophysical structure. This is not to say that all necessary conditions are grounds, but (i) seems as plausible a candidate as any. As to (ii), the grounding claim could refer to a number of physical laws or principles. We do not need to go into the details here – the stability of macrophysical objects clearly requires microphysical stability and (ii) may be considered to refer to whatever physical laws the stability of matter requires.38 So, there are good reasons to think that certain fundamental physical laws at least partially ground many claims concerning macrophysical stability. Now, by transitivity, weget: (iii) The fact thata particular bottle of beer , b, exists is partially grounded in the fact that certain fundamental laws of physicshold. 37 38

See Tahko, ‘Truth-Grounding and Transitivity’ for further discussion of this example. For instance, we could refer to the Pauli Exclusion Principle, which states that two fermions in a closed system cannot be in the same quantum state at the same time. This principle is sometimes said to be responsible for the space-occupying behaviour of matter, as it ‘prevents’ atoms from collapsing. We encountered this principle in passing in Chapter 4 and will refer to it again in Chapter 6.

111

112

Grounding and ontological dependence

It seems that (iii) states a part of the ‘ultimate ground’ of the existence of b. This is perfectly in accordance with the usual treatment of grounding – we only need the transitivity of grounding to get this result. But at the same time one might think that (iii) is a rather odd grounding claim. The reason for this is that we do not immediately see how fundamental laws of physics are relevant for the existence of a particular bottle of beer. The clash of intuitions at hand here– between a certain type of intuitive relevance and ontological ground– may highlight a deeper issue concerning ground. The issue is that of epistemic as opposed to ontological relevance. The notion of ground is generally considered to be ontological, in which case the clash of intuitions described here should not cause much concern, at least insofar as the opposition to claims like (iii) is based only on considerations of epistemic relevance. So if one asks whether there’s any beer left and someone answers that there is a bottle of beer in the fridge, one would not consider it relevant that the macrophysical stability of the bottle depends, in some sense, on certain fundamental laws of physics. Considerations such as these may prompt us to question whether ground is truly a single ontological relation or are there instead several distinct notions of ground, perhaps some of them ontological and some epistemic. If this were the case, then it might 39 seem that grounding is unable to do all the work it is often claimed to do.

We turn to an overview of some this work, but we will briefly return to the issue concerning distinct notions of ground towards the end of the chapter.

5.5 Grounding, causation, reduction, and modality The title of this subsection suggests that a lot of ground will be covered. This is true, but we will not need to spend too much time on any given issue.40 The reason is partly because we are about to move to a rather 39

40

Besides Tahko, ‘Truth-Grounding and Transitivity,’ see for instance Jessica Wilson, ‘No Work for a Theory of Grounding,’ Inquiry 57.5–6 (2014), pp. 1–45 for discussion regarding the work that grounding is supposed to do in different contexts. For further discussion regarding the formal features of ground, see Gonzalo Rodriguez-Pereyra, ‘Grounding is Not a Strict Order,’ Journal of the American Philosophical Association (forthcoming). For a more in-depth discussion of many of the topics of this subsection, see for instance Gideon Rosen, ‘Metaphysical Dependence: Grounding and Reduction,’ in B. Hale and A. Hoffman (eds.), Modality: Metaphysics, Logic, and Epistemology (Oxford University Press, 2010), pp. 109–135.

5.5 Grounding, causation, reduction, and modality

controversial area, so there are no clear standard views to be presented. For instance, take the relationship between grounding and causation. We have been avoiding this issue so far, but it is quite typical to describe grounding as non-causal explanation. Similarly, although many of the examples might tempt one to think that the grounded entity reduces to whatever it is that grounds it, this is not how grounding is typically understood. The same goes for modality: grounding does not seem to be just modal supervenience, one thing necessitating another. The following illustrative example from Fine concerning the relationship of the physical and the mental serves to summarize all these aspects of grounding:41 It will not do, for example, to say that the physical iscausally determinative of the mental, since that leaves open the possibility that the mental has a distinct reality over and above that of the physical. Nor will it do to require that there should be an analytic definition of the mental in terms of the physical, since that imposes far too great a burden on the anti-realist. Nor is it enough to require that the mental should modally supervene on the physical, since that still leaves open the possibility that the physical is itself ultimately to be understood in terms of the mental.

Fine’s suggestion, unsurprisingly, is that the right question about the relationship of the physical and the mental is whether the latter is grounded in the former. Let us now look at each of the mentioned aspects of grounding in some more detail. As Fine suggests, even if physical states causally determine mental states, this does not mean that mental states could not have a separate existence over and above the physical. However, this does make an assumption about what physicalism involves – one that not everyone would necessarily accept. The assumption is that physicalism rules out the possibility that the mental is separate from but depends on the physical. 42 To establish physicalism in this sense, a stronger connection between physical and mental states should be demonstrated, such as grounding. From this, we can infer that grounding is, in some sense, stronger than causation – although

41 42

Fine, ‘Guide to Ground,’ p. 41. For relevant background concerning physicalism in philosophy of mind, see for instance Daniel Stoljar, ‘Physicalism,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2015 edn); see http://plato.stanford.edu/archives/spr2015/entries/ physicalism/.

113

114

Grounding and ontological dependence

it should be mentioned that this is not universally accepted. Jonathan Schaffer, for instance, has argued that grounding is ‘something like metaphysical causation’.43 If this were correct, it would have certain important ramifications: Schaffer himself argues that just as there are counterexamples to the transitivity of causation, there are counterexamples to the transitivity of grounding as well.44 However, one major challenge for the position that grounding is just a type of (metaphysical) causation is that many of the examples we have considered would appear to involve a type of explanation that does not seem to be purely causal. Another challenge is that grounding seems to be possible between entities of a purely abstract nature or between abstract and concrete entities, such as sets and their members, whereas causation is, on the face of it, a relation that concerns only physical entities. While some attempts are currently in progress to characterize grounding as a variety of causation, for the time being this remains a marginal view. What about reduction: is there a connection to grounding? If we ask whether the mental reduces to the physical, how does that compare to the grounding question? There are some good reasons to think that these are not one and the same question.45 However, one problem in settling this issue is that reduction, just like grounding, allows for a variety of different interpretations. This can be naturally illustrated with the example at hand, namely the relationship between the mental and the physical.proponent A of strict reductive physicalism might (but does not have to) subscribe to the view that mental and physicalpredicates or concepts are interchangeable in such a way that the mental predicate applies if and only if the physical predicate applies. This type of strict reductive physicalism may not be very popular in contemporary philosophy of mind, but if reductionism is understood in these terms, then it is clearly distinct from grounding. Another alternative is to conceive of reduction asidentity.46 In that case we would automatically lose irreflexivity 43 44

45

46

Schaffer, ‘Grounding, Transitivity, and Contrastivity,’p. 122. But for a defence of the ‘orthodoxy’ and against the possibility of counterexamples to the transitivity and irreflexivity of grounding, see Raven, ‘Is Ground a Strict Partial Order?’ For discussion, see Rosen, ‘Metaphysical Dependence: Grounding and Reduction,’ especially pp. 122 ff. For a defence of the view that reduction should be conceived as identity, see Paul Audi, ‘A Clarification and Defense of the Notion of Grounding,’ in Correia and Schnieder (eds.), Metaphysical Grounding, pp. 101–121.

5.5 Grounding, causation, reduction, and modality

and asymmetry, since identity is reflexive, symmetric, and transitive. So, it seems that grounding would again turn out to be a different relation. However, we do not necessarily need to interpret reductionism in terms of identity. There are at least two further alternatives that have been entertained in the grounding literature: reduction characterized in terms of fundamentality and reduction characterized in terms ofessence.47 Roughly put, the first view suggests that when two facts explain the same thing, but one

does so in more fundamental terms, then the less fundamental explanation reduces to the more fundamental. The notion of fundamentality will be examined in detail in thenext chapter, but for now it will be sufficient to note that the more fundamental terms would be thosethat are further in the grounding chain, that is, towards the beginning of the hierarchy. The most fundamental terms would then be at the very bottom of the hierarchy. The second, essence-based characterization suggests that if we have an account of the essence of, say, water in terms of its microstructure, 2HO, then we might propose that being waterreduces to being H2O. This view (or at least this example) certainly runs into some problems, some of which we will consider in subsequent chapters, but on this conception of reduction, the link to grounding could perhaps be maintained:if water is reduced to its microstructure, then we may explain what it is to be water in terms of 2HO. This results in what has been called agrounding–reduction link: If q reduces to p, then p grounds q.48 In sum, our judgement about the grounding–reduction

link will depend heavily on how we prefer to understand reduction. In this connection, we may anticipate the discussion in Chapter 7 regarding a proposed link between metaphysical modality and essence, suggested by Fine, Lowe, and others: they argue that all metaphysical necessities are grounded in essence, even though a given metaphysically necessary truth may not be an essential truth about any one thing. One way to understand this proposal is in terms of the grounding–reduction link, namely that metaphysical modality reduces to essence. But this would certainly require understanding reduction in terms not of identity, but of something else, as some metaphysically necessary truths may require several distinct essences to ground them.

47 48

For an overview ofthese options, see Trogdon, An ‘ Introduction to Grounding.’ This type of characterization can be found in Rosen, ‘Metaphysical Dependence: Grounding and Reduction.’

115

116

Grounding and ontological dependence

Finally, we should briefly examine the relationship between grounding and modality. The two options here are that grounding is either necessary or contingent.49 It is perhaps more popular to think about grounding as necessary: if x (fully) grounds y, then x modally entails y. The modal content would likely be interpreted as metaphysical modality (since grounding is understood as metaphysical explanation), that is, if x (fully) grounds y, then x metaphysically necessitates y. Or, in other words, grounding holds across metaphysically possible worlds. The connection to ontological dependence is obvious here, since we saw that at least certain types of dependence can be characterized in terms of metaphysical necessitation. If grounding were contingent, then it seems that we would lose this type of link to ontological dependence. Most philosophers working on grounding assume that grounding is necessary, but systematic treatments of the issue are still scarce. We do not need to engage with all the details here. Instead, we might consider an example, building on the passage quoted from Fine. Fine claims that to understand physicalism, it is not enough to require that the mental modally supervenes on the physical. This, he continues, leaves open the possibility that the physical could still be somehow understood in terms of the mental. Such a result may seem peculiar, but it is perfectly consistent with the modal picture being proposed: the physical could (metaphysically) necessitate the mental, but this still leaves open what grounds the physical itself, and it is possible that it is the mental that somehow helps to explain the physical. So it appears that metaphysical necessitation is not sufficient for grounding, but if grounding is necessary, then grounding of course entails necessitation. Hence, if the physical grounds the mental, then the physical also metaphysically necessitates the mental.

5.6 Grounding and truthmaking This final section of the chapter deals with an important issue, the relationship between grounding and truthmaking. The truthmaking literature is somewhat older than the modern grounding literature, but the issue is complicated by the fact that some of the early grounding literature was 49

For a much more comprehensive discussion of this issue, see Kelly Trogon, ‘Grounding: Necessary or Contingent?’, Pacific Philosophical Quarterly 94 (2013), pp. 465–485.

5.6 Grounding and truthmaking

also truthmaking literature.50 Truthmaking is often characterized roughly

as follows. Some entity x makes it true that p, where x is a truthmaker and p is a truthbearer, typically a proposition. The relation between a truthmaker and a truthbearer is thought to be the truthmaker relation. The literature on truthmaking is vast and there is not all that much agreement, for instance, about what kind of things could be truthmakers (they could be, e.g., ‘states of affairs’ or just ‘entities’). What interests us are the attempts to clarify the truthmaking relation, sometimes said to be the in virtue of relation: we say that p is true in virtue of x. An influential attempt to clarify this ‘in virtue of’ relation is in terms of truth-grounding: some entity x grounds the truth of some proposition p if p is true in virtue of the existence of x. So, the entities which ground the truth of p are p’s truthmakers. The idea here is that the ‘in virtue of’ relation could be defined as the converse of grounding: x grounds p if and only if p is true in virtue of x. If this

is correct, then grounding would seem to have an important role to play in truthmaker theory. To be more precise, on an understanding of grounding as a relation between facts, we may define truth-grounding as follows: Truth-grounding If the fact that x exists helps ground the fact that p is true, then x helps ground the truth of p, i.e., x is a (partial) truth-ground for p.

However, it is highly controversial whether this type of connection between truthmaking and grounding in fact holds. One reason for the controversy is that truth-grounding violates an important feature of traditional truthmaking theory, namely that truthmaking is generally understood to be a cross-categorical relation between an entity or entities and a truth, not merely a relation between facts, as wehave taken grounding to be.51 Friends of truth-grounding could of course insist that it is a mistake to conceive of truthmaking as cross-categorical and argue in favour of a unified account of explanation in terms of grounding, but it is unclear whether it is possible 50

51

In particular, see Gonzalo Rodriguez-Pereyra, ‘Why Truthmakers?’, in H. Beebeeand J. Dodd (eds.), Truthmakers: The Contemporary Debate (Oxford University Press, 2005), pp. 17–31, and Benjamin Schnieder, ‘Truth-Making Without Truth-Makers,’ Synthese 152.1 (2006), pp. 21–46. However, the contemporary truthmaking literature dates from as early as Kevin Mulligan, Peter Simons, and Barry Smith, ‘Truth-Makers,’ Philosophy and Phenomenological Research 44.3 (1984), pp. 287–321. On this, see David M. Armstrong, Truth and Truthmakers (Cambridge University Press, 2004), p. 5.

117

118

Grounding and ontological dependence

to maintain enough of the srcinal motivation for truthmaker theory if this move is made.52 Many grounding theorists are in fact sceptical about the link between grounding and truthmaking. Fine, for instance, is very much opposed to such a link.53 At best, Fine thinks, we could point to a necessary connection of the following type: if x grounds p, then the fact x will be a truthmaker for p. But such a link would not be very informative, for it might not hold to the other direction. Even if we discover that some propositionp is made true by some fact x, it might turn out thatx is not a sufficient groundfor p. The shaky link between truthmaking and grounding serves to highlight one final and important question about grounding, which we have not directly addressed yet: is grounding univocal? In other words, are there several different types of grounding or are all genuine examples of grounding analysable in terms of one common notion of ground? Univocalism about grounding is commonly assumed without argument.54 To be precise, univocalism as we are here understanding it suggests that ‘grounding’ should be considered in terms of a single notion of dependence, which may or may not be analysable in terms of other notions, depending on the type of univocalism. We have been treating grounding as univocal so far, but the case of truthmaking in particular might drive some towards distinguishing between different types of grounding. This view comes with certain risks, however, as there are those who would regard non-univocalism to be evidence of the incoherence of the whole notion of ground and hence driving us towards scepticism about grounding.55 However, given that there is disagreement even about some of the formal features of ground, it could be that we can arrive at a better sense of grounding if we consider it a family of relations (if it is a relation!), like ontological dependence. That way we could define different subspecies of grounding even with distinct formal features, while they would all fall under the general notion of ground. Such an approach has its appeal, but the question then becomes how to

52 53 54

55

Again, see Tahko, ‘Truth-Grounding and Transitivity.’ Fine, ‘Guide to Ground,’ pp. 43–46. For instance, Trogdon assumes univocalism in his ‘An Introduction to Grounding,’ suggesting that it i s a reasonable starting point. Trogdon also associates this type of univocalism with Rosen. For a sceptical response, see Chris Daly, ‘Scepticism About Grounding,’ in Correia and Schnieder (eds.), Metaphysical Grounding, pp. 81–100.

5.6 Grounding and truthmaking

specify the general notion of ground. If any given example of grounding is an example of some subspecies of the general notion, but the different examples vary to a large extent, then it is not clear why we should consider them to be subspecies of a general notion of ground rather than distinct notions altogether. Indeed, if this route is taken, then the general notion of ontological dependence might serve us better. But we should not end with such a sceptical note. Grounding has quickly become a central area of research in analytic metaphysics and it is no doubt one of the most intriguing themes in metametaphysics. At its best, it promises to give us a thorough understanding of metaphysical explanation. Even failing that, the numerous applications of grounding certainly claim to do a lot of metaphysical work. As we have already seen in passing, one the most important applications both of grounding and ontological dependence concerns fundamentality and the hierarchical structure of reality. This is the topic that we turn to in the next chapter.

119

6

Fundamentality and levels of reality

This chapter concerns the view that reality comes with a hierarchical structure of ‘levels’. This view has a long history and it remains very popular. Our everyday experiences as well as scientific practice seem, prima facie, to strongly support such a view, since scale is a major factor in both of them. The existence of a scale suggests that there is some way to structure reality and this idea is naturally combined with the thought that this structure can be described in terms of ontological dependence – a tool which is now at our disposal. The relevance of scale becomes apparent when we consider parts and wholes, which are studied in mereology: we talk about subatomic particles constituting atoms, atoms constituting molecules (and other superatomic structures), and molecules constituting everything we see around us. To express this in terms of ontological dependence we might say that a whole depends for its existence on its parts (although not everyone would agree that this is the correct direction of the dependence relation, as we will see). Fundamentality comes in when we consider whether there is an end to this chain of dependence: do we ever reach the smallest parts? That is, is there a fundamental, ‘bottom level’, or does the hierarchical structure of reality continue ad infinitum? The received view has long been that there indeed is a fundamental level that everything else ‘stands on’. The fundamental level is usually thought to be at the smaller end of the spectrum: atomism suggests that there are certain (subatomic) indivisible simples, particles that are fundamental or ontologically independent. But

we must immediately note that the fundamental level must not necessarily be at the bottom, the smaller end – the fundamental end could also be at the top; that is, the universe as a whole could be considered fundamental, to be prior to its parts. We will look into all of these options in much more detail below. But before that, it should be mentioned that some recent work on this topic, 120

Fundamentality and levels of reality

taking heed of contemporary physics, has gone some way towards refuting the idea of ‘levels’ altogether. Without the hierarchical view introduced by the levels metaphor, talk about a fundamental level seems problematic. So before we can get started, we will have to clarify a number of issues, such as the levels metaphor itself. Further problems are introduced when we try to make sense of the possibility that there could be no fundamental level at all. While future physics may help to clarify the issue, it is not clear that any amount of empirical work will conclusively settle it. An important question here is whether it is possible that the levels not only go on ad infinitum, but are also infinitely complex. So far we have seen that each consecutive level brings with it certain new properties, but could it turn out that at some point it’s just ‘turtles all the way down’? This is the idea that there is an infinity of levels but the same structure repeats and hence we would have infinity without complexity.1 But before we attempt to make sense of these ideas, we should get clearer about the state of the art regarding fundamentality. The cone in Figure 6.1 represents the hierarchical structure of reality, with the smallest thing(s) at the bottom and the universe as a whole at

Figure 6.1

1

This type of repetitive structure has been coined as ‘boring’ in Jonathan Schaffer, ‘Is There a Fundamental Level?’, Noûs 37 (2003), pp. 498–517. For further discussion on this type of boring infinite descent, see Tuomas E. Tahko, ‘Boring Infinite Descent,’ Metaphilosophy 45.2 (2014), pp. 257–269.

121

122

Fundamentality and levels of reality

the top. Why a cone? Because it reflects the idea that there is a scale to the structure of reality. Another reason: it seems natural to think that there are fewer different kinds of things at the smaller end of the spectrum. The Standard Model of particle physics currently postulates that there are 61 elementary or fundamental particles (if we count particles and their corresponding antiparticles as well as the various colour states of quarks and gluons). In contrast, the larger end of the spectrum may be considered to encompass all the different kinds of things there are in reality – perhaps even infinitely many different kinds. The lines at each end of the cone represent the hypothetical bottom and top levels. The dotted sections represent the possibility of infinite descent and infinite ascent, beyond the hypothetical bottom and top levels. Four options immediately arise regarding the hierarchical structure of reality: (1) Closed (i.e., the chain of dependence terminates) at both ends of the cone. (2) Open at the top, but not at the bottom. (3) Open at the bottom, but not at the top. (4) Open at both ends. If both ends of the cone are closed, then we might consider either one of them as fundamental. If only one end is closed, then it appears that only the closed end could be considered fundamental, whether it is the top or the bottom. This follows directly from the idea that at the fundamental end the chain of dependence must terminate. The first three options enable different varieties of metaphysical foundationalism, which, in its broadest sense, states simply that there is a fundamental level. The fourth option, however, cannot support metaphysical foundationalism, as there is neither a top nor a bottom level. In this case, only some sort of metaphysical infinitism would seem to be available. However, this description is somewhat simplified, as we have not yet said much about what ‘fundamentality’ amounts to. We will do so in due course. There are three questions that this chapter is primarily concerned with: 1. Is there a fundamental level of reality, at either end of the spectrum? 2. If there is, what is ‘fundamental’ about it? 3. Supposing there is such a level, how can we know what it is like?

Fundamentality and levels of reality

Many philosophers working in metaphysics may be inclined to defend a positive answer to the first question, or at least they are likely to give it a shot. It would be quite a task to conclusively settle the first question – that is not the aim of this chapter – but it will be worthwhile to research whether there is any hope of settling it in the first place. But it is perhaps the second, methodological question, as well as the third question which is closely related to the second, that are the most interesting, and they have indeed received increasing attention in recent literature. Given the recent explosion in the literature concerning fundamentality on the one hand and the relevant scientific issues on the other, it is not possible to present a comprehensive survey. It is, however, necessary to discuss both scientific and philosophical issues in order to study each of the three questions. We might start with a critical view by James Ladyman and Don Ross,2 who argue that reality is not organized into levels in the first place and that we have some good reasons to think that there is no fundamental level. However, before the discussion can rightly begin, two issues must be clarified. First, is there a plausible interpretation of the ‘levels’ metaphor? Secondly, what is the relevant understanding of ‘fundamentality’? Both of these issues will receive a preliminary treatment in the first section of this chapter, where we will also use some of the tools that we learned about in Chapter 5. In the second section we will consider some arguments in favour of and against so-called mereological fundamentality , which is the usual or at least a very common understanding

of ‘fundamentality’.3 After this, the discussion gets somewhat more controversial and technical, and the reader is advised to refer to cited material where needed or simply skip some sections of the chapter unless interested in the more speculative content. The third section attempts to specify some of the more technical assumptions behind fundamentality, especially one notion which is often relied upon when trying to specify fundamentality, namely well-foundedness. This is a notion with set-theoretical srcins. We will consider the role of well-foundedness in this discussion and see if it can indeed be of use in clarifying the key idea behind fundamentality. We will also take advantage 2

3

See J. Ladyman and D. Ross (with D. Spurrett and J. Collier), Every Thing Must Go (Oxford University Press, 2007), especially pp. 4, 53–57, and 178–180. See for instance Schaffer, ‘Is There a Fundamental Level?’

123

124

Fundamentality and levels of reality

of the discussion in Chapter 5 to identify what type of ontological dependence is relevant in the present discussion. With these specifications in place, in the fourth section we will attempt to define a broader understanding of ‘fundamentality’, namely generic ontological fundamentality. One motivation to search for a more generic understanding is that mereological fundamentality faces some serious challenges. This takes us to a discussion of an alternative understanding of the notion of ‘part’. Finally, an analysis of evidence from physics regarding fundamentality will follow in the fifth section. This section is obviously heavy on the physics and may be safely skipped by readers who do not have a specific interest in the issue.

6.1 The ‘levels’ metaphor Our inquiry begins with an analysis of the ‘levels’ metaphor: what do we mean when we say that reality is organized into levels? In philosophy, the classic account of a hierarchical, layered understanding of reality comes from Paul Oppenheim and Hilary Putnam.4 They suggest that levels of reality are based on science and that the hierarchy is manifested by a reduction: macrophysical phenomena will be reduced to microphysical phenomena as a part of a general physicalist reduction, and the ‘levels’ metaphor describes this layered, reductive structure. Specifically, the entities of a higher level of reality can be reduced to their parts – entities of a lower level of reality – and in fact are already contained in the lower level. Therefore, all entities, as they ultimately reduce to the fundamental level, are already contained in it. For example, a biological organism occupies a relatively high level of reality. We can explain the biological processes of the organism in terms of lower-level phenomena such as biochemical reactions, reducing the biological organism to its cellular structure. But we can go even further, reducing the cellular structure to molecular structure and the chemical reactions to quantum chemistry, taking into account the atomic and subatomic constituents of the molecules in the organism’s cells. Before long, we will have reduced the biological organism to fundamental physics, presumably reaching the lowest level of reality. 4

Paul Oppenheim and Hilary Putnam, ‘Unity of Science as a Working Hypothesis,’ in H. Feigl et al. (eds.), Concepts, Theories, and the Mind-Body Problem, Minnesota Studies in the Philosophy of Science (Vol. II, pp. 3–36) (Minneapolis: University of Minnesota Press, 1958).

6.1 The ‘levels’ metaphor

While this description may appear plausible at first sight – and no doubt there are plausible reductions of this type – the Oppenheim–Putnam (henceforth OP) account has been almost universally rejected. The primary worry is the built-in reductionism of the account, for it is very difficult indeed to provide a reductive principle which would explain all macrophysical phenomena in terms of fundamental physics. One could perhaps appeal to future physics here, suggesting that eventually we will be able to account for everything in terms of fundamental physics, but any argument that has to appeal to future physics will suffer from the obvious problem that we do not yet have a clear idea what future physics will look like. In any case, before such a reduction can be demonstrated, there is no reason to just assume that it is possible. There are also other concerns: Ladyman and Ross reject this picture because of its commitment to atomism – the thesis that all matter is constituted of indivisible ‘atoms’, familiar from the pre-Socratic philosopher Democritus. This is a view which they deny.5 All in all, the worry is that the OP account imposes much too strong a requirement: it requires that each level of reality corresponds to a theory in such a way that we can start, say, from psychology, and proceed all the way to microphysics, via biology and chemistry, reducing each discipline to the theory on the next level down in the hierarchy. It is easy to see that this requirement is extremely difficult to fulfil. Despite tremendous attempts, completely reducing psychology to, say, neurochemistry has not been achieved. This is the case even with seemingly more promising reductions, such as the reduction of chemistry to physics. One problem here is that it would first have to be specified what kind of reduction we even have in mind. For instance, can we account for the notion of ‘chemical substance’ in quantum mechanics and would this constitute an appropriate reduction?6 This is far from straightforward, and certainly not as straightforward as the OP account suggests. However, more recently there has been a surge of alternative accounts concerning the ‘levels’ metaphor, some of which may fare better. Some of the recent accounts of levels are indeed non-reductive, and also deny the commitment to atomism that Ladyman and Ross find troublesome.

5 6

See Ladyman and Ross, Every Thing Must Go, p. 47. For discussion, see Jaap van Brakel, ‘Chemistry and Physics: No Need for Metaphysical Glue,’ Foundations of Chemistry 12 (2010), pp. 123–136.

125

126

Fundamentality and levels of reality

One such approach comes from Alexander Rueger and Patrick McGivern.7 They suggest that the hierarchy of reality should be understood as a sequence, not of entities, but of behaviours of entities. These behaviours, they claim, are not ordered according to spatial part–whole relations as in the OP account – this is what helps to avoid reductionism. This type of account might even be motivated by current physics: When physicists talk about levels, they often do not have in mind a mereological ordering of entities. Instead, what they describe is best understood as a stratification of reality into processes or behaviours at different scales. To describe a system’s behaviour at a particular scale, we first specify a set of equations that represent the relevant features of the system at that scale. It is then the solutions to those equations – for instance, the integration of an equation over some time interval – that describe the behaviour of the system on that scale. Note that ‘behaviour’ is understood very broadly here as the distribution of properties of a system over space and/or time.8

As Ladyman and Ross are opposed to the ‘levels’ metaphor in general, they would presumably not be satisfied with this alternative approach. They are not alone in questioning the metaphor.9 Moreover, the notion of ‘scales’ used by Rueger and McGivern would seem to do much of the work of the srcinal ‘levels’, so it is not entirely clear that this account manages to avoid the mereological connotations of levels. There are also those who deny that there is a fundamental level, but accept the levels metaphor in some form.10 But why exactly do Ladyman and Ross abandon the levels metaphor – what is so problematic with the mereological ordering? One of their major concerns appears to be that the classic accounts simply assume atomism and take granularity to be the primary criterion for distinguishing levels. However, if something along the line suggested by Rueger and McGivern is feasible, then the atomistic assumption may not be necessary for the 7

8 9

10

Alexander Rueger and Patrick McGivern, ‘Hierarchies and Levels of Reality,’ Synthese 176 (2010), pp. 379–397. Rueger and McGivern, ‘Hierarchies and Levels of Reality,’ p. 382. For further discussion, see John Heil, From an Ontological Point of View (Oxford: Clarendon Press, 2003), Ch. 2. E.g., Schaffer, ‘Is There a Fundamental Level?’; Andreas Hüttemann and David Papineau, ‘Physicalism Decomposed,’ Analysis 65 (2005), pp. 33–39.

6.2 Mereological fundamentality

levels metaphor. Another concern that Ladyman and Ross have is that the mereological ordering is not supported by current physics. They contend that the levels metaphor understood as a mereological structure ordered by part–whole relations fails as there is no good evidence for mereological atomism coming from physics.11 We’ll look into their objection in more detail below. But as we will see in due course, mereological atomism is not built-in to the notion of fundamentality or ‘levels’. Let us now try to specify the sense of ‘fundamentality’ at play in this discussion.

6.2 Mereological fundamentality We have already been talking about mereological atomism and its connection to fundamentality, but we have not yet defined this conception of fundamentality. Here is a simple definition: Mereological fundamentality (MF) The world is organized into mereological levels and the fundamental level is at one end of the mereological scale.

Note that (MF) has two parts: the mereological hierarchy, and the idea that fundamentality is a thesis about the fundamental mereological level. (MF) comes in two primary forms, depending on which end of the mereological scale is considered fundamental. An additional commitment that those in support of (MF) may have is that the entities at the fundamental level have the highest degree of ‘reality’. This type of view about ‘degrees of being’ can be traced all the way to Aristotle, but it has proven somewhat difficult to explicate this view and it remains unclear what the link is between such a view and recent work on fundamentality.12 Metaphors abound, but typically the idea is that x is fundamental or ontologically independent in this sense if and only if nothing grounds x.13 The notion of onto-

logical independence is familiar to us from Chapter 5, but we can now make some further use of it.

11 12

13

Ladyman and Ross, Every Thing Must Go, pp. 53–57. For a recent attempt to clarify this link, see Kris McDaniel, ‘Degrees of Being,’ Philosophers’ Imprint 13.19 (2013). McDaniel argues that there is indeed such a link and that we can (and should) make use of it. See Jonathan Schaffer, ‘On What Grounds What,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.),Metametaphysics (Oxford University Press, 2009), p. 373.

127

128

Fundamentality and levels of reality

Furthermore, although the above definition of (MF) does not state it, (MF) is generally combined with the idea that the there is an asymmetric ontological dependence relation from one end of the scale to the other. It is the direction of this dependence which divides proponents of (MF) into pluralists and monists , although we should note that pluralism and monism are views independent of (MF) and hence variations of pluralism and monism without (MF) are possible. 14 When combined with (MF), the pluralists hold that the direction of the dependence is from the larger to the smaller, resulting in mereological atomism

– this is, per-

haps, the standard view. The monists think that the parts are dependent on the whole and hence that there is only one fundamental entity, namely the universe. One proponent of the monist view is Jonathan Schaffer, who considers only substances to be fundamental, and further, that there is exactly one substance, namely the cosmos. On this understanding, a substance is a basic or fundamental, ontologically independent entity. This is roughly how Aristotle might characterize the notion of ‘substance’. It is the cosmos, conceived as a substance, which is prior to its (arbitrary) parts. Schaffer’s Spinoza-inspired

priority monism , how-

ever, does not assume mereological atomism. In fact, one of Schaffer’s arguments in favour of priority monism is that it is compatible with the possibility of atomless gunk , whereby matter is infinitely divisible ‘gunk’ and matter or objects have no smallest parts; there are no mereological simples. The question of gunk is of some importance here, as those committed to (MF) + pluralism and the idea that the dependence relation goes from the larger to the smaller will struggle to accommodate the infinite divisibility of matter entailed by a ‘gunky’ ontology

– we will

return to this shortly. Even though (MF) combined with mereological atomism is something of a default view when it comes to fundamentality, it is also the most resisted: as we saw above, Ladyman and Ross strongly oppose the view. Most opponents of (MF) are worried about the additional commitment to mereological atomism. The specific worry that Ladyman and Ross have is directed precisely against mereological atomism: none of the various atomistic conceptions can stand the test of contemporary physics. The 14

For further discussion on the monism/pluralism issue as well as an alternative account of fundamentality, see Kelly Trogdon, ‘Monism and Intrinsicality,’Australasian Journal of Philosophy 87 (2009), pp. 127–148.

6.2 Mereological fundamentality

phenomenon to which they appeal is famous: ‘quantum entanglement’. Since Ladyman and Ross consider the srcinal formulation of the idea by Erwin Schrödinger to be difficult to improve upon, let us quote it here as well: When two systems, of which we know the states by their respective representatives, enter into temporary physical interaction due to known forces between them, and when after a time of mutual influence the systems separate again, then they can no longer be described in the same way as before, viz. by endowing each of them with a representative of its own. I would not call that one but rather the characteristic trait of quantum mechanics, the one that enforces its entire departure from classical lines of thought. By the interaction the two representatives [the quantum states] have become entangled.15

The details regarding the phenomenon of quantum entanglement can be set aside here. What is important is that the question boils down to one regarding individuality. If the atomistic conception of fundamentality were correct, then we would expect to find some individuals, some mereological atoms which are, in the relevant sense, independent. But given the phenomenon of quantum entanglement, it looks as if the best candidates for individuals, the elementary particles mentioned in the Standard Model, do not behave as we would expect on the basis of the philosophical discussion.16 The core of the critique by Ladyman and Ross is that there is nothing in fundamental physics that corresponds with the atomistic idea of ‘simples’ – instead you will find a very complex arrangement of systems relating to each other in ways that we can only model probabilistically, such as the one described by Schrödinger in the passage above. However, in fairness it should be noted that the interpretation of these issues is still controversial – eighty years after the publication of Schrödinger’s paper. For instance, building on the work of the physicists David Bohm and Basil 15

16

Erwin Schrödinger, ‘Discussion of Probability Relations Between Separated Systems,’ Proceedings of the Cambridge Philosophical Society 31 (1935), p. 555; quoted in Ladyman and Ross, Every Thing Must Go, p. 19. However, the inference from quantum entanglement to the failure of individuality and hence mereological atomism can be questioned. For a discussion of this issue, see Mauro Dorato and Matteo Morganti, ‘Grades of Individuality. A Pluralistic View of Identity in Quantum Mechanics and in the Sciences,’ Philosophical Studies 163 (2013), pp. 591–610.

129

130

Fundamentality and levels of reality

Hiley, interpretations of quantum entanglement that do not drop individuality have been and are still being developed.17 But why is the atomistic conception of fundamentality so popular in the first place, if we know that contemporary physics is unlikely to support the picture? Sometimes a non-atomistic conception of fundamentality is considered. Ned Markosian discusses such a conception, but quickly dismisses it, because abandoning mereological atomism would have unwanted implications for the proponent of fundamentality – the choice would then appear to be between (MF) combined with atomism, and non-fundamentality. 18 One of the unwanted implications mentioned by Markosian is that it would supposedly be highly implausible to say that nothing is ‘maximally real’ – assuming that it is the fundamental level that has the highest degree of reality.19 What Markosian has in mind is that if the world were ‘gunky’ and there would be no fundamental mereological level, we would never reach the highest degree of reality – metaphysical grounding would, as it were, never ‘bottom out’. A related line of thought is discussed by Ross Cameron.20 Cameron resists the idea that there could be infinite chains of ontological dependence, also because of some non-intuitive implications. Applying the idea to the case of composition, Cameron notes that in a gunky world, composition would never get ‘off the ground’. In other words, if complex objects are ontologically dependent on their mereological parts, then composition never ‘bottoms out’ in gunky worlds. The intuitive result, according to Cameron, is that complex objects are not possible in gunky worlds. But such a strong result may be alarming, for it would rule out the possibility that we live in a gunky world, at least if we accept the plausible idea that complex objects are ontologically dependent on their parts. Indeed, the intuitive appeal of

17

18

19

20

For a very recent example, see Paavo Pylkkänen, Basil J. Hiley, and Ilkka Pättiniemi, Individuals Across ‘Bohm’s Approach and Individuality,’ in A. Guay and T. Pradeu (eds.), the Sciences, Ch. 12 (Oxford University Press, 2015). See Ned Markosian, ‘Against Ontological Fundamentalism,’ Facta Philosophica 7 (2005), pp. 69–84. However, degrees of reality do not necessarily have to be associated with mereological fundamentality. Ross P. Cameron, ‘Turtles All the W ay Down: Regress, Priority and Fundamentality,’ Philosophical Quarterly 58 (2008), pp. 1–14.

6.2 Mereological fundamentality

this approach has been questioned.21 There are various approaches that one could take in order to resist the result and we will consider one of them in detail below, but let us first briefly examine the intuition at stake. Cameron and many others writing about fundamentality and the idea that composition couldn’t ‘get off the ground’ without a fundamental level typically acknowledge that this appeal to intuition hardly constitutes an argument in favour of fundamentality. An obvious problem here is that our ‘folk’ intuitions may not be a reliable source of evidence when it comes to matters such as the constraints on the fundamental structure of reality – something that we will discuss further in Chapter 8. One complication concerns the way in which we think about the structure of space-time. If space-time itself is made up of zero-dimensional space-time points and hence has no internal structure, no smaller parts, then this ‘pointy’ space-time would seem to rule out a certain type of infinite regress by its very nature: surely those space-time points themselves do not depend for their existence on anything and hence are, in one sense, fundamental? On the other hand, it’s standard to think that any given collection of these space-time points composes a space-time region. These regions are identified in terms of their composite space-time points. This discussion is related to various issues in the metaphysics of space and time, but what matters to us now is that there are two different ways to conceive of the relationship between the points and the regions; as John Hawthorne puts it: In antiquity it was common for philosophers to admit the existence of points only as derivative entities and, relatedly, to think of facts about points as derivative from more fundamental facts about extended objects. On this conception, points exist only derivatively, as limits of lines, and surfaces without thickness exist only derivatively as limits of bulky objects. Nowadays, however, it is common for metaphysicians to hold both that space-time regions are less fundamental than the space-time points that compose them and that facts about the intrinsic character of space-time regions are less fundamental that the facts about the intrinsic character of 22 space-time points and facts about how the space-time points are arranged.

21

22

See for instance Matteo Morganti, ‘Dependence, Justification and Explanation: Must Reality Be Well-Founded?’,Erkenntnis 60.3 (2015), pp. 555–572. John Hawthorne, ‘Three-dimensionalism vs. Four-Dimensionalism,’ in T. Sider, J. Hawthorne and D. W. Zimmerman (eds.), Contemporary Debates in Metaphysics (Oxford: Blackwell Publishing, 2008), p. 264.

131

132

Fundamentality and levels of reality

So there is at least one open question regarding the relative fundamentality of the regions and the points. The relevant intuition that Cameron and others are appealing to may presuppose the view that points are more fundamental – this is of course what Hawthorne identifies as the more popular view, but there is also the empirical question regarding the structure of space-time. General relativity (GR) treats space-time as ‘pointy’ and hence may support the view at hand, but there are other options as well. For instance, if we apply quantum field theory to this debate, the result could very well be quite different – but this is something that we have to leave aside here. In fact, Ladyman and Ross also point out that, as is well known, the ‘conflict’ between (GR) and quantum mechanics is particularly problematic when considering small scales, such as the supposed space-time points.23 In any case, there is also a separate metaphysical question of how composition works for space-time regions. In particular, if we forget about the ‘pointy’ conception and just ‘zoom in’ to a spacetime region, then what would be problematic in assuming that we encounter infinitely many smaller regions, without ever reaching the supposedly fundamental space-time points? Here, the srcinal intuitive pull of the failure of composition to get ‘off the ground’ may be much weaker, unless it is simply assumed that it is built in to the notion of composition that everything is made up of fundamental, ‘pointy’ objects. Now, to sum up the discussion so far: infinite chains of ontological dependence are typically considered problematic, but the primary motivation for finding them problematic is that they would have non-intuitive results, such as causing problems for an intuitive view about complex objects. This has led some, like Cameron, to reject the possibility of the actual world being gunky, as there clearly are complex objects. But it is not obvious that infinite sequences of ontological dependence in themselves are necessarily problematic – we did not encounter any reasons to think so in the discussion concerning ontological dependence in Chapter 5 and there may be ways to interpret composition that have no such consequences. Moreover, what Cameron and many others seem to be concerned about are infinite sequences of mereological dependence.24 This is dependence 23 24

Ladyman and Ross, Every Thing Must Go, p. 19. Compare with Jaegwon Kim, Essays in the Metaphysics of Mind (Oxford: Oxford University Press, 2010), p. 183. Kim prefers ‘mereological dependence’ instead of the term ‘mereological supervenience’ and contrasts it with causal dependence.

6.3 Further specifications: well-foundedness and dependence

between a whole and its parts, and it is, of course, the sense of dependence that seems to be underlying (MF). One upshot of all this is that there is a wide agreement about the incompatibility of pluralism + (MF) + gunk (or metaphysical infinitism more generally). Recall that pluralism in this context refers to the idea that there is a plurality of fundamental objects (or kinds) at the fundamental level – a view that goes naturally together with mereological atomism. Most seem to be inclined to drop gunk and keep pluralism + (MF). But recall also that one could drop pluralism and adopt monism instead, as Schaffer does – Schaffer’s priority monism suggests that the universe as a whole is prior to its parts. In fact, Schaffer considers it an argument in favour of his priority monism that it is able to accommodate gunky ontologies. But we’ve also seen reasons to doubt (MF) above; the preceding considerations would seem to undermine at least one motivation for (MF), because one could accept that mereological dependence cannot be infinite while denying that the fundamental level consists of mereological simples. Given this, as well as the forceful objections against (MF) from physics by Ladyman and Ross (and others), the present status of (MF) is in doubt. But rather than try to pass judgement about any of this, we should open up the assumptions behind (MF) and contrast it with a more general sense of fundamentality.

6.3 Further specifications: well-foundedness and dependence We saw in the previous section that (MF) concerns mereological dependence. But the notion of fundamentality can be associated with another type of dependence – we already know from Chapter 5 that ontological dependence is in fact a rich family of notions. The idea that some notion of ontological dependence will help us make sense of the fundamental level of reality or of levels of reality in general came up with our discussion of Cameron’s view. Following this discussion, we might say that if there is a fundamental level, it could be defined in terms of ontological independence, that is, the existence of anything fundamental cannot depend

on something else. This idea is commonly expressed with reference to the notion of well-foundedness; we will now turn to its analysis. The notion has its srcins in set theory and the axiom of foundation, which states that a set cannot contain an infinitely descending sequence (of membership).

133

134

Fundamentality and levels of reality

In slightly different terms, the axiom of foundation requires that every non-empty set S contains an element that is disjoint from S; as a consequence, a set cannot be its own element. The upshot of the axiom of foundation is well-foundedness, which can be applied to any order (which we have generally called a ‘chain’ above), such as proper parthood. Something x is a proper part of y when x and y are not identical, but y contains x in its entirety. This is really just the intuitive conception of parthood, but the

qualifier ‘proper’ is needed to rule out reflexivity, that is, something being a part of itself. To put this in terms of the notion of grounding, which is typically used in this context, we could say that everything must be fully grounded in something fundamental.25 With this in place, we can define well-foundedness as follows: Well-foundedness (WF) An order is said to be well-founded if every non-fundamental element of the order is fully grounded by some fundamental element(s).26

If an order is thought to be well-founded, then infinite descent is ruled out in that order. As we have seen, an atomless, gunky ontology does not satisfy (WF), as in such an ontology we have an infinite descending proper parthood order. However, although in this context it is the classic notion of parthood that has received the most attention, the notion of well-foundedness is not tied to parthood. In the next section we will specify a more general understanding of fundamentality, but for the case of (MF) it is exactly well-foundedness with regard to proper parthood that helps us to formulate the idea of a fundamental level. Note that (WF) is typically not considered to rule out chains that are infinite just so long as they terminate eventually, even if it would take an infinite number of steps (or an infinite amount of time) to reach the fundamental level. To see this, consider the idea that the universe is infinitely large, without a ‘top level’, but having a foundation, say, as required by (MF) (i.e., mereological atomism). This entails that an infinite chain directed towards the mereological atoms is possible without violating (WF): there is a fundamental level, it just takes infinitely many steps to reach it! This type of

25 26

For the distinction between full and partial ground, see section 5.4. There are various other possible formulations of well-foundedness, for a recent survey, see Scott Dixon, ‘What Is the Well-Foundedness of Grounding?’, Mind (forthcoming).

6.3 Further specifications: well-foundedness and dependence

idea is partially analogous to the familiar Zeno paradoxes, such as the race between Achilles and a tortoise. As the story goes, Achilles gives a head start to the tortoise, but in the time that it takes Achilles to reach the starting point of the tortoise,p1, the tortoise has already reached pointp2. So Achilles must reach point p2, but since he does not move instantaneously, the tortoise has reached point p3 in the meanwhile; this way it seems that Achilles will never be able to overtake the tortoise. Of course, we know very wellthat this sequence, although infinite, does terminate– Achilles will indeed overtake the tortoise. The case of infinite chaining is somewhat different in the sense that it would take an infinite amount of ‘time’ to reach the fundamental level, so it is not directly analogous. But as long as there is a fundamental level, infinity can perhaps be dealt with. This idea can be seen in Schaffer’s work, as well as in Karen Bennett’s.27 For instance, Bennett requires that a well-founded chain must not be infinite at the fundamental end, even though well-foundedness is compatible with an infinity of grounded entities. A chain can be infinite and well-founded, that is, grounded in something fundamental. As Ricki Bliss puts it:‘A finite grounding chain is well-founded buta well28 founded grounding chain need not be finite’. The upshot is that (WF) only

requires that the grounding chain eventually terminates, even if it takes infinitely many steps to reach the fundamentallevel. Given that we are now laying the foundations for a notion of fundamentality broader than (MF), we also have to characterize a notion of ontological dependence that would be able to maintain the type of hierarchical chain of dependence or order that could explicate it. We are looking for asymmetric sequences of ontological dependence that could start, for instance, from macrophysical objects and terminate at a fundamental level. It seems that only a rather general sense of ontological dependence will do the trick. One possibility which would appear to be sufficiently general would be a sequence of generic existential dependence:29’ Generic existential dependence (GXD) x depends for its existence upon Fs =df Necessarily, x exists only if some Fs exist.

27

28

29

See Schaffer, ‘Is There a Fundamental Level?’, pp. 509–512; Karen Bennett, ‘By Our Bootstraps,’ Philosophical Perspectives 25 (2011), p. 34. Ricki L. Bliss, ‘Viciousness and the Structure of Reality,’Philosophical Studies 166.2 (2013), p. 416. Originally introduced in section 5.1.

135

136

Fundamentality and levels of reality

Where F is a general term: necessarily, x exists only if another object of the type F exists. (GXD) refers to an object’s existence requiring the existence of an object of a certain sort rather than a specific object (as opposed to rigid existential dependence). For instance, any given w ater molecule cannot exist unless hydrogen atoms exist. Or, more precisely, any given water molecule cannot exist unless the physical constraints required for the existence of hydrogen atoms are fulfilled, which in turn requires the possibility of a certain electron configuration, among other things. Since we are typically interested in the ontological basis of macrophysical objects, it is perhaps more natural to think about dependence on a specificrange of values for the fundamental physical constants, or on certain physical principles – we will consider these requirements in the final section of this chapter. With (WF) and (GXD) in place to specify a more general sense of hierarchy, compatible with the mereological hierarchy of (MF) but also with other ways to understand ‘levels’, we can move on to specify a generic sense of fundamentality.

6.4 Generic ontological fundamentality The term ‘generic ontological fundamentality’ is not established in the literature, even though the idea of a more general sense of fundamentality in addition to the mereological version (MF) is widely recognized. Typically, discussion about fundamentality focuses on versions of (MF). One reason for this is simply the historical dominance of atomism. But if we are open to the possibility of conceiving the hierarchical structure of reality in terms of something other than the mereological scale and hence mereological dependence, then we must make room for different conceptions of fundamentality as well. As we saw in the previous section, the underlying notions of well-foundedness and ontological dependence do enable this. However, rather than attempting to list all the possible ways to understand levels and the relevant chains of dependence that structure them, we will here try to define a very general sense of fundamentality, which could capture (MF) but also many, if not all, other conceptions of fundamentality. We define this general conception of fundamentality as follows: Generic ontological fundamentality (GOF) The world is organized into levels of ontological elements and the fundamental level consists of ontologically minimal elements.

6.4 Generic ontological fundamentality

This definition is only provisional and it is obviously in need of further explication, as nothing has yet been said about ‘ontological elements’ in general and ‘ontologically minimal elements’ in particular. But since (GOF) is supposed to be able to express (MF) as well, we can already guess at least one thing that could take the place of ‘ontological elements’ in (GOF), namely mereological elements. In that case, we would get the familiar hierarchy of mereological levels and the ontologically minimal elements would be mereological atoms – at least if (MF) is combined with pluralism. Of course, (GOF) is just as compatible with pluralism as it is with monism, since there would be just one ontologically minimal element, the universe. But what else could the ontological elements be? That would seem to depend on three things. First, what kind of hierarchy are we interested in? Secondly, what kind of dependence structures that hierarchy? Thirdly, how is the hierarchy manifested in the world, or in other words, what is the relevant physics? All of these are difficult questions, especially the last one. If the critique of (MF) by Ladyman and Ross that we discussed in section 6.2 is correct, then much of the relevant work in determining the correct answers to these three questions will come from physics. However, even if we will eventually need more input from physics, this does not mean that we can’t say anything more at the moment. At the very least, we can try to get a clearer idea about what is required from ontologically minimal elements – the theoretical background of ‘ontological minimality’. To get an initial idea, the ontologically minimal elements could be compared to the idea of ‘minimal truthmakers’, familiar from David Armstrong: If T is a minimal truthmaker for p, then you cannot subtract anything from T and the remainder still be a truthmaker for p.30

A minimal truthmaker for a given proposition can be understood as the smallest or least encompassing portion of reality that fully grounds the truth of that proposition. This analogy between minimal truthmakers and ontologically minimal elements suggests that the fundamental level consists of the least encompassing portion of reality. In usual accounts of minimal truthmakers, the least encompassing portion of reality is understood

30

David M. Armstrong, Truth and Truthmakers (Cambridge University Press, 2004), pp. 19–20.

137

138

Fundamentality and levels of reality

in terms of parthood, which would naturally take us back towards (MF) rather than (GOF). But there is another way to understand it, which we might compare with what Ted Sider calls ‘ideological’ fundamentality, taking the notion of ‘ideology’ from Quine. 31 Sider uses the notion of fundamentality interchangeably with ‘joint-carving’ and ‘part of reality’s structure’. He specifies that his conception of fundamentality is ideological rather than mereological or propositional. This can be further illustrated in terms of grounding: Every non-fundamental truth is grounded by some fundamental truth. It should be noted though that Sider distances himself from certain popular treatments of ground, which we discussed in Chapter 5. In contrast to Schaffer, Sider is not committed to entity-grounding , which points towards (MF). Indeed, Sider claims that his ideological fundamentality is compatible with infinite mereological descent – a gunky ontology. Similarly, Sider claims to be able to accommodate infinite propositional descent, for instance a proposition about a gunky object having a certain mass is true in virtue of a proposition about the masses of its parts going one level down, and that proposition in turn is true in virtue of a proposition about the masses of the object’s parts on a further level down, and so on, ad infinitum. So how should we understand Sider’s ideological fundamentality? He specifies that some notions are fundamental in the sense that there is a perfectly fundamental description of reality expressible by terms that carve perfectly at the joints – the ‘joint carving’ metaphor was initially discussed in Chapter 4. Note also that Sider is committed to a type of levels of reality view, as he thinks that there are degrees of fundamentality and hence degrees of joint-carving – perfect joint-carving being at the ideological fundamental level. The connection between Sider’s approach and (GOF) is that the ontologically minimal elements can also be understood in terms of perfect joint-carving: a given term is part of the fundamental level if there is no other term that describes reality more minimally. Note however that Sider is primarily interested in which concepts carve at the joints (as discussed in Chapter 4). This brings with it a linguistic

31

For an account of minimal truthmakers, see Tuomas E. Tahko and Donnchadh O’Conaill, ‘Minimal Truthmakers,’ Pacific Philosophical Quarterly (2015 [online], forthcoming). For the details of Ted Sider’s ‘ideological fundamentality,’ see his Writing the Book of The World (Oxford University Press, 2011); see also section 4.4 of this book.

6.4 Generic ontological fundamentality

connotation, which we may wish to distance ourselves from given the present goal. Sider’s picture goes some way towards clarifying the idea of ontological minimality, but there may be a more neutral way to avoid the commitment to (MF), with the help of the analogy concerning minimal truthmakers and parthood. We might try to understand the notion ofparthood itself in a more liberal manner. Arecent attempt to do this comes from Kit Fine, who develops an alternative to the classic notion of ‘part’: When one object is a part of another, there is a sense in which it is in the other – not in the sense of being enclosed by the other, as when a marble is in an urn, but more in the sense of being integral to the other. When parts are in question, it is also appropriate to talk of a given object being composed of or built up from the objects that it contains.32

Fine suggests that all manner of things, from sentences to symphonies to sets, are composed of other things. Accordingly, he proposes to define the general relation of part in terms of the operation of composition: ‘The parts of an object are the object itself, or its components, or the components of the components, and so on’.33 A question that remains open even if we were to adopt this more liberal sense of parthood is whether infinite chaining of these components is subject to similar ‘non-intuitive’ consequences as the more traditional notion is. This does not follow automatically though: it would seem possible that the components of a composite object one level down are integral to the object, but if the components going down one level further are no longer integral to it, then the infinite regress causing the typical non-intuitive consequences wouldn’t obviously follow. So it turns out that there are ways to understand parthood which do not necessarily imply the non-intuitive consequences that have been associated with infinite chains of mereological dependence. This opens up the possibility of interpreting ontologically minimal elements quite openly indeed: the smallest, minimal ‘parts’ of reality do not need to be mereological elements at all, they can be anything that count as components, such as structures, relations, objects, or whatever. This may enable us to overcome at least some of the complications regarding the tension between a classic, object-oriented metaphysics such as mereological atomism and the 32 33

Kit Fine, ‘Towards aTheory of Part,’Journal of Philosophy 107.11 (2010), p. 560. Ibid., pp. 567–568.

139

140

Fundamentality and levels of reality

type of view defended for instance by Ladyman and Ross, according to which reality is fundamentally relational or structural.34 Now that we have a clearer idea about what fundamentality amounts to, let us briefly look at one outlier: the possibility of a repetitive version of metaphysical infinitism – the view that the grounding chain does not terminate. The idea is that the same structure could repeat infinitely. This is an idea also entertained by Schaffer; he calls this type of infinite descent ‘boring’, or in other words, the same structure keeps repeating – it’s ‘turtles all the way down’. 35 In other words, there is no novelty in the structure after a certain point. The part of the structure that repeats could be of any length, as long as it starts anew at some point. What makes this possibility interesting is that such a repeating structure might even be enough to satisfy (GOF). To see this, consider the idea of ontologically minimal elements again, as introduced with the definition of (GOF). The infinite descent of a repetitive structure could allow for an ontologically minimal description in the sense that a description of the repetitive part only needs to be supplemented with an instruction to continue as before, for instance ‘the world stands on four elephants, the four elephants stand on a turtle, the turtle stands on two camels, the camels stand on four elephants, the four elephants stand on a turtle … and repeat ad infinitum’. No other terms than these four elephants, a turtle, and two camels

can be introduced that would describe reality more minimally – they carve perfectly at the joints and hence could be understood to constitute the fundamental level in the sense of (GOF). But we are treading on very controversial terrain here – this option is not yet established in the literature, even though it is receiving more and more attention.36 Summarizing, it appears that much remains to be done in order to specify the range of options with regard to fundamentality. The generic notion 34

35

36

For further discussion of this type of structuralist view and fundamentality, see for instance Kerry McKenzie, ‘Priority and Particle Physics: Ontic Structural Realism as a Fundamentality Thesis.’ British Journal for the Philosophy of Science 65 (2014), pp. 353–380; Steven French, The Structure of the World: Metaphysics and Representation (Oxford University Press, 2014). Schaffer, ‘Is There a Fundamental Level?’, p.505. See also Tahko, ‘Boring Infinite Descent.’ See for instance Matteo Morganti, ‘Dependence, Justification and Explanation: Must Reality Be Well-Founded?’; Matteo Morganti, ‘Metaphysical Infinitism and the Regress of Being,’ Metaphilosophy 45.2 (2014), pp. 232–244; Ricki Bliss, ‘Viciousness and Circles of Ground,’ Metaphilosophy 45.2 (2014), pp. 245–256.

6.5 Fundamentality and physics

(GOF) sketched here is designed to be as liberal as possible, in order for the various possible conceptions of fundamentality to be captured with one general notion, but it must be stressed that there is not yet an established usage for such a notion.

6.5 Fundamentality and physics We have already discussed some of the limitations that contemporary physics appears to impose on our discussion of fundamentality. But the idea of a fundamental level is certainly present in contemporary physics, even if it is rarely discussed in much detail. This is of course partly because of empirical limitations. For instance, the physicist David Bohm discusses the possibility of an ‘infinity of levels’, stating that current physics is entirely compatible with this possibility.

37

He seems to think

that this is a purely empirical matter (this is what motivates Schaffer’s liberal attitude towards the idea as well). Bohm makes one general observation, concerning the requirement of ‘autonomy and stability’ in modes of being, which suggests that regardless of whether there is infinite descent or not, it is clear that the very idea of a ‘thing’ is intimately connected with some kind of stability and autonomy of the thing (object, entity, process …) in question – leaving aside the concern about individuality that we came across when discussing the critical points about fundamentality from Ladyman and Ross. Perhaps we can say something more about this requirement of stability, but it will require going into some detail about the relevant physics. However, the reader should not be alarmed if some of the details seem obscure, as this is all very speculative. Since this is not a physics textbook, we will not define all the relevant scientific terminology, but will instead assume that interested readers have some familiarity with the topic. Those who do not wish to engage with the remaining material may safely skip this last section of the chapter. Our starting point is very simple. For there to be any macrophysical objects, the forming of such objects must be possible. Consider the minimal conditions for the possibility of stable macrophysical objects.

37

David Bohm, Causality and Chance in Modern Physics, 2nd ed. (London: Routledge, 1984), p. 95.

141

142

Fundamentality and levels of reality

A natural way to spell out these minimal conditions is in terms of the physical laws that govern the forming of macrophysical objects. For instance, it is necessary for the existence of (most) stable macrophysical objects that molecules can bind together to form molecular complexes, albeit some covalently bonded networks of carbon atoms such as diamonds or graphite lattices are examples of macrophysical, extremely large molecules. At any rate, it is necessary for any macrophysical object that atoms are able to form bonds to create stable molecules, and further, that subatomic particles are able to form stable atoms. So what enables atoms to form bonds and subatomic particles to form atoms? Well, the binding of molecules and atoms is dependent on the electron configuration of individual atoms, which in turn depends on the energy levels of specific electrons and is moderated by the Pauli Exclusion Principle (PEP). 38 Similarly, the manner in which subatomic particles form atoms is dependent on the individual charges of subatomic particles, namely the negative charges of the electrons and the positive charges of the protons, where each proton consists of three quarks which make up the total charge of the proton. At this point it should be noted that although we have been referring to subatomic particles , there is no reason to assume mereological atomism here, bearing in mind the discussion of individuality in the second section and the further point later on regarding the tension between a classic, object-oriented metaphysics such as mereological atomism and the type of view defended by Ladyman and Ross, according to which reality is fundamentally relational or structural. All we need is a microphysical arrangement which enables the possibility of macrophysical objects. Accordingly, whether or not we view electrons as particles with some sort of individuality, we know that their behaviour is subject to (PEP). More specifically, if we have two identical and indistinguishable electrons, the wavefunction for the system of those two electrons must be antisymmetric. As an example, let us consider a case of covalent 38

The Pauli Exclusion Principle states that no two identical fermions, such as electrons, can have all the same quantum numbers (e.g., spin, angular momentum, energies of electrons) at the same time. This means, for instance, that if two electrons in an atom occupy the same orbital and hence have the same energy, then they must differ with regard to another quantum number, namely spin.

6.5 Fundamentality and physics

bonding in a hydrogen molecule. For a covalent bond to form between two hydrogen atoms, the complete wavefunction of the system, which combines the spin and spatial wavefunctions, must be antisymmetric. Accordingly, if the spin wavefunction is symmetric, then the spatial wavefunction must be antisymmetric. However, only a symmetric spatial wavefunction will lead to bonding, as we need an attractive force between the hydrogen atoms, enabling them to share a pair of valence electrons. Hence, since (PEP) requires the complete wavefunction to be antisymmetric, we know that the spin wavefunction must be antisymmetric for a bond to form. We also know that a third hydrogen atom cannot bond with the two-atom hydrogen molecule, as it would necessarily have an antisymmetric wavefunction with one of the hydrogen atoms, and would therefore be repelled. As should be clear from this description, the Pauli Exclusion Principle is crucial for the process of covalent bond forming. In the case of ionic bonds, such as the bond between sodium and chlorine in sodium chloride molecules, (PEP) is responsible for the repulsive force between Na+ and Cl ions. As the ions come closer and the wavefunctions of −

their electrons start to overlap, (PEP) requires that these electrons cannot be in the same quantum state. The situation is resolved by a change in the energy levels of the electrons so that no two identical electrons occupy the same quantum state. The change in the energy levels of electrons requires energy and results in the repulsive force, typically called Pauli repulsion. This repulsive force prevents the ions from coming any closer together. The result is a stable sodium chloride lattice. Again, (PEP) is absolutely central for this bonding process. Since (PEP) governs both the bonding behaviour of atoms and the electron configuration of individual atoms, it is crucial for the emergence of stable macrophysical objects. It is sometimes said that (PEP) is responsible for the space-occupying behaviour of matter, as it prevents atoms from collapsing together: the electrons must occupy successively higher orbitals to prevent a shared quantum state and hence not all electrons can collapse to the lowest orbital. Incidentally, it is (PEP) that explains why subatomic particles can behave in a manner which is so different from macrophysical objects. The principle is key to understanding why fundamental physics cannot be viewed as a network of ‘microbangings’ – something that

143

144

Fundamentality and levels of reality

Ladyman and Ross accuse metaphysicians of doing in their search for ‘genuine causal oomph’.39 The ‘microbangings’ model of fundamental physics is indeed not very fruitful, but the idea of levels of reality and of a fundamental level does not need to be intimately connected with that model – this is what we are in the process of showing here. We’ve seen that macrophysical objects depend for their existence on certain principles. These principles themselves depend on a distribution of fundamental physical constants that fall within a specific range. The dependence here is transitive: ultimately macrophysical objects depend for their existence on the distribution of fundamental physical constants, as will become clear in what follows. One important fundamental physical constant associated with the stability of matter (and hence the existence of macrophysical objects) is the elementary charge, namely 1.6021892

×

10

−

19

coulombs. This is the charge

of a proton, whereas an electron has a negative charge of equal strength. Interestingly, the charges of all other freely existing subatomic particles that have a charge are either equal to or an integer multiple of the elementary charge. Quarks, which are the constituents of protons, have charges that are integer multiples of one third of the elementary charge, but they are not freely existing as individiual particles. The total charge of the atom is of course neutral. The picture gets somewhat more complicated when details about the underlying fundamental forces are introduced; for instance, the nucleus holds together in virtue of the strong nuclear force , which overpowers the repulsive forces between the (net) positively charged quarks that make up protons. (Quarks can form quark–antiquark pairs – mesons – and combinations of three quarks – baryons, such as protons and neutrons – which are collectively known as hadrons.) What is interesting for us is whether fundamental physical constants, such as the elementary charge, are physically necessary for the stability of matter. In other words, if the fundamental constants had been different, would atoms still be stable? It has been suggested that at least some of the fundamental physical constants do, or at least could vary over time, specifically the fine-structure constant (sometimes called the electromagnetic force coupling constant), 39

Ladyman and Ross, Every Thing Must Go, p. 4. By ’microbangings’, Ladyman and Ross mean something like what has traditionally been called ’billiard-ball physics’, i.e., radically simplified physics in which particles are imagined to be tiny spheres bouncing off each other.

6.5 Fundamentality and physics

which characterizes the strength of the electromagnetic interaction, and the electron-to-proton mass ratio. However, there are experimental limitations in determining whether fundamental physical constants do indeed vary over time: observations commonly entangle a certain set of constants, or assume that certain constants do not vary. It is also worth noting that if all the fundamental physical constants were to vary together so that their ratios would remain the same, it would presumably be impossible to establish this observationally.40 If we hope to establish some limitations concerning the possible variation of fundamental physical constants so that this variation would not jeopardize the stability of matter, then we should direct our attention towards the values of the constants relative to each other. It is plausible that if all the constants were to change so that their proportions would not change, then the stability of matter might not be in jeopardy. However, if we change just one of the constants, this would be much more likely to cause problems. Take the fine-structure constant, α: it is a dimensionless constant which is expressed in terms of other physical constants, namely the elementary charge, the electric constant, the Planck constant, and the speed of light; the numerical value of α 1 is just over 137. A dimension−

less constant serves our purposes very well, as a change in a dimensionless constant implies that the proportions between constants have changed. The electron-to-proton mass ratio, β, is another good example. Indeed, had either one of these fundamental physical constants been of a slightly different value, then macrophysical phenomena would not be possible: For example, if we were to allow the ratio of the electron and proton masses β = me/mN and the fine structure constant α to change their values (assuming no other aspects of physics are changed by this assumption – which is clearly going to be false!), then the allowed variations are very constraining. Increase β too much, and there can be no ordered molecular structures because the small value of β ensures that electrons occupy well-defined positions in the Coulomb field created by the protons in the nucleus. If β exceeds about 5 × 10 3 α2, then there would be no stars. If −

modern grand unified gauge theories are correct, then α must lie in the

40

For further discussion regarding fundamental physical constants especially as they are related to laws of nature, see Tuomas E. Tahko, ‘The Modal Status of Laws: In Defence of a Hybrid View,’The Philosophical Quarterly (2015 [online], forthcoming).

145

146

Fundamentality and levels of reality

narrow range between about 1/180 and 1/85 in order that protons not decay too rapidly and a fundamental unification of non-gravitational forces can occur. If, instead, we consider the allowed variations in the strength of the strong nuclear force, αs, and α then roughly αs < 0.3α1/2 is required for the stability of biologically useful elements like carbon. 41

It should be emphasized that it would not be physically possible to vary individual constants in the simplistic manner imagined in this scenario: a change in one constant would have ramifications throughout the range of physical constants, as Barrow notes.42 However, it does seem clear that fundamental physical constants must fall within a specific, fairly narrow range to enable the emergence of macrophysical objects. What does all this tell us about the fundamental level? The answer depends on what we take to be the ontological requirements for the described stability of matter. Note that, on the face of it, we can only make inferences about the actual world based on actual physics. If we could produce an argument to the effect that this stability is not possible without a fundamental level, then we would have a good case at least for the existence of a fundamental level in the actual world – or, perhaps, any world where macrophysical objects exist. Yet it seems that no amount of empirical research could establish that a fundamental level is metaphysically necessary. So unless there is some a priori argument available to the effect that

the actual structure of reality is necessary, then we ought to draw inferences only regarding the actual world. We do however have a fairly good idea concerning the actual requirements for the forming of macrophysical objects. This brings us back to the physical requirements regarding the sequence of generic existential dependence (GXD), which were mentioned in passing in the third section. One way to interpret these requirements is exactly in terms of the type of constraints that have been listed above: namely a list of requirements for the stability of matter, including physical principles such as the Pauli Exclusion Principle and a limited fluctuation of the fundamental physical constants. As the quotation from Barrow demonstrates, macrophysical objects are extremely fragile and the range for 41

42

John D. Barrow, ‘Cosmology, Life, and the Anthropic Principle’, Annals of the New York Academy of Sciences 950 (2001), p. 147. This is not the only factor that we simplify for the purposes of illustration. For a more rigorous account, see for instance Marc Lange, ‘Could the Laws of Nature Change?’, Philosophy of Science 75.1 (2008), pp. 69–92.

6.5 Fundamentality and physics

the appropriate values of physical constants is narrow. Such precision, one might think, would not be possible with full-blown infinity. The philosophical idea here is similar to one expressed by John Heil, who suggests that ruling out infinite complexity is a necessary constraint for scientific theorizing; this leads him to deny actual infinities altogether.43 The approach outlined here is more liberal, as the generic sense of fundamentality that we defined in the previous section (GOF) allows for actual infinities, provided that they are of the ‘boring’ sort. Since boring infinite descent requires that there is no novelty in the structure after a certain point, it appears that there is no need for infinite complexity either. Hence, rather than infinitism per se, it may be that it is only infinite complexity that we need to rule out. There is a worry though: if it cannot be ruled out a priori that the lower-level structure involves infinite complexity, it is not clear

whether anything more can be said. Certainly, the resulting ontological picture is easier to swallow without infinite complexity, and this seems to be what motivates Heil to abandon actual infinites as well, but the evidence from physics does not settle the matter conclusively. We could attempt to motivate the idea further. Take for instance the elementary charge, which was briefly discussed above. Quarks have charges that are integer multiples of one third of the elementary charge, and if they had internal structure then this structure would have to add up to the appropriate charge as well. Now, if we continue this process indefinitely, then eventually the charges of the constituent parts would approach zero. Perhaps this oddity can be overcome, but how? At the very least, it is plausible that some as-yet-unknown forces that bind these near-chargeless parts together would have to be postulated. Since the complex sequence goes on infinitely, this would give rise to infinitely many fundamental forces. It is the dream of physicists to unify all the fundamental forces, but this may not be possible if there are infinitely many of them. Accordingly, it is not surprising that many physicists find the idea of infinite descent unattractive. Here is a recent example: In the last century, there have been repeated discoveries of underlying structure. Moving from macroscopic objects, to atoms, to components of the nuclei, to quarks, it has been demonstrated repeatedly that the differences between supposedly fundamental particles are, in fact, 43

John Heil, The Universe As We Find It (Oxford University Press, 2012), pp. 38 ff.

147

148

Fundamentality and levels of reality

merely consequences of the composite structure of underlying reality. It only seems a natural progression that such an approach of looking for underlying structure be used to explain the particles of the standard model. Attempts towards this end, dubbed preon models, met with many obstacles, but still there was something deeper that presented itself as a difficulty. The difficulty is that, as such a process does not have an end, we can continue to suppose that below the currently understood structure is another set of more fundamental particles. This idea quickly becomes unappealing at a philosophical level, or even a practical level, as the question then becomes ‘What could make it end?’44

Such speculation is certainly no more conclusive than the philosophical discussion so far, but it is interesting to note that there are physicists whose intuitions about the matter match those of many philosophers. However, these intuitions have not gone unchallenged in physics either. For instance, the physicist Howard Georgi, drawing on the issues surrounding renormalization in quantum field theory, suggests that effective quantum field theories could ‘go down to arbitrarily short distances in a kind of infinite regression’.45 However, Georgi adds that the mostprobable scenario is one in which ‘at some very short distance the laws of relativistic quantum mechanics break down and an effective quantum field theory is no longer adequate to describe the physics’.46 He does not seem to have considered, or else has rejected, the possibility that the infinite regress could be benign because of being ‘boring’, as was discussed earlier. What we might take from Georgi’s line of thought is that whatever happens when we approach ‘very short’ distances, it must be something that is able to support the laws of relativistic quantum mechanics. With some effort, this may be considered to reflect the requirements that we associated with (GXD). A further, speculative account that could perhaps accommodate infinitism in physics comes from the Nobel Prize winner Hans Dehmelt. Dehmelt proposes an intriguing quark/lepton substructure based on the model of the triton – the nucleus of hydrogen’s radioactive isotope tritium (3H):

44

Jonathan Hackett, ‘Locality and Translations in Braided Ribbon Networks,’Classical and Quantum Gravity 24 (2007), p. 5757.

45

46

Howard Georgi, ‘Effective Quantum Field Theories,’ in Paul Davies (ed.),The New Physics (Cambridge University Press, 1989), p. 456. Ibid.

6.5 Fundamentality and physics

I propose to extend the triton substructure scheme to an infinite number of layers. Below the four layers listed above [up to subquarks], they contain higher order dN subquarks, with N = 5 → ∞. In each layer the particles are not identical but resemble each other in the same way as quarks and leptons do, with masses varying as much as a factor 10 8. In an infinite regression to simpler particles of ever increasing mass, they asymptotically approach Dirac point particles.47

Up to N = 3, the level of electrons, Dehmelt’s model is motivated by current physics, but it is speculative from N = 4 onwards, where electron substructure is postulated. Importantly, it appears that the particles in each new layer – as N increases – will be held together by new, stronger and shorter-range forces (compare with the strong nuclear force, which easily overpowers the electromagnetic repulsion in the nuclei of atoms). As N approaches infinity, we must also postulate infinitely many, arbitrarily

stronger and shorter-range forces. While this may be relatively unproblematic mathematically, the ontological costs may be problematic. Notice also that on the face of it, the model would appear to involve mereological levels all the way down, although Dehmelt does not discuss it in these terms. However, it is open to interpretation what exactly happens at the point when we reach N = 5 →

. One option would be to inter-

∞

pret the mereological chain as terminating at this point and the further chain being ‘boring’ (that is, repetitive) and satisfying (GXD) in the sense discussed earlier. What about the ontological cost of introducing new forces at each level? The worry here is that if ‘boring’ infinite descent requires an infinite number of novel forces, its plausibility as a candidate for something that could satisfy (GOF) may diminish. In fact, Dehmelt’s model would also appear to require the introduction of new exchange particles at each level, which is perhaps even more problematic for (GOF). Consider the strong nuclear force again: it binds protons and neutrons together in the nucleus, overpowering the electromagnetic repulsion between protons. Like all the fundamental forces, the strong nuclear force is an exchange force, that is, it involves the exchange of one or more particles. At the level of protons and

47

Hans Dehmelt, ‘Triton,… Electron,… Cosmon,…: An Infinite Regression?’, Proceedings of the National Academy of Sciences 86 (1989), p. 8618. See also Tahko, ‘Boring Infinite Descent’ and Morganti, ‘Metaphysical Infinitism and the Regress of Being.’

149

150

Fundamentality and levels of reality

neutrons, the strong nuclear force relies on the exchange of mesons, but it is merely a residual force of the strong force (that is, the colour force) that binds together positively charged quarks. In the latter case, the relevant exchange particles are gluons. A full examination of the strong force would require delving into quantum chromodynamics, but all we are interested in here is the apparent need for independent exchange particle(s) at each level where the force is active (e.g., mesons and gluons). The core of the issue is this: if any physical model with infinite descent requires postulating infinitely many new forces and corresponding exchange particles, can such a model ever be well-founded? Ultimately, that depends on our understanding of a viable fundamental level given the complications of contemporary physics – which is something that has not been examined in much detail so far. So as should be obvious from the preceding discussion, research on these issues is still very young, and also rather challenging because of the limited empirical data. But one thing is clear: the search for a fundamental level will no doubt remain a central puzzle in metametaphysics and physics alike for the foreseeable future.

7

The epistemology of metaphysics: a priori or a posteriori?

Throughout this book, we have encountered various issues that fall within the purview of epistemology. We have seen that the substance, significance, and progress of many a debate depends on the success of our attempts to gain knowledge about the subject matter of that debate. In the two preceding chapters, we have acquired various further tools to express more precisely what it is that we are trying to gain knowledge about, but these tools are of limited help when it comes to the epistemic side. Indeed, the discussion has quickly turned to matters so abstract that it is difficult to even evaluate the relative merits of different views. So it is high time to address the epistemic issues that loom large over the metametaphysics literature. These issues are not commonly considered to be a part of metametaphysics per se, but rather just a part of first order metaphysics or epistemology. The present author holds a different opinion: in many cases the epistemic part of a debate in metametaphysics is so intimately connected with the ontological part that it is simply impossible to distinguish between metametaphysics ‘proper’ and epistemology. Hence, we cannot avoid delving into epistemic issues when we pursue questions in metametaphysics; they are a central part of the methodology of metaphysics. One of the central questions here is whether metaphysical knowledge is gained by a priori or a posteriori means, or by some combination thereof. If a posteriori elements are involved, then the obvious question that follows is to what extent are these elements due to science; that is, what is the relationship between science and metaphysics? However, this question is so broad and important that we will dedicate a chapter of its own to it, namely Chapter 9. But the a priori vs. a posteriori distinction itself ought to be discussed briefly. We will do so in the first section of this chapter. In no other area is the epistemic issue more pressing than in questions that deal with modal truths. We have already seen plenty of examples 151

152

The epistemology of metaphysics: a priori or a posteriori?

regarding modal truths in metametaphysics, such as the speculative discussion regarding the (physical or metaphysical) necessity of a fundamental level in Chapter 6. The second, third, and fourth sections of the present chapter focus on examining the potential sources of this type of modal knowledge; that is, we focus on modal epistemology. Of course, modal knowledge does not exhaust metaphysical knowledge, but it is a particularly important and illuminating case and hence serves as an appropriate case study. We can divide the accounts of the source of modal knowledge into two rough categories: modal rationalism and modal empiricism. As the names suggest, the first is a group of views that attempts to explain modal knowledge primarily in terms of a priori resources, whereas the latter emphasizes a posteriori resources. A further complication involves the relationship between essence and modality, as one currently popular account takes essence to be ontologically prior to – or reducible to – modality. The question arises: does the ontological priority of essence entail epistemic priority as well? We will discuss this issue in the third section. As we will see, distinguishing the a priori and the a posteriori also causes complications, for it is quite difficult to point out examples of ‘pure’ a priori or a posteriori knowledge. In fact, much of the fourth section of this chapter

will be spent in trying to clarify what the status of ‘armchair’ methods in metaphysics is. The fifth section takes stock of the preceding debate between modal rationalism and modal empiricism. In the course of our discussion, we will see that there is strong pressure towards a type of hybrid view: if the distinction between a priori and a posteriori knowledge is not as straightforward as is often assumed, then we’ll have to consider alternative ways of formulating the debate. On the positive side, this may prove to be a methodologically interesting development, especially if it allows us to make progress with regard to the primary issue as well, namely modal epistemology.

7.1 A priori vs. a posteriori A priori and a posteriori knowledge are typically defined as knowledge

acquired by non-experiential and experiential means, respectively. This distinction is known to be problematic, but since philosophers will no doubt continue to use it (and we will do so as well), we ought to do our best to clarify it. Consider two rather obvious problems, discussed for instance

7.1 A priori vs. a posteriori

by Laurence BonJour: the problem of how to define ‘experience’ and how the a priori is supposed to be ‘independent’ of it.1 In the first case the problem is how to determine the correct scope of experience. Do mental processes count as experience? How about mathematical or philosophical reasoning that relies on certain learned patterns? Should only perceptual information count as experiential? How does memory fit in with all of this? The second problem involves issues concerning concept acquisition as a precondition for a priori knowledge, but also independence of experience, namely whether a proposition is indefeasible by experiential information.2 The basic underlying problem here is the following. On the one hand, if we are too strict in how we define independence of experience, then nothing whatsoever will count as a priori knowledge and hence the distinction breaks down. On the other hand, if we are more liberal about what counts as non-experiential, then the distinction becomes vague and we might have cases where it is equally correct to say that some knowledge is a priori or that it is a posteriori. However, as we will see in what follows, the labels of ‘a priori’ and ‘a posteriori’ are currently used in such a broad fashion that we are effectively already in a situation where the distinction has become rather vague. This is not a reason for giving the distinction up altogether though, for we can still identify certain characteristics that point towards apriority or aposteriority and hence, at least arguably, we have some use for the distinction. Note that besides knowledge, we talk about justification and reasoning . There is an extensive literature on a priori justification, which we will not engage with here.3 Justification concerns the process of how a person’s beliefs are justified. The same belief could be justified by a posteriori means

for one person and by a priori means for another. Moreover, a belief might not be true even if it is justified. But not all issues concerning justification are directly relevant for us here, as we are primarily interested in how

1

2

3

Laurence BonJour, In Defense of Pure Reason (Cambridge University Press, 1998), pp. 7–11. For further discussion regarding these and other problems, see Tuomas E. ahko, T ‘A New Definition of A Priori Knowledge: In Search of a Modal Basis,’Metaphysica 9.2 (2008), pp. 57–68, and Tuomas E. Tahko, A‘ Priori and A Posteriori: A Bootstrapping Relationship,’ Metaphysica 12.2 (2011), pp. 151–164. For one influential account, see Albert Casullo, A Priori Justification (Oxford University Press, 2003).

153

154

The epistemology of metaphysics: a priori or a posteriori?

it is possible to distinguish between apriority and aposteriority in the abstract. Regarding reasoning, one might ask whether it makes sense to talk about anything except a priori reasoning, since reasoning is presumably a (deductive) mental process. Hence, if one arrives at a belief purely by reasoning, then the belief would appear to be justified by a priori means. But perhaps it is not altogether impossible to conceive of something called ‘ a posteriori reasoning’: generalizing from empirical observations by inductive inference would seem to be a form of ‘reasoning’ distinct from the deductive process we typically refer to by ‘reasoning’. So empirical generalization of this type could perhaps be called justification by ‘ a posteriori reasoning’. In any event, these are not issues that we have to set-

tle here, but the reader is urged to keep in mind that there are different ways to approach the topic. Given that the distinction between a priori and a posteriori knowledge / justification / reasoning / inquiry / methods, and so on does not appear to be particularly sharp, we will resort to stipulation. Specifically, we will stipulate that inquiry or justification involving deductive reasoning is a characteristic of apriority, whereas the involvement of empirical elements is, naturally enough, a characteristic of aposteriority. Sometimes it is not perfectly clear whether some process of inquiry does involve the mentioned characteristics, but this is a question of a different level. What is more important is that very often, if not always, a process of inquiry will involve both characteristics. Typically, when any characteristics of aposteriority are present, philosophers are inclined to call the resulting piece of knowledge a posteriori (e.g., the case of a posteriori necessities). But we have already noted that this is likely to be too strong a condition, for absolutely nothing might be totally ‘unpolluted’ by a posteriori characteristics, and hence the a priori vs. a posteriori distinction would simply break down altogether. Some might welcome this result, but it should be a last resort, given that the notion of apriority features heavily in metaphysics. Thus, in such cases it may be better to conclude that a priori and a posteriori elements are intertwined to such an extent that we simply can’t classify such a piece of knowledge one way or another. Yet, this does not mean that we couldn’t talk about the different characteristics of apriority and aposteriority that are present in a given case. If one aspect dominates, say, a given process of justification, then we might say that the resulting belief was justified primarily by a priori means. But since we have acknowledged that

7.2 Modal rationalism and a priori methods

the distinction is unlikely to be sharp, then not that much hangs on this issue in the first place. In sum, the reader should be aware that the approach taken in this book holds there to be a place for the notion of apriority, even if it is sometimes very difficult to determine whether a process or a proposition is a posteriori, a priori , or both. Given this, some mention should be made of the possibility of a more deflationary approach: one might think that no useful purpose is being served by the a priori vs. a posteriori distinction. For instance, Timothy Williamson has recently argued that since experience plays a role that is more than purely enabling but less than strictly evidential both in clear cases of a priori knowledge (or justification) and in clear cases of a posteriori knowledge, the distinction really does not have much theoretical significance.4 Often, it is the notion of apriority which comes out as the underdog when such challenges are posed, probably because empirical elements have such a strong standing in contemporary philosophy. Many of these challenges derive from logical empiricism, Quine’s denial of the analytic/synthetic distinction, or more recent radical empiricism. It is not clear that these challenges capture the subtlety of the intertwined conception of apriority and aposteriority that we have stipulated above. But there have been many recent challenges concerning the coherence or significance of the notion of apriority, so they should be taken seriously. Albert Casullo has recently made an effort to address some of these challenges, although he is also concerned about the problems regarding the experiential/non-experiential distinction.5 However, discussing these issues in more detail here would take us too far from our primary topic, even though ultimately any account of the epistemology of metaphysics that resorts to these notions will have to face the difficult questions regarding the a priori vs. a posteriori distinction that we have outlined here.

7.2 Modal rationalism and a priori methods Let’s start by distinguishing two approaches: 4

5

Timothy Williamson, ‘How Deep is the Distinction Between A Priori and A Posteriori Knowledge?’ in A. Casullo and J. C. Thurow (eds.), The A Priori in Philosophy (Oxford University Press, 2013). Albert Casullo, ‘Four Challenges to the A Priori–A Posteriori Distinction,’ Synthese (2013 [online], forthcoming).

155

156

The epistemology of metaphysics: a priori or a posteriori?

Permissive modal rationalism (PMR) suggests that our epistemic access to modality generally involves rationalist, a priori elements. Strict modal rationalism (SMR) suggests that modal epistemology is entirely a matter of a priori reasoning. 6

Modal rationalism encompasses intuition- and conceivability-based approaches of the type defended, for instance, by George Bealer 7 and David Chalmers8, and also the essence-based account of E. J. Lowe.9 Bealer defends an intuition-based account, whereas Chalmers defends a conceivability-based account as a part of a broader rationalist picture. Lowe’s approach is slightly different, since he thinks, following Kit Fine, that metaphysical modality is grounded in essence – we will return to this idea in a moment.10 An upshot of this view is that the epistemology of modality becomes the epistemology of essence, as discussed in more detail in the third section of this chapter. According to Lowe, the epistemology of essence is an a priori process of understanding what an entity is (or would be). In fact, Bealer also talks about ‘understanding’, but his usage differs from Lowe’s: Bealer focuses on the possession of concepts whereas Lowe seems to have a more general type of understanding in mind, albeit he does not specify it very carefully.11 The literature on modal epistemology, however, is enormous, and these are just a few prominent examples. In fact, because of the size of the literature, it would require a series of books to fully engage with the extant views. Thus, we will instead just take a small but hopefully illuminating peek into the debate, with special attention to the

6

7

8

9

10 11

For an excellent selection of articles dealing with conceivability in particular, see T. S. Gendler and J. Hawthorne (eds.),Conceivability and Possibility (Oxford University Press, 2002). See for instance George Bealer, ‘Modal Epistemology and the Rationalist Renaissance,’ in Gendler and Hawthorne (eds.), Conceivability and Possibility, pp. 71–125; and his ‘The Origins of Modal Error,’Dialectica 58.1 (2004), pp. 11–42. David Chalmers, ‘Does Conceivability Entail Possibility?’, in Gendler, and Hawthorne (eds.), Conceivability and Possibility, pp. 145–200. E. J. Lowe, ‘Whatis the Source ofour Knowledge of Modal Truths?’,Mind 121 (2012), pp. 919–950. Kit Fine, ‘Essence and Modality,’ Philosophical Perspectives 8 (1994), pp. 1–16. For a more systematic theory of understanding and essence, see Anand Vaidya, ‘Understanding and Essence,’ Philosophia 38 (2010), pp. 811–833. Vaidya’s account, however, may not be compatible with the type of (realist) essentialism that Lowe hopes to defend, so it represents a somewhat different approach to the topic.

7.2 Modal rationalism and a priori methods

methodological consequences.12 We’ll start from the conceivability-based account and continue with the essence-based account, taking into account how they relate to (PMR) and (SMR). The idea that conceivability might be a guide to possibility has a long history, but we will not dwell on its Cartesian roots here. Rather, we will focus on one particular issue concerning the contemporary usage of conceivability. But first, the claim obviously needs qualification: what kind of conceivability and what kind of possibility are we talking about? Typically, we are interested in whether conceivability might entail metaphysical possibility, whereas it is taken as granted that conceivability entails at least conceptual possibility. In other words, if you can conceive of a scenario x, then x

must at least be a scenario that is not conceptually contradictory. Hence, in its simplest form, conceivability refers to our ability to coherently imagine or conceive the scenario in question. Note, though, that strictly speaking we should perhaps distinguish conceivability and imaginability, even though most authors use them interchangeably.13 There are various reasons for this, one of them being the fact that ‘imaginability’ typically has a very strong connection to our ability to visualize something, so it is quite tightly associated with just one sensory faculty. Conceivability, however, is typically understood in a broader fashion, not so tightly connected with vision. Sometimes further requirements are added as well, for instance that we should be able to imagine or conceive the relevant scenario as if it were actual, that is, to imagine an alternative history or a metaphysically pos-

sible world that could have been actual. The theory of two-dimensional modal semantics develops this line of thought, but this is a topic that we will omit

here, since a detailed presentation would take us quite far from our main topic.14 Given our restricted focus, we do not need to discuss all the intricacies of the various modal rationalist accounts. Rather, we might concentrate on those particularly problematic cases where seemingly metaphysically 12

13

14

For a selection of state-of-the-art articles on the topic, see R. W. Fischer and F. Leon (eds.), Modal Epistemology After Rationalism, Synthese Library (Dordrecht: Springer, forthcoming). For an argument to the effect that conceivability and imaginability should be distinguished, see Marcello Oreste Fiocco, ‘Conceivability, Imagination, and Modal Knowledge,’ Philosophy and Phenomenological Research, 74.2 (2007), pp. 364–380. For an overview of two-dimensional modal semantics, see David Chalmers, ‘The Foundations of Two-Dimensional Semantics,’ in M. Garcia-Carpintero and J. Macia

157

158

The epistemology of metaphysics: a priori or a posteriori?

necessary truths such as ‘Gold is the element with atomic number 79’ are to be reconciled with conceivable yet metaphysically impossible scenarios such as ‘Gold might have turned out not to be the element with atomic number 79’. One reaction to this issue is to say that it is always conceivable that things might have been otherwise, while another would be to insist that conceivability is restricted by the current a posteriori framework. A proponent of (SMR) might favour the first option, whereas (PMR) is more naturally combined with the restricted use. Plausibly, we can distinguish between different types of conceivability here, as Chalmers and Stephen Yablo, among others, have done.15 However, no generally accepted convention about the use of conceivability in such cases exists despite the vast literature, and in fact it may even be possible to be a strong modal rationalist and accept that conceivability is always restricted – this would seem to be a coherent view if it is emphasized that we cannot simply step out of our current epistemic situation and so must always reason within certain ‘a posteriori bounds’. To put the distinction between the two possible reactions

more precisely, the first issue is whether we should fix conceivability in terms of how the world might be before we have any a posteriori knowledge, limited only by a priori consideration, or how the world might be given the a posteriori framework. Here we can immediately see that the relationship

between a priori and a posteriori elements comes into question. On the basis of the above discussion, a further possibility of a dismissive reaction emerges: if there is any sense in which a posteriori elements are required to get the modal rationalist view off the ground then it would not seem to be rationalism as it is traditionally understood. This type of dismissive reaction would lead to either (SMR) or a flat denial of modal rationalism. However, this may not be an entirely fair reaction, for we should

15

(eds.), Two-Dimensional Semantics: Foundations and Applications (Oxford University Press, 2006), pp. 55–140. Chalmers, ‘Does Conceivability Entail Possibility?’, and Stephen Yablo, ‘Is Conceivability a Guide to Possibility?’, Philosophy and Phenomenological Research, 53 (1993), pp. 1–42. Chalmers thinks that there might be up to eight types of conceivability; it is his distinction between primary and secondary conceivability that reflects the issue at hand: S is primarily conceivable when it i s conceivable that S is actually the case, and it is secondarily conceivable when S conceivably might have been the case (p. 157). We will set aside the complications regarding ideal conceivability, which Chalmers also discusses, i.e., the case of an ideal conceiver with perfect rational capabilities.

7.2 Modal rationalism and a priori methods

presumably allow for at least some a posteriori elements without having to conclude that the epistemic framework we are dealing with is empiricism. For instance, we clearly learn about gold – that gold is the element with atomic number 79 – through empirical means, but even once we know this, the necessity of the proposition ‘Gold is the element with atomic number 79’ may only be derived by a priori means, as the modal rationalist would have it. Similarly, although we learn about the concept ‘element’ via our senses, a posteriori, this type of concept acquisition may be considered a part of our background knowledge rather than speaking against (SMR). So a charitable reading would seem to be that both (PMR) and (SMR) are valid candidates and the question will hang on something other than whether there is any empirical exposure whatsoever. But if we understand conceivability in this charitable, more restricted sense, taking into account the a posteriori framework, then it seems difficult to reconcile it with conceivable yet metaphysically impossible scenarios. Given our knowledge about, say, gold, namely that it is the element with atomic number 79, it ought to be impossible even to imagine gold not being an element with that atomic number without making radical changes in the a posteriori framework. So this might push us back towards understanding conceivability in the unrestricted, a priori sense, as then it appears that conceivability is a necessary requirement for (knowledge of) possibility. This is because conceivability could then be interpreted simply as an initial requirement for a lack of contradictions in the imagined possibility. Yet, while this may save a sense of modal rationalism, it does appear to be a weaker thesis than we were initially looking for. Perhaps, in addition to the requirement of a lack of contradictions in conceivable scenarios, we could give a corresponding negative requirement for conceivability? This could be something like the following: everything not ruled out by a priori reasoning is conceivable. On the face of it, this does seem to help at least a little, as plausibly a priori considerations rule out contradictory propositions, so this qualification for conceivability encompasses the previous requirement as well. The focus of the question then moves to the epistemic: if conceivability is a prima facie guide to possibility in this sense, could it be a useful tool in modal epistemology? Of course, given that there is an overlap between what is conceivable and what is metaphysically possible, but also between what is conceivable and what is metaphysically impossible, the previous overlap may not entail a

159

160

The epistemology of metaphysics: a priori or a posteriori?

reliable epistemic link between conceivability and metaphysical possibility.

In fact, there are infinitely many conceivable metaphysically possible scenarios (albeit that in the case of non-ideal conceivers such as us mere mortals, this is subject to limitations in cognitive capacity). But there are also infinitely many conceivable metaphysically impossible scenarios! To see this, take any metaphysically necessary truth, such as ’Gold is the element with atomic number 79’, and start generating metaphysically impossible scenarios where gold is the element with atomic number 80, or 81, or 82, or … If we hope to use conceivability as a reliable guide to metaphysical possibility, we should somehow be able to distinguish between metaphysically possible and metaphysically impossible scenarios and hence restrict our focus somehow. One of the most influential attempts to do this is due to Williamson, who discusses conceivability as a part of his counterfactual-based account of modal epistemology. 16 Note that Williamson attempts to explain our knowledge of modality in terms of our knowledge of counterfactuals and wishes to stay neutral about the epistemic status of our modal judgements, so we are not suggesting that his view is necessarily a version of modal rationalism. Nevertheless, he does have some interesting things to say about the issue at hand. Figure 7.1 illustrates what the problem is: Inconceivable

Conceivable and metaphysically impossible

Conceivable and metaphysically possible

Figure 7.1

16

Timothy Williamson, The Philosophy of Philosophy, (2007, Oxford: Blackwell Publishing), especially Ch. 5.

7.2 Modal rationalism and a priori methods

Everything outside the circle is inconceivable and everything inside it is conceivable – this is a relatively unproblematic distinction. But the problematic distinction is inside the circle, between what is conceivable but metaphysically impossible and what is conceivable and metaphysically possible. How are we able to divide the circle?

On Williamson’s line, the key is that we need to take advantage of our background knowledge. In other words, Williamson thinks that our ability to conceive of counterfactual scenarios, such as gold failing to be the element with atomic number 79, must be restricted by at least some empirical knowledge: If we know enough chemistry, our counterfactual development of the supposition that gold is the element with atomic number 79 will generate a contradiction. The reason is not simply that we know that gold is the element with atomic number 79, for we can and must vary some items of our knowledge under counterfactual suppositions. Rather, part of the general way we develop counterfactual suppositions is to hold such constitutive facts fixed.17

This would already seem to concede that modal rationalism needs to be supplemented with some empirical knowledge. But if most of the work in modal epistemology is done with the help of our ability to entertain counterfactual suppositions – that is, our ability to conceive of alternative scenarios – then perhaps this would still be acceptable to some modal rationalists. Williamson at least provides an answer to the problem at hand, as we can dismiss the conceivable but metaphysically impossible scenario of gold having some other atomic number than 79 with the help of ‘constitutive facts’ – background knowledge that we do not vary when considering counterfactual scenarios. The manner in which the method reveals that we must hold constitutive facts fixed is based on the idea that any counterfactual supposition which fails to hold such a fact fixed will generate a contradiction. This enables us to distinguish between conceivable metaphysical impossibilities and metaphysical possibilities. Promising as it is, one obvious problem with this proposal is that we should somehow be able to know which facts are constitutive. Cases such as ‘Gold is the element with atomic number 79’ may be rather good candidates for this analysis, as there are no obvious alternatives for what gold 17

Williamson, The Philosophy of Philosophy, p. 164.

161

162

The epistemology of metaphysics: a priori or a posteriori?

could be, but unless we have a more general story about the class of such constitutive facts, then we have not yet discovered a reliable method. In fact, Williamson does not even attempt to come up with such a story, as he believes that we do not need to know which facts are constitutive; it suffices that we know some constitutive facts.18 To see that it is important for Williamson’s method that we hold the correct facts fixed, consider what would happen if we held some non-constitutive facts fixed. What would happen is that we would end up ruling out (as metaphysically impossible) certain possibilities that we wish to include: say, if we held it fixed that Earth is the third planet from the Sun, we would end up ruling out the possibility that the planet Venus never formed, which is surely possible! Similarly, if we fail to hold some constitutive fact as fixed, say that elements are defined by their atomic number (if that is indeed the case), then we would erroneously include metaphysical impossibilities, such as gold failing to be the element with atomic number 79. So it does seem that something more needs to be said about constitutive facts. However, attempts to come up with an analysis of constitutive facts have been made; one candidate, not surprisingly, is that constitutive facts of the relevant type concern essences. If Williamson’s account were to be supplemented with a satisfactory analysis of constitutive facts in terms of essences, then we would seem to have a fairly strong case for modal rationalism. But note that we would also lose part of Williamson’s srcinal motivation, namely the possibility of a uniform account– an account that explains our knowledge of modal truths in terms of a single rational faculty, namely the ability to handle counterfactuals. So if essences are needed to undergird the account, then why not go all out for an essence-based approach? According to the essence-based approach, which was popularized by Fine, essentialist facts can be understood as non-modal constitutive facts (which entail modal facts).19 Now, if we have some prior epistemic access to these constitutive, essentialist facts, then we can address the previous concern. For instance, if we consider atomic number to be essential for elements, then the atomic number of gold being 79 is a constitutive fact that we know independently of Williamson’s counterfactual analysis, hence enabling adistinct 18

19

For further discussion, see Sonia Roca-Royes, ‘Conceivability and De Re Modal Knowledge,’ Noûs 45.1 (2011), pp. 22–49, and Tuomas E. Tahko, ‘Counterfactuals and Modal Epistemology,’Grazer Philosophische Studien 86 (2012), pp. 93–115. Fine, ‘Essence andModality.’

7.3 The epistemology of essence

source of constitutive facts that could serveto restrict conceivability as well. Yet it should be immediately noted that this may only push the problem one iteration further, for now we ought to present an account of the epistemology of essence.

7.3 The epistemology of essence What we are now interested in is the process of coming to know modal facts via our knowledge of essentialist facts. Fine says less about the epistemic dimension than the ontological one, but his account of the ontological relationship between essence and modality has been hugely influential. On the 20 epistemic side, Lowe has done much of the pioneering work. Both Lowe’s

and Fine’s understandings of essence follow a ‘neo-Aristotelian’ line whereby the essence of an entity is its real definition. This idea is not easily elaborated on, but one way to understand real definitions is to take them to be expressed by propositions which tell us what a given entity is orwould be – we can also state 21 the real definitions of things that are non-existent. Lowe considers essence

to be prior to existence both ontologically and epistemically. In Lowe’s own words: ‘instead of trying to explicate the notion of essence in terms of that of modality, as on the Kripkean account of essence, the very reverse needs to be done’.22 Note that unless we assume that essence can be understood in non-modal terms, the area of the epistemology of essence would seem to collapse into modal epistemology proper, as the essentialist truths would simply be a proper subset of the modal truths. But if essence can be understood in non-modal terms, then the question about our epistemic access to essentialist truths is very pressing indeed. Note further that Lowe holds that essences cannot be entities in themselves, on pain of infinite regress. For if every entity had an essence (as Lowe also holds) and essences were entities, then essences ad infinitum. themselves would have to have essences as well, and so on 20

21 22

For further discussion, see Tuomas E. Tahko, ‘Empirically-Informed Modal Rationalism,’ in Fischer and Leon (eds.), Modal Epistemology After Rationalism, Synthese Library (Dordrecht: Springer, forthcoming). See Lowe, ‘What is the Source ofour Knowledge of Modal Truths?’, p.935. E. J. Lowe, ‘Essence vs. Intuition: An Unequal Contest,’ in A. R. Booth and D. P. Rowbottom (eds.), Intuitions (Oxford University Press, 2014), p. 264. See also E. J. Lowe, ‘Two Notions of Being: Entity and Essence,’ Royal Institute of Philosophy Supplements 83.62 (2008), pp. 23–48. In the Kripkean account of essence, essential properties are defined simply as necessary properties.

163

164

The epistemology of metaphysics: a priori or a posteriori?

How do we come to know the propositions that express essences – the real definitions – according to Lowe? The core of Lowe’s suggestion is that to grasp an entity’s essence is simply to understand what the entity is: To know something’s essence is not to be acquainted with some further thing of a special kind, but simply to understand what exactly that thing is. This, indeed, is why knowledge of essence is possible, for it is a product simply of understanding – not of empirical observation, much less of some mysterious kind of quasiperceptual acquaintance with esoteric entities of any sort. And, on pain of incoherence, we cannot deny that we understand what at least some things are, and thereby know their essences. 23

Even though this passage may not convince everyone about the possibility of our knowledge of essences, it does at least make clear what Lowe’s take on the issue is, namely that the epistemology of essence is a simple a priori process of understanding what an entity is (or would be). On Lowe’s view, this is the basis of all modal knowledge, including a posteriori necessities. To see how essence and metaphysical modality come apart on this view, consider the difference between two geometrical examples, as discussed by Lowe: (E1) An ellipse is the locus of a point moving continuously in a plane in such a fashion that the sum of the distances between it and two other fixed points remains constant […] (E2) An ellipse is the closed curve of intersection between a cone and a plane cutting it at an oblique angle to its axis greater than that of the cone’s side.24

Lowe suggests that (E1) gives us the generating principle of ellipses – and hence their essence – whereas (E2) states merely a necessary property of ellipses. We can understand properties such as (E2) once we know the generating principle of ellipses, but Lowe insists that (E1) is not contained in (E2); we need an understanding of the essences of cones as well as ellipses before we can produce (E2). This is how essence and

23 24

Lowe, ‘Two Notions of Being: Entity and Essence,’p. 39. Lowe, ‘What is the Source of ourKnowledge of Modal Truths?’, p.936.

7.3 The epistemology of essence

metaphysical necessity come apart: we may state numerous necessary properties that ellipses have, but only something like (E1) will give us their essence. It would be tempting to conclude that all metaphysically necessary truths are true in virtue of the essence of something or other, but Lowe suggests instead that ‘any essential truth is ipso facto a metaphysically necessary truth, although not vice versa: there can be metaphysically necessary truths that are not essential truths’. 25 However, the view needs

to be qualified further, as this is not to say that (E2) would not also hold in virtue of the essence of something – it just does not hold strictly in virtue of the essence of ellipses. Lowe thinks that (E2) holds in virtue

of the

essences of ellipses and cones together . So following this line of thought, all metaphysical necessities are grounded in essence, even though a given metaphysically necessary truth may not be an essential truth about any one thing. The structure of the epistemic process, then, is as follows: (1) We know that the essence of x is expressed by proposition p (which states the real definition of x). (2) We know that if p expresses the essence of x, then p is metaphysically necessary. (3) We know that p is metaphysically necessary. (2) is trivially true on the Finean analysis of essence, so it is (1) that does the real work here. To put all this to together, consider the following example, which features one of the tools we also learned about in Chapter 5, namely essential dependence:

Consider the following thing, for instance: the set of planets whose orbits lie within that of Jupiter. What kind of thing is that? Well, of course, it is a set, and as such an abstract entity that depends essentially for its existence and identity on the things that are its members – namely Mercury, Venus, Earth, and Mars. Part of what it is to be a set is to be something that depends in these ways upon certain other things – the things that are its members. Someone who did not grasp that fact would not understand what a set is.26

25 26

Ibid., p. 938. Lowe, ‘Two Notions of Being: Entity and Essence,’p. 37.

165

166

The epistemology of metaphysics: a priori or a posteriori?

So Lowe thinks that in many cases knowing how a thing is related to other things is central to our knowledge of what a thing is. But note that Lowe thinks that for the set to depend essentially for its and identity on its members is only a

existence

part of what it is to be a set. In

general, knowing the complete essence of a thing may not be a very simple affair – perhaps it is almost impossible – but knowing a part of a thing’s essence is often sufficient for talking or thinking about it comprehendingly, and for being able to distinguish it from other things. This is what seems to be the basis of Lowe’s idea of

understanding .

While Lowe’s line of thought seems to hold some promise, we have hardly been provided with a full epistemology of essence so far. One complication emerges immediately when we try to apply this account to concrete objects: just how much of a thing’s essence do we need to grasp – to understand – in order to have a sufficient picture of its existence and identity conditions? In particular, is it plausible that this

will be an a

priori process rather than a primarily empirical one in the case of things

such as gold and other natural kinds? Lowe’s picture remains incomplete in this regard. Because of these difficulties, a more straightforward solution may be simply to abandon strict modal rationalism (SMR), as it was defined in the previous section. If all or part of our modal knowledge is acquired by some less powerful means than the conceivabilityand essence-based accounts would have it, then perhaps the risk of over-generating possibilities and hence running the risk of including metaphysically impossible scenarios can be avoided altogether

– this

might still be compatible with some versions of permissive modal rationalism. Moreover, once empirical information is allowed in the picture, it may be easier to account for modal facts concerning concrete objects. In any case, there is an obvious motivation to examine alternative epistemic strategies in more detail; the recent interest in ‘modal empiricism’ is symptomatic of the difficulties that we’ve encountered with modal rationalism. 27

27

Note however that attempts to develop different versions of the essence-based account (although not necessarily modal rationalist in the strong sense) continue. For a very recent attempt, see Bob Hale, Necessary Beings: An Essay on Ontology, Modality, and the Relations Between Them (Oxford University Press, 2013), Ch. 11.

7.4 Modal empiricism and the status of armchair methods

7.4 Modal empiricism and the status of armchair methods The term ‘modal empiricism’ is not yet established in the literature, but since we are looking into an approach that contrasts with modal rationalism, calling it modal empiricism would seem to be a natural28choice; the notion has previously been used for instance by Carrie Jenkins. Roughly speaking, modal empiricism encompasses views that suggest our modal knowledge derives primarily from experience, or, to put it somewhat differently, that experience is what ensures the reliability of our modal knowledge. The latter is close to what Jenkins suggests:experience provides an epistemic grounding for our concepts, the basis of our conceptual abilities. Hence, somewhat surprisingly, the reliability of conceivability itself is based on experiential knowledge! Williamson’s account, one aspect of which we discussed above, has a similar starting point. It starts from the evolutionary basis of our ability to entertain counterfactual suppositions and proceeds to assess the reliability of conceivability on this basis. More precisely, Williamson reduces our capacity for modal knowledge to our capacity for assessing counterfactual conditionals, which he claims to have an evolutionary basis; it has been extremelyuseful for us as a species to be able to entertain counterfactual suppositions (and to do so successfully). For instance, if a tiger threatens you, your survival may depend on your ability to predict the tiger’s behaviour on the basis of itspossible behaviour. This leads Williamson to conclude that the worries regarding conceivability and especially the over-generation of possibilities that we discussed in the previous section fail to acknowledge the evident successes of conceivability in more mundane scenarios: ‘Once we recall its fallible but vital role in evaluating counterfactual conditionals, we should be more open to the idea 29 that it plays such a role in evaluating claims of possibility and necessity’. Thus, it seems that Jenkins’s and Williamson’s accounts– and many others as well – have elements both of modal rationalism and modal empiricism. One reason for this peculiar situation – the fact that many recent accounts attempt to develop an empiricist basis for conceivability or other 28

29

C. S. Jenkins, ‘Concepts, Experience and Modal Knowledge,’ Philosophical Perspectives 24 (2010), pp. 255–279. Williamson, The Philosophy of Philosophy, p. 163.

167

168

The epistemology of metaphysics: a priori or a posteriori?

rational capabilities – is that, historically and still today, there is an exceptionally strong consensus that experience can only be a guide to what is actual. So it seems only natural that experience has to be supplemented

with something if it is to give us access to matters relating not only to the actual, but also to the merely possible or the necessary. This situation also tells us something important about the methodological issues surrounding modal epistemology. Initially, we hope to explain a given phenomenon, such as our apparent ability to grasp modal truths, by resorting to a single explanation – we attempt to develop a uniform account. But if it turns out that such an account faces insurmountable difficulties, as strict modal rationalism perhaps does, the first reaction is not to drop uniformity but to attempt a reduction of the srcinal account. This is not to suggest that the accounts we have mentioned in this section, namely Jenkins’s and Williamson’s, have necessarily developed in this manner. The debate concerning modal epistemology as a whole, in any case, certainly shows signs of such a development, for attempts to reduce our ability to conceive and imagine into other, supposedly less mysterious sources of knowledge are numerous. The driving idea is that, other things being equal, a uniform account of modal epistemology is preferable to a fragmented account. We arrive at a somewhat problematic result: we’ve attempted to distinguish between modal rationalism and modal empiricism roughly in terms of the a priori vs. a posteriori distinction. But in both cases the situation is certainly not as simple as the a priori vs. a posteriori distinction (supposedly) is. An additional caveat, which we have not yet discussed, is the fact that if we were to defend a uniform account of modal epistemology, then all our modal knowledge would have to be explained by resorting to just one of these areas of knowledge. On the face it this seems implausible, because at the very least our knowledge of modal facts concerning abstract objects would appear to require some knowledge that could plau-

sibly be labelled ‘ a priori’, whereas knowledge about modal facts concerning concrete objects plausibly requires empirical knowledge. But perhaps some alternative understanding of the a priori vs. a posteriori distinction would enable us to make progress? For instance, Jenkins suggests that although empirical knowledge is crucial in our efforts to gain knowledge of the mind-independent world, the crucial elements of that knowledge retain ‘many characteristic features of the a priori: it is knowledge of conceptual truths through conceptual examination alone; no tests need to be

7.4 Modal empiricism and the status of armchair methods

conducted beyond the examination of one’s concepts, so it can be secured in the armchair (by people whose concepts have previously been suitably grounded) rather than the laboratory’.30 So Jenkins, like Williamson, is inclined to play down the a priori vs. a posteriori distinction, as it appears that in many cases the process of inquiry that leads to modal knowledge involves elements that could readily be called either a posteriori or a priori. This is certainly a valuable methodological lesson. Before we examine the consequences of this lesson, let us pause for a moment, since here we seem to be at the core of an important methodological issue. Here’s one way to put the puzzle. We philosophers primarily engage in ‘armchair’ philosophy, just as Jenkins suggests above. 31 We read philosophical texts and arguments, think about them in our armchairs, and report our findings in journals and at conferences. So, on the face of it, our philosophical work seems to lean strongly towards the a priori. Yet, at the same time we believe that we can make progress in philosophy and make interesting (and true!) observations about the world . These include but are of course not limited to the judgements we make about modality. Perhaps particularly troubling in this regard are our judgements about natural kinds; for instance, the famous judgement that gold is, necessarily, the element with atomic number 79, or that water is necessarily H2O. These judgements are philosophically of the utmost importance, but philosophers do not typically consult chemists or other scientists when making these judgements, regardless of their empirical content. We do all this without conducting experiments in the laboratory, and often even without any in-depth knowledge about such experiments. There are several ways that one might react to this puzzle. They include a sceptical reaction, which we will set aside here (cf. the discussion on conventionalism in Chapter 4), a type of reliabilism regarding our a priori tools (perhaps along the lines of Lowe’s work discussed briefly above, i.e., that it’s generally quite unproblematic and easy to grasp essences a priori), and conceptualism, which emphasizes conceptual analysis as the primary area

of philosophical work (we can see elements of this in Jenkins’s account of modal epistemology). A further reaction, proposed by Daniel Nolan, is to

30 31

Jenkins, ‘Concepts, Experience and Modal Knowledge,’ p. 266. This puzzle and some of the discussion that follows is adapted from Daniel Nolan, ‘The A Posteriori Armchair,’ Australasian Journal of Philosophy 93.2 (2015), pp. 211–231.

169

170

The epistemology of metaphysics: a priori or a posteriori?

look at our armchair methods from a slightly different perspective: perhaps they’re not a priori at all, but a posteriori. What Nolan means by this is that our armchair methods generally involve the senses playing a role that is not merely ‘enabling’ – a role that is something more than a necessary part of the acquisition of concepts et cetera. So what more can senses do when it comes to the armchair? Nolan identifies four possible tasks, but we’ll just look at two of them. The first is assembling and evaluating commonplaces. This task amounts to the analysis of such stories as the one about (the statue of) Goliath and the Lumpl, a piece of clay of which Goliath is made. 32 The debate about material constitution can be introduced with the help of such commonplaces. Nolan’s idea in this regard is simple enough: that there are statues made of clay and that such statues can be smashed without destroying the clay are, according to Nolan, pieces of a posteriori knowledge. But the real interest of the suggestion is that philosophers can make surprising ‘discoveries’ on the basis of such commonplaces, which are seemingly available to everyone on the basis of a posteriori knowledge. For instance, the realization that Goliath and Lumpl have different persistence conditions has important consequences for metaphysical debates about composition; indeed, this might be considered to go some way towards answering Peter van Inwagen’s Special Composition Question, which we have already discussed extensively in Chapters 2 and 3. While Nolan is no doubt correct about the involvement of a posteriori methods in our assessment of cases such as the one concerning Goliath and Lumpl, there is still room to argue that the relevant philosophical work is nevertheless a priori. For one might insist that we need to have some grasp of the kind of thing that statues and lumps are before we are in any position to draw the philosophically important conclusions. And if one follows something like Lowe’s line of thought on how we come to know at least a part of the natures of different kinds of things, it would seem that the process starts from a priori inquiry (ignoring the possible ‘enabling’ role that a posteriori knowledge may have). So this task may perhaps be read in two

different ways, depending on one’s prior commitments. Another armchair task that Nolan discusses is applying theoretical virtues. By ‘theoretical virtue’, Nolan means things like internal consistency,

32

For the srcinal example, see Alan Gibbard, ‘Contingent Identity,’ Journal of Philosophical Logic 4 (1975), pp. 187–221.

7.4 Modal empiricism and the status of armchair methods

external coherence, simplicity, explanatoriness, fertility, unificatory power, and other such comparative as well as internal virtues. 33 The need for such a task is of course to assess the respective virtues of different theories, which may help us to choose between them; this task thus consists primarily of comparative judgements. It must be noted immediately that it is typical in the literature discussing theoretical virtues to consider them as a priori.34 Nolan protests that this is just an assumption, because the process of ‘applying theoretical virtues’ is rarely elaborated on. He suggests that the role that theoretical virtues play is primarily epistemic: it’s epistemically better to accept a theory that better satisfies theoretical vir-

tues. Of course, it is another question why, exactly, the theoretical virtues that we usually cite, such as simplicity, are epistemically better. One explanation is simply that such theories are more likely to be true. But other attempts to justify theoretical virtues can be made; Nolan mentions that they can sometimes be justified with reference to other theoretical virtues. For instance, unificatory power may often promote further simplicity, and unifying two theories may also increase explanatory value. Hence, the appeal to unificatory power could be justified in terms of these other virtues, provided that they are considered valuable. Perhaps a more direct justification could sometimes be drawn from predictive success: it has often turned out that a simpler theory makes more accurate predictions, which is at least a pragmatic consideration that speaks in favour of simplicity. So Nolan would suggest that this type of comparative assessment between various theories can be conducted in the armchair since it relies on established evidence rather than any direct empirical work. But it will be left to the reader to judge whether this constitutes a reason not to regard this type of comparative assessment as a priori work, as Nolan argues. Regardless of what we conclude from Nolan’s analysis of armchair philosophy, if we abandon the dogma that modal knowledge must be secured purely by a priori or a posteriori inquiry, perhaps we can make further progress. It is to the credit of the pioneering modal empiricist views

33

34

For some further discussion regarding theoretical virtues, see Chris Daly, Introduction to Philosophical Methods (Peterborough, ON: Broadview Press, 2010). For example, see L. A. Paul, ‘Metaphysics as Modeling: The Handmaiden’s Tale,’ Philosophical Studies 160.1 (2012), pp. 1–29.

171

172

The epistemology of metaphysics: a priori or a posteriori?

that this lesson is clearly taken seriously. Whether or not the resulting view is, properly speaking, modal empiricism is partly just a terminological matter. What seems to be needed, in any case, is some sort of a hybrid view.

7.5 Combining a priori and a posteriori methods There are at least two reasons to think that both in the area of modal epistemology, which we have been focusing on, as well as in the epistemology of metaphysics more generally, there is a need to employ both a priori and a posteriori methods. The first reason is that the area of research

encompasses entities of many different kinds and it is plausible that we cannot acquire knowledge of all these different kinds of things in exactly the same manner. In particular, metaphysics investigates both concrete and abstract entities. When we are dealing with abstract entities, such as sets, it is difficult to see how any judgements we might make about the nature of sets could be anything but a priori , since we have no direct contact with sets (even if we may have direct contact with the members of sets). In contrast, when our research concerns concrete things – say, instances of natural kinds such as water – it seems that we simply have no alternative but to resort to empirical research in order to learn about their properties. The second reason to think that some kind of a ‘hybrid’ view is likely to be correct is due to the issues we discussed in the first section of this chapter and towards the end of the previous section:the a priori vs. a posteriori distinction itself would appear to allow for some vagueness. Hence, a priori and a posteriori elements in metaphysical inquiry may be intertwined to the extent that it is quite difficult if not impossible to take them apart at all. This, at any rate, is the possibility that we will briefly explore in this section. As a first pass, let us briefly return to Lowe’s view regarding modal epistemology, which we classified under modal rationalism. Lowe’s view, at least at first glance, appears to be uniform:he holds that our access to modal knowledge is purely a priori. However, it should be noted that Lowe himself apparently never used the label ‘modal rationalism’ for his view. In fact, Lowe is very critical of the other modal rationalist approaches, such as intuition- and conceivability-based approaches. He argues that a view taking intuitions as evidential in metaphysics, quite generally, is fundamentally ‘ misguided and

7.5 Combining a priori and a posteriori methods

35 leads inexorably to an anti-realist conception of metaphysical claims’. We

will return to the role of intuitions as a source of metaphysical knowledge in Chapter 8, but for now let us focus on Lowe’s positive proposal. We saw that it is based on grasping essences, understanding what a thing is. But how exactly does this happen if not with the help of intuitions or conceivability? That is, what other purely a priori sources of knowledge are there? Positing a distinct rational faculty, that of understanding, is unlikely to be very helpful in answering this question, as we concluded earlier. Yet Lowe makes an important qualification to how he conceives of this process: he holds that our inquiry (into essence and modality at least) should notbe consideredcompletely independent of experience, but rather as proceeding in a ‘cyclical man-

ner, by alternating stages ofa priori and a posteriori inquiry’.36 On this point, the present author is sympathetic to Lowe’s suggestion, having called this type of ‘cyclical’ process a ‘bootstrapping relationship’ betweena priori and a posteriori elements.37 The bottom line is that we should befallibilists about a priori knowledge just as we are about empirical knowledge, and thus deny the

existence of ‘pure’ a priori or a posteriori knowledge. If any work at all is left for the notion of apriority, it seems that we must qualify our conception of it and move beyond the naïve Cartesian account ofinfalliblea priori knowledge. In order to see how a hybrid account of metaphysical inquiry might work, let us look at the classic example of water and H2O. 38 Suppose that ‘water’ designates a genuine, mind-independent natural kind. A genuine natural kind must have a determinable set of identity and existence-conditions and we should generally be able to state them. Whether water in fact is a genuine kind (or whether there are any such kinds) is in fact open to debate, but since water is the

locus classicus , it

serves our purposes well. According to the Kripke–Putnam framework, we know that samples of water are made up of H2O molecules. 39 If it is 35 36 37 38

39

Lowe, ‘Essence vs. Intuition: An Unequal Contest,’ p. 256. Ibid., p. 257. Tahko, ‘ A Priori and A Posteriori: A Bootstrapping Relationship.’ The following discussion is adapted from Tuomas E. Tahko, ‘Natural Kind Essentialism Revisited,’ Mind 124.495 (2015), pp. 795–822. See Saul Kripke, Naming and Necessity (Cambridge, MA: Harvard University Press, 1980) and Hilary Putnam, ‘The Meaning of “Meaning”’ (1975), reprinted in his Mind, Language and Reality: Philosophical Papers, Vol. 2 (Cambridge University Press, 1979), pp. 215–271. Here we will not attempt to be entirely faithful to the views of Kripke and Putnam, but will instead focus on the received view based on their earlier work, here labelled the Kripke–Putnam framework.

173

174

The epistemology of metaphysics: a priori or a posteriori?

also the case that water has its actual microstructure essentially

– even

though empirical work is needed to determine what individual samples of water are made up of – then ‘Water is H 2O’ is a metaphysically necessary a posteriori essentialist truth. That water does have its actual microstructure essentially is usually considered to be knowable

a priori , but

we will see that this assumption must be clarified. Accordingly, the core of the Kripke–Putnam framework is that the combination of an essentialist a priori truth about a given natural kind essence and empirical,

a

posteriori information about the microstructure of that natural kind are

needed to establish such metaphysically necessary theoretical identity sentences. Typically, one would conclude that since empirical information is needed to establish such metaphysically necessary truths, the process of inquiry is, by and large, a posteriori. After all, the content of the a priori part here is very simple indeed, it only amounts to saying that if water is H2O, then this is necessarily so. This a priori part is quite generally taken to be entirely unproblematic, since it is thought that natural kinds such as water are obviously defined by their intrinsic properties, which is to say their microstructure. But once we start to unpack this assumption, it turns out to

be rather more complicated. For it must be recalled here that our initial familiarity with water (and other natural kinds), is not via its microstructure, but rather via its macroscopic, phenomenological properties, such as its boiling point and its ability to dissolve other compounds. So it seems that we need an additional assumption to complement the a priori element to get the desired result, namely the metaphysical necessity of the theoretical identity statement ‘Water is H 2O’. This assumption is that the following principle is true: Microstructural determination (MD) Microstructure determines the macroscopic properties of chemical substances.

Naturally, (MD) is thought to apply to cases such as water. This determination is manifested by chemical properties, that is, properties of a chemical substance that typically become evident in chemical reactions such as oxidation (these are purely chemical properties), or else relate to the chemical structure of the substance (these are sometimes referred to as physical properties). Thus, using philosophical terminology, we might define a

chemical property as a property of a chemical substance in virtue of which

7.5 Combining a priori and a posteriori methods

the substance can undergo chemical reactions. The question that emerges now is: what is the epistemic status of (MD)? For it is this principle that would seem to do most of the work in securing the metaphysical necessity of ‘Water is H2O’ rather than the relatively simple a priori principle we started with – or indeed even the empirical discovery that water is in fact H2O. Moreover, (MD) also seems to be a core part of the criteria of identity for chemical substances. Here we at the heart of the ‘hybrid’ approach: it turns out that the relevant principle – that microstructure determines the macroscopic properties of chemical substances – is not easily established either by ‘pure’ a priori or empirical inquiry. Why is this? Let’s attempt to unpack what it

means that a chemical property is a property of a chemical substance

in

virtue of which the substance can undergo chemical reactions. This should

ring a bell for those who recall our discussion of the ‘in virtue of’ relation in Chapter 5. Take a concrete example: electronegativity – the ability of an atom or a functional group of a molecule to attract electrons. For the purposes of understanding how microstructure and macroscopic properties are assumed to be related according to the picture at hand, tracking the source of this ability is important. In the case of electronegativity, the ability of an atom to attract electrons is influenced by its nuclear charge. Atoms with a higher electronegativity attract valence electrons more strongly; hence the distance from the atom’s nucleus to its valence electrons tends to be shorter. There is a straightforward way in which electronegativity is related to the microstructural properties of the substance, to its nuclear charge in particular. So in this case it seems that we do have a reasonably good idea about how electronegativity and microstructure are related. However, electronegativity is itself something that is not directly observed, and so it is not macroscopic in the sense that chemical properties such as boiling point, solubility, and flammability are. We do of course have a pretty good idea about some of the microstructural properties that are correlated with properties such as these, but correlation is not determination and hence not enough to establish (MD). The problem can be highlighted by asking whether we are really dealing with physical or chemical properties: it used to be (and often still is) common practice to consider properties such as boiling point to

be phys-

ical properties, not necessarily connected with the chemical properties of

175

176

The epistemology of metaphysics: a priori or a posteriori?

a substance, rather than grouping both sets together under ’chemical properties’ as is now often done.

40

Moreover, if microstructure deter-

mines chemical properties, then it seems that these properties should be reducible to microstructure. But this is certainly a problematic assumption, one that requires an argument. The case of water and H2O was thought to be a simple combination of a straightforward a priori principle and an empirical discovery, rendering the relevant metaphysical necessity an a posteriori truth. But the upshot of our discussion seems to be that there is a third part to the story, (MD), which underlies the bridging of the srcinal a priori and a posteriori parts. What is the epistemic status of that principle? This still remains unclear. Certainly, it is a controversial principle, so at the very least we can conclude that we are not dealing with a simple a priori truth. But since current empirical data does not establish the truth of (MD) either, it appears that it cannot be a purely empirical principle either. Yet, the principle has remained largely unchallenged in mainstream metaphysics (albeit it has been repeatedly challenged by philosophers of chemistry). None of this entails that metaphysical inquiry quite generally proceeds as it has apparently done in the case of ‘Water is H 2O’. However, given the prominence of the example, this should at least make us alert to the possibility of such ‘hidden’ premises, and hence also about drawing quick conclusions about the epistemic status of given pieces of philosophical knowledge. In the next chapter, we will look further into the epistemic tools available to the metaphysician. 40

For further details, see Tahko, ‘Natural Kind Essentialism Revisited.’

8

Intuitions and thought experiments in metaphysics

This chapter discusses two general, closely related epistemic tools used in metaphysics: intuitions and thought experiments. As we saw in Chapter 7, both of these tools make an appearance in modal epistemology, which was our primary case study in the epistemology of metaphysics. But since intuitions and thought experiments are used in philosophy more broadly – not just in metaphysics – they deserve a treatment of their own. Note however that we will be focusing on the usage of these two tools in metaphysics. Metaphysicians routinely cite intuitions as prima facie evidence for a given view, but sometimes the appeal to intuitions seems to serve an even stronger role – especially in cases where there do not appear to be any other epistemic tools at our disposal. Thought experiments likewise often

rely on an intuitive reaction, such as in the case of Hilary Putnam’s Twin Earth scenario (discussed in Chapter 4): we are presented with a thought experiment, but instead of an empirical test the relevant test seems to be a test of intuitions. On the face of it, thought experiments are thus a type of ‘intuition pump’, as discussed by Daniel Dennett.1 Given that the literature on intuitions spans the whole discipline of philosophy, we cannot engage with the topic in anything like a thorough manner.2 Instead, we might focus on the status of intuitions with regard to metaphysical thought experiments in particular. Perhaps a better understanding of how intuitions work will also help us to better understand

1

2

Daniel Dennett, Intuition Pumps and O ther Tools for Thinking (New York: W. W. Norton & Co, 2013). For a general overview, see Joel Pust, ‘Intuition,’ in E. N. Zalta (ed.), The Stanford Encyclo pedia of Philosophy (Fall 2014 edn); see http://plato.stanford.edu/archives/ fall2014/entries/intuition/ . A recent anthology of state-of-the-art articles that may prove useful is A. R. Booth and D. P. Rowbottom (eds.), Intuitions (Oxford University Press, 2014).

177

178

Intuitions and thought experiments in metaphysics

the role of a priori knowledge in metaphysics. One reason for addressing intuitions and thought experiments in the same chapter is that, as Darrell Rowbottom puts it, it is a prima facie attractive view that intuitions are to thought experiments as perceptions or observations are to experiments (and hence play an evidential role comparable to experiments).3 While initially attractive, this view does of course have obvious challenges: although we have clear standards of what counts as a good experiment, repeatability being one of the primary ones, intuitions are not, on the face of it, restricted by any such standards. So to get any mileage out of this idea, we ought at least to attempt to set some criteria for what counts as a ‘well-formed’ intuition analogously to those for a respectable experiment. To this effect, we need to be more precise about what is meant by ‘intuition’, for the term is certainly not used in an entirely unambiguous fashion in philosophy. This will be the task of the first section in particular, where two conceptions of intuition are examined – these are the ‘common-sense conception’ and the ‘a priori conception’. In the second section, we will very briefly discuss the study of intuitions in so-called ‘experimental philosophy’. This is strictly speaking not of great relevance in metametaphysics, but we will consider a potential structural issue that may influence discussions of intuitions in metaphysics as well. The third section introduces a further specification regarding intuitions, which is of special importance in metaphysics: ‘experience-based intuition’. These are typically based on the phenomenological aspects of experience. In the fourth section we will consider a version of the ‘ a priori conception’ of intuition in more detail, drawing especially on the work of George Bealer. The fifth and sixth sections concern thought experiments in science and metaphysics, respectively. While we are of course primarily interested here in the philosophical role of thought experiments, they are used not just by philosophers but also by scientists – some of the most famous thought experiments are familiar from the work of Einstein and Galileo, for example. So one question is: do philosophical thought experiments differ from scientific thought experiments? The answer might depend on what the thought experiment is supposed to test. In the case of science, some of

3

Darrell P. Rowbottom, ‘Intuitions in Science: Thought Experiments as Argument Pumps,’ in Booth and Rowbottom (eds.), Intuitions, p. 119.

8.1 Specifying ‘intuition’

the most famous thought experiments have actually been recreated in the laboratory (e.g., the Einstein–Podolsky–Rosen thought experiment regarding quantum entanglement, which we’ll discuss in the fourth section), but in philosophy this would not always be possible. Certainly, many thought experiments in epistemology have been recreated, most famously the Gettier cases – they are quite easy to recreate.4 But there are many metaphysical thought experiments we probably would not like to see realized, at least not the ones involving zombies! (Of course, philosophical zombies are nothing like Hollywood zombies, but some might regard qualia-lacking copies of ourselves rather scary …) However, as it will turn out, the apparent differences between scientific and philosophical thought experiments may not be as drastic as it might first seem.

8.1 Specifying ‘intuition’ Let us focus on two candidate precisifications for the term ‘intuition’ as it is used in philosophy; call them the ‘common-sense conception’ and the ‘a priori conception’.5 These are by no means the only candidate meanings, but they do capture a substantial portion of the uses of ‘intuition’ in philosophy and in metaphysics in particular. The first, common-sense conception, is fairly straightforward: intuitions in philosophy operate like ‘folk’ intuitions generally seem to do, namely as initial, pre-theoretical reactions to various scenarios. In philosophy, this applies specifically to philosophical thought experiments, but ‘folk’ intuitions may of course be applied more broadly. We might take Putnam’s famous Twin Earth thought experiment as our example, bearing in mind that a full analysis of the scenario ought to take into account the discussion in Chapter 7 regarding ‘Water is H2O’.6 The Twin Earth thought experiment could be summarized as follows. Let 4

5

6

Gettier cases are examples of justified true belief (JTB) which we would not intuitively consider as knowledge, hence challenging the classic JTB analysis of knowledge. These and other conceptions of ‘intuition’ are discussed in more detail in C. S. I. Jenkins, ‘Intuition, “Intuition”, Concepts and the A Priori,’ in Booth and Rowbottom (eds.), Intuitions, pp. 91–115. For the srcinal Twin Earth scenario, see Hilary Putnam, ‘The Meaning of “Meaning” ’ (1979), reprinted in his Mind, Language and Reality: Philosophical Papers, Vol. 2 (Cambridge University Press, 1979), pp. 215–271. For further discussion regarding the analysis of the scenario, see Tuomas E.Tahko, ‘Natural Kind Essentialism Revisited,’ Mind 124.495 (2015), pp. 795–822.

179

180

Intuitions and thought experiments in metaphysics

us imagine that the chemical properties of water are reproduced by some molecular structure other than H2O – say, XYZ. This could perhaps happen in another possible world or in a remote location of our own universe – either option introduces some complications, but we can set them aside for the purposes of this discussion. The question is: should we consider this substance, XYZ, to be water? The typical intuitive reaction is that XYZ is not water. The empirical details of the thought experiment are rarely discussed, at least by metaphysicians. But the role of empirical knowledge regarding the existence of metaphysically possible microstructures that could replicate the chemical properties of water is certainly debatable, partly for reasons that we discussed in the previous chapter. Chemistry will presumably be of some help here, but it is often thought that we also need metaphysical a priori work to determine what is metaphysically possible. In later work Putnam himself expressed serious doubts about extending the Twin Earth thought experiment across metaphysically possible worlds and even said that the question about the possible variation of the laws of physics with regard to water ‘makes no sense’.7 Setting Putnam’s worries aside, it is quite common to take the result of our intuitive reaction to the Twin Earth thought experiment to be that water has its actual microstructure by metaphysical necessity. But isn’t this an overly strong result to arrive at on the basis of our ‘folk’ intuitions? Here we arrive at the primary worry regarding the common-sense conception of intuition. Insofar as intuitions are pre-theoretical reactions like the one described in the case of the Twin Earth thought experiment, their evidential role cannot be on a par with that of observations in regular experiments. For one thing, the setting and interpretation of the thought experiment itself will cause complications: in the Twin Earth thought experiment the modal status of the scenario is particularly problematic, as Putnam’s own later reaction makes clear. Moreover, as we discussed in Chapter 7, the intuitive reaction seems to be based on assuming the truth of an underlying, more controversial a priori principle, namely (MD) (which was introduced in Chapter 7): microstructure determines the macroscopic properties of chemical substances. But once we have identified all such complications, our reaction to thought experiments may very well change.

7

Hilary Putnam, ‘Is Water Necessarily H2O?’, in J. Conant (ed.), Realism with a Human Face (Cambridge, MA: Harvard University Press, 1990), p. 70.

8.1 Specifying ‘intuition’

So there are reasons to think that pre-theoretical intuitions understood in the lines of the common-sense conception should not serve an evidential role. Of course, this does not mean that intuitions understood in this fashion could not serve a different role, but it would have to be significantly weaker than we might desire if intuitions are to be considered an important tool in the epistemology of metaphysics. Are there reasons to think that ‘intuition’ can be better specified in terms of apriority? A feature of intuitions that might at least initially point towards the a priori conception is that intuitions are typically considered immediate.8 According to Carrie Jenkins’s analysis, this could mean two

different things. First, intuitions can be immediate in the sense of being direct or non-inferential – or, at any rate, if there is some kind of an inference underlying an intuition, it is not explicit and, perhaps, not conscious. Jenkins considers this type of view about intuitions to be quite common, especially among epistemologists. There is, however, an obvious vagueness to this understanding of immediacy, for it is not clear what kind of a role is played by the purely psychological factor of not being aware of any explicit inference. Jenkins distinguishes this psychological criterion for immediacy from the stronger requirement that immediacy means that there is no underlying inference, not even an implicit, non-conscious one. She does not further engage with this issue, but we might note that the stronger requirement seems, perhaps, too strong, for if even non-conscious inference isn’t allowed, then it might be impossible for a person to know whether her immediate, intuitive reaction to something truly is immediate in the required sense. If this has an influence on the epistemic value of intuitions, then we do not seem to have made much progress. The second sense of immediacy that Jenkins identifies is a more liberal one: intuitions are immediate in the sense of being (or at least seeming to be) obvious, spontaneous, natural, compelling… These features are more straightforwardly psychological (orphenomenological, relating to how the subject perceives the intuitive process), for better or worse. We may be in a risky area here. If we start from the idea that intuitions ought to or could count as evidence, then special attention should be given to the question whether intuitions counting as evidence depend on phenomenological factors. An initial reason for concern is that unless we are very confident about our ability 8

See Jenkins, ‘Intuition, “Intuition”, Concepts and the A Priori,’ p. 94.

181

182

Intuitions and thought experiments in metaphysics

to identify the relevant phenomenological aspects that are associated with intuitions in ourselves, then they may be unlikely to produce a sufficiently reliable method. This is not to suggest that intuitions couldn’t be fallible (more on this shortly). But unless we have some further input on the degree of confidence regarding our ability to identify the relevant phenomenological aspects, then the process seems fallible on two counts. These are our ability to identify something as an intuition in the first place and the reliability of intuitions themselves. If we lack criteria of reliability on both counts, then the use of intuitions might become hopelessly unreliable. But before this sceptical challenge is deemed to refute intuitions as evidence, it ought to percepbe noted that similar considerations might just as well be applied to tion: whether a perceptual state counts as evidence or not likewise seems to

depend on phenomenological factors, at least tosome degree. Sometimes we have phenomenological reasons to think that a certain perception could be unreliable. For instance, if one has blurred vision caused by watery eyes, this would be an immediate phenomenological reason to lower the reliability of visual evidence. So it is not obvious that intuitions are relevantly different from perceptions in this regard. Elijah Chudnoff has recently argued to the effect that intuitions share many of the important epistemic and phenomenological aspects of sensory perception.9 There is at least one important difference as well, though; according to Chudnoff, ‘sensory perception includes concrete reality and excludes abstract reality’, while ‘the subject matter of intuition includes abstract reality and excludes concrete reality’. 10 This clearly sets intuition and perception apart, even if they do share important factors. We will not attempt to settle the abstract/concrete question here though. Chudnoff himself emphasizes that what’s important is not the stronger claim that intuition excludes concrete reality, but that it includes abstract reality. This is enough to distinguish sensory perception and intuition. Let us return to other general conditions for intuition. If immediacy without further qualification seems too crude to characterize intuitions on its own, then perhaps some additional characterization will help. One possibility is to look for a conceptual or linguistic basis, whereby intuitions are associated with conceptual analysis, linguistic competence, or understanding.

9 10

Elijah Chudnoff, Intuition (Oxford University Press, 2013). Ibid., p. 11.

8.1 Specifying ‘intuition’

It’s fairly easy to see where this type of view might srcinate from. For instance, when presented with propositions that are analytically true – true in virtue of the meanings of the concepts involved– we seem to have a relevantly immediate reaction to them. Consider the proposition ‘All vixens are female’. Since vixens are female foxes, the concept of ‘female’ would appear to be contained in the concept of ‘vixen’. So we might say that it is purely by understanding what the word‘vixen’ means that we know ‘All vixens are female’ to be true. An interesting question is whether this involves an inference of some sort. Certainly, in most cases we would consider such cases to be obvious in the sense discussed earlier, so any inference involved would likely be implicit. But it is a matter of debate whether conceptual analysis in general could be understood as free from inference in the relevant sense. In any case, there is clearly a limitation with the view that intuitions are closely connected with conceptual or linguistic competence, as very often intuitions are used in connections where there seems to be more at stake than conceptual or linguistic issues. This is arguably alsothe case in the Twin Earth scenario, even though it should be noted that Putnam was srcinally interested in the semantic aspects of the scenario, namely whatwe mean by ‘water’. But since the upshot of the scenario even when considered on purely semantic terms is supposed to be thatmeanings are fixed externally (it’s the microstructure of water that counts), it appears that the intuitive reaction to the scenario is richer than just a grasp of something like the truth of ‘All vixens are female’. Indeed, Ernest Sosa emphasizes that focusing on purely linguistic features in connection with intuition is not sufficient. Rather, we should be more interested in the underlying process ofunderstanding, which Sosa specifies as follows: Fundamental, intuitive rational beliefs are based atleast on understanding of the propositions believed […]It is not, however,just the understanding of a proposition, whatever its content, that gives a proper basis for believing it. Otherwise, it would also constitute a basis for believing its negation, which must be equally well understood. […] What sufficesis rather the being understood (shared by a proposition and itsnegation pretty much equally) along with the specific content of that very proposition .11

11

Ernest Sosa, ‘Intuitions: Their Nature and Probative Value,’ in Booth and Rowbottom (eds.), Intuitions, pp. 36–49.

183

184

Intuitions and thought experiments in metaphysics

Given that Sosa includes the content of propositions as a crucial part of the basis of intuitive rational beliefs, it would seem that the type of ‘intuition’ at work in the Twin Earth scenario could be salvaged, for it seems possible to build enough information (about the microstructure of water, for instance) into the content of propositions. However, this certainly shifts the focus from pure conceptual or linguistic ability towards something richer. If we need both understanding and some further grasp of the content of propositions to reach intuitive rational beliefs, then the immediacy criterion might once again be questioned. A recent extensive study of the use of intuitions in philosophy, by Herman Cappelen, focuses on this type of conceptual competence view about intuitions, noting their specific phenomenology and epistemic status.12 Cappelen suggests that intuitions understood in this fashion are commonly considered to be used extensively as evidence in contemporary analytic philosophy. But Cappelen’s key claim is that these types of intuitions are in fact not used as evidence, contrary to what many contemporary analytic philosophers think. This calls for further reflection on the methods actually in use in contemporary analytic philosophy. While Cappelen’s project is certainly interesting, we cannot engage with it in detail here. By way of justifying this omission, we have already seen some shortcomings in the conception of intuitions according to which they are based on conceptual competence. In his commentary on Cappelen’s book, David Chalmers has argued to the same effect, noting for instance that even if in some areas of philosophy intuitions seem to stem from conceptual competence, there are other areas where the intuitions relied upon are clearly not similar, such as moral intuitions and other normative intuitions.13 It is more plausible (even if not uncontroversial) that conceptual competence is central for linguistic intuitions, which are naturally of interest in philosophy of language. More to the point, there may be specific reasons to avoid this conception of intuitions when it comes to metaphysical inquiry. For unless we have prior reasons to think that concepts reliably mirror reality, then we have little reason to think that intuitions based on conceptual competence will provide reliable evidence of

12 13

Herman Cappelen, Philosophy without Intuitions (Oxford University Press, 2012). David J. Chalmers, ‘Intuitions in Philosophy: A Minimal Defense,’ Philosophical Studies 171.3 (2014), pp. 535–544.

8.2 Intuitions and experimental philosophy

the structure of reality.14 Admittedly, this requirement may already be too strong: is it even possible to have good reasons to think that our conceptual framework, quite generally, mirrors reality? Since we can’t very well check this by resorting to some other framework, it looks as if the burden of proof here should be on the sceptic. In other words, it may be acceptable to assume that, by and large, our concepts do reflect reality unless there are good reasons to think that they do not. In some cases, such good reasons undoubtedly exist and indeed it would be a move comparable to the sceptic’s to assume that our concept always reliably mirror reality. Accordingly, establishing some criteria to evaluate the reliability of concepts in this sense without assuming either of the extreme views would seem like a reasonable way to make progress. Attempts to engage in this type of work certainly exist; we will discuss some relevant work below.

8.2 Intuitions and experimental philosophy Before we move on, let us briefly note one interesting aspect related especially to the study of linguistic or conceptual intuitions (and of special relevance from the point of view of epistemology). A movement which we have not yet discussed is so-called experimental philosophy or ‘x-phi’.15 This is an area which focuses on, among other things, testing popular philosophical thought experiments and the intuitions that they generate in the general population, especially transculturally. Among popular test cases are the Gettier cases, having to do with intuitions regarding the concept of knowledge. But there is also a growing literature on experimental philosophy applied to metaphysics and to causation in particular. 16 Debate

14

15

16

However, for a defence of the view that experience provides an epistemic grounding for our concepts and hence the needed mirroring, see C. S. Jenkins, ‘Concepts, Experience and Modal Knowledge,’ Philosophical Perspectives 24 (2010), pp. 255–279. For a survey of experimental philosophy, see for instance J. Horvath and T. Grundmann (eds.), Experimenta l Philosophy and its Critics (London: Routledge, 2012). A very recent analysis focusing especially on the source of intuitions is Helen de Cruz, ‘Where Philosophical Intuitions Come From,’ Australasian Journal of Philosophy (2015). See for instance David Rose and David Danks, ‘Causation: Empirical Trends and Future Directions,’ Philosophy Compass 7.9 (2012), pp. 643–653. Another recent area of focus, which we’ll discuss in more detail below, is temporal experience, see L. A. Paul, ‘Temporal Experience,’Journal of Philosophy 107.7 (2010), pp. 333–359.

185

186

Intuitions and thought experiments in metaphysics

about the merits and usefulness of experimental philosophy is heated. One central issue concerns the difference, or lack thereof, in ‘folk’ intuitions as opposed to the intuitions of professional philosophers. In other words, what kind of weight should we put on studies focusing on ‘folk’ intuitions? Also, does it matter if there is transcultural variation? An obvious concern is also the setup of the experiments themselves, since it may be especially difficult to describe philosophical thought experiments to non-experts without biasing the interpretation one way or the other. Further, for someone who does not put much weight on intuitions in the first place, the project of experimental philosophy may seem to be beside the point. For instance, why would it matter whether there are, say, several notions of causation at play in folk ‘intuitions’ if the relevant research is to be done primarily with reference to fundamental physics? Indeed, it is difficult to even discuss the topic of experimental philosophy unless we have some more general agreement about the method of metaphysics, such as that ‘folk’ intuitions do and/or should play a role in it. Because of these difficulties, we might fare better in the present context by focusing on the general role of intuitions in philosophical methodology. To this end, a recent note by Anand Vaidya is of special interest, since he purports to have identified an element that underlies much of the experimental philosophy literature; he calls it the Dependency Thesis (DT): (DT) Experimental inquiry that aims at the discovery of variation in intuitions about the application of philosophical concepts essentially depends on non-experimental intuitions about the application of philosophical concepts. And non-experimental intuitions are indispensible [sic] for experimental inquiry into the reliability of intuitions about the application of philosophical concepts.17

If Vaidya is correct about there being certain ‘non-experimental’ intuitions regarding the application of philosophical concepts, then those intuitions are likely applicable to any philosophical debate concerning philosophical concepts. As Vaidya points out, (DT) itself is a

meta-

physical thesis about the relation between intuitions and inquiry and

hence it appears to be of special relevance from the point of view of metametaphysics. 17

Anand Vaidya, ‘Intuition and Inquiry,’ Essays in Philosophy 13.1 (2012), Article 16.

8.2 Intuitions and experimental philosophy

The crucial notion in (DT) is ‘non-experimental intuition’. Vaidya gives a formal definition of this notion, but an informal sketch suffices here; the central idea is that non-experimental intuitions are something that one must eventually appeal to in order to establish a criterion for determining whether a (philosophical) concept is shared by two groups. Now, in the setting of experimental philosophy, the two groups are likely to be, say, Europeans and Asian, or some other culturally or geographically separated groups. But if we are to apply the idea more generally, we can think of the two groups simply as proponents of two different philosophical views. A good example might be an anti-realist about something and a realist about the same issue. We can immediately bring to mind several examples discussed in earlier chapters that could be analysed in this fashion, such as the debate between nihilists and universalists about mereological composition discussed in Chapters 2 and 3, or the debate between perdurantists and endurantists in Chapter 3. What we are interested in is whether these two

groups share the philosophical concept(s) relevant for the debate, such as the concept of composition. The upshot of Vaidya’s analysis is as follows. If we hope to make progress in any debate possibly involving a variation of intuitions regarding a given concept, we need to distinguish two possibilities: (i) The proponents of two competing views both possess the relevant concept and disagree about its application in a given case. (ii) The proponents of two competing views do not both possess the relevant concept and their disagreement is actually due to applying different concepts in the given case. One could attempt various strategies in distinguishing between these cases, such as conducting an experimental study about the relevant concept to generate the criterion for its usage or looking at a distinct theory which proposes such a criterion.18 But each of these options, as suggested by Vaidya in the case of the concept of ‘knowledge’, is likely to involve an appeal to some further non-experimental intuitions. If this is indeed the case, then intuitions are pervasive in metaphysical inquiry as well. Note that here the proposed role of intuitions is not evidential, but rather

18

For discussion of such attempts with regard to the concept of causation, see Rose and Danks, ‘Causation: Empirical Trends and Future Directions.’

187

188

Intuitions and thought experiments in metaphysics

something that is used to ensure that a debate is substantial, that is, nonlinguistic. How this is to be done, exactly, is another matter. In any case, one consequence of this is that intuitions seem to come into play in metaphysics whenever there is a possibility of a variation in the possession of relevant concepts between proponents of competing views. But let us now move to a more direct usage of intuitions in metaphysics.

8.3 Experience-based intuitions The use of intuitions in metaphysics has not been studied as widely as their use in epistemology and philosophy of language, for instance. One reason for this may be the fact that many of the relevant intuitions do not neatly fit into the specifications that we discussed above. For instance, metaphysical inquiry into the nature of time and space would typically start from our experience of time and space rather than from, say, the analysis of the concepts of ‘time’ and ‘space’. A very recent survey of the usage of such

experience-based intuitions has been made by Jiri Benovsky. 19 Benovsky analyses a number of case studies regarding the use of experience-based intuitions in metaphysics, ultimately drawing the conclusion that the intuitions in use – if they are to be called that – are by and large based on our (varying) phenomenological experiences, which are unable to stand their ground in the face of scientific scrutiny. The upshot of his analysis is that we ought to not take such ‘intuitive data’ as serious metaphysical evidence. But let’s take a look at one of Benovsky’s case studies so that we can decide for ourselves. Take the case of temporal experience with regard to the judgements we make about the A-theory and B-theory of time. 20 Very briefly, the A-theory amounts to the view that tense properties, meaning those involving time, are irreducible – that is, ‘yesterday’, ‘last year’, ‘a million years ago’, and so

19

20

Jiri Benovsky, ‘From Experience to Metaphysics: On Experience-based Intuitions and their Role in Metaphysics,’ Noûs (2013 [online], forthcoming). This is not the place to introduce the reader to the details of these theories, but for a compact introduction, see Ned Markosian, ‘Time,’ in E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Spring 2014 edn); see http://plato.stanford.edu/archives/ spr2014/entries/time/. The locus classicus in this area is J. M. E. McTaggart, ‘The Unreality of Time,’ Mind 17 (1908), pp. 457–473; reprinted in R. Le Poidevin and M. McBeath (eds.), The Philosophy of Time (Oxford University Press, 1993), pp. 23–34.

8.3 Experience-based intuitions

on are fixed by the passage of time and correspond to genuine properties; call these A properties. The B-theory, in contrast, denies the reality of tense and holds that all these properties involving time are reducible to two-place relations such as ‘a day earlier than’, ‘a year earlier than’, ‘amillion years earlier than’; call these B relations. Each theory can be supported with various metaphysical (and scientific) arguments, but most people would agree that the A-theory has a certain intuitive advantage because on the face of it only the A-theory can account for the passage of time. Temporal passage is of course a key part of how we experience the world, so it would appear that the intuitive support for the A-theory is derived from the phenomenology of time, that is, temporal experience.21 The role of this phenomenologi-

cal element is larger than one might initially think: it is at least arguable that if we did not perceive change in our surroundings or even just in our thoughts, then time would not seem to be passing. The force of this point is easy to see when we consider how subjective our perception of the passage of time is: time can appear to pass very slowly or very fast depending on external factors such as changes (or lack thereof) in our environment, or depending on internal factors such as stress or chemically induced states (i.e., drugs). Indeed, our sense of how quickly the time passes does not seem to be very reliable at all. Building on the point that change typically involves movement, Benovsky observes that these phenomenological aspects regarding the passage of time have peculiar consequences: Take the case of the hour hand on a mechanical watch:it moves so slowly that we just do not perceive it as moving. We can of course observethat it has moved, if we look at it after some time, but such a “perception” is no experience of movement at all. Thus, we have a case where there is movement, but we do not experience it as such, simply because the hour hand’s continuous movement is far too slow for us to be able to perceive it. Indeed, what we realize here, is that our capacities to notice change and movement have a lower limit and that anything that moves too slowly will not be registered by our perceptual system as moving.22

Similarly, there are cases of ‘apparent motion’, where weperceive something as moving without there being any real motion at all – various visual 21 22

See Paul, ‘Temporal Experience.’ Benovsky, ‘From Experience to Metaphysics,’ p. [5].

189

190

Intuitions and thought experiments in metaphysics

illusions are a good example. Since our sense of the passage of time is closely linked to change and hence motion, this seems once again to highlight the unreliability of the phenomenology of time perception. What is the upshot? Well, since the phenomenology related to our experience of time is clearly very badly biased and often misleading, any intuitive support that it provides for a general metaphysical theory about the nature and reality of tense is surely in serious doubt. Based on this and other examples, Benovsky states that we can draw no metaphysical conclusions from the nature of our experience. So this purports to be a strong case against using experience-based intuitions as evidence – or even as heuristic tools. There are fairly obvious metametaphysical implications to be drawn from this, one of them being that at least certain types of intuitions are likely to be ‘polluted’ by the phenomenological framework of our experience. If this is the case, we need to seriously reconsider the value of data drawn from these intuitions. But perhaps there are different types of intuitions? If so, such intuitions might be able to avoid being polluted in this sense. One possible avenue are pure ‘ a priori intuitions’, which would appear to be more directly tied to our cognitive or rational abilities.

8.4 Rational intuition It is not untypical to propose a strong connection between intuitions and rationality.23 The result is an intriguing characterization of the a priori conception of intuitions, according to which our belief in certain propositions is justified by rational intuition rather than by the senses, introspection, or memory. A good example is Bealer’s work: he proposes that we ought to distinguish between the ‘a priori intuitions’ used in philosophy and the ‘physical intuitions’ used in science.24 If intuitions were to provide us with similar warrant to a priori knowledge more generally, then their evidential role would clearly be different from that of empirical warrant. Indeed, on Bealer’s line, something that we intuit, philosophically, seems necessary to us. This further tightens the connection between apriority and intuition. 23

24

E.g., George Bealer, ‘Intuition and the Autonomy of Philosophy,’in M. DePaul and W. Ramsey (eds.), Rethinking Intuition: The Psychology of Intuition and Its Role in Philosophical Inquiry (Lanham, MD: Rowman and Littlefield, 1998), pp. 201–240; Laurence BonJour, In Defense of Pure Reason (Cambridge University Press, 1998). Bealer, ‘Intuition and theAutonomy of Philosophy,’p. 165.

8.4 Rational intuition

Yet, it also produces an immediate limitation: if what we intuit is necessary, then the subject matter of intuitions must be the realm of necessary truths, for instance analytic, conceptually necessary truths, such as the earlier ‘All vixens are female’. But it can’t be as simple as this, for surely intuitions are often also fallible , in which case we might sometimes think that what we’ve intuited is necessary (and a priori), but it is in fact false. This is how it should be, though: one can be justified a priori in believing that something is necessary, on the basis of an intuition for instance, but that justification could later be defeated, for instance in light of empirical evidence. Bealer has put forward a number of arguments to the effect that it is incoherent not to consider intuitions as evidence. 25 He pitches them against ‘Quinean’ empiricism, which is committed to three principles: empiricism, holism, and naturalism. The core of their combination is that experience/ observations form the prima facie basis of a person’s evidence, so that a theory is justified if and only if it is the simplest complete theory that justifies that evidence, and the natural sciences comprise the simplest theory. The first of Bealer’s arguments, the Starting Points Argument, is very simple; it points out that even empiricists rely on intuitions when following justificatory procedures. A likely reply would be that even if intuitions are relied upon initially, they do not participate in the justificatory part. So intuitions could serve as a guide in the formulation of theories (the discovery part), but not in their justification. Yet, Bealer points out, if the intuitions involved in the starting points of the theories – the discovery part – were not reliable, then surely they would also influence the justification part. Hence, even the empiricist must admit that the intuitions involved must be able to serve as evidence. Bealer’s second argument to this same effect, the Argument from Epistemic Norms, runs as follows. We have certain epistemic norms that we routinely

use in our justificatory procedures. If the empiricist wishes to rule intuitions out as evidential, then it should be shown that intuitions fail when it comes to this standard assessment of justificatory procedures. More precisely, Bealer argues that to abandon some source of prima facie evidence, it should be demonstrated that this piece of evidence fails to satisfy the

25

George Bealer, ‘The Incoherence of Empiricism,’ Aristotelian Society Supplementary Volume 66 (1992), pp. 99–138.

191

192

Intuitions and thought experiments in metaphysics

‘three cs’: consistency, corroboration, and confirmation. As an example of a clear failure of all three cs, consider prediction made with the help of tea-leaf readings: First, to the extent that we have looked, we find no particular consistency among the tea-leaf readings made by a single person. Second, a person’s readings are not corroborated by other people. Third, there is no pattern of confirmation of the tea-leaf predictions or other tea-leaf claims by our experiences, observations, and intuitions. Indeed, there is a pattern of disconfirmation by these sources of prima facie evidence.26

In contrast, Bealer argues that intuitions manage to satisfy all three cs. First, if you consider your intuitions, it is likely that they are largely a consistent set. But even if some of your intuitions do conflict each other, you

will likely be able to come up with a more complete description of what is happening in the problematic scenario in such a way that the apparent conflict of intuitions can be resolved. Secondly, even though there may be cases where your intuitions conflict with those of someone else, who therefore appears to fail to corroborate your own intuitions, there are also plenty of clear test cases where our intuitions corroborate each other extremely strongly. For example, in the areas of elementary logic and mathematics, intuitions are widely shared. But even in more complex cases, such as the Twin Earth thought experiment, the intuitive reaction to the scenario is widely shared, at least when specified further. So it may be that often the initial conflict in intuitions is simply due to an under-described scenario. Finally, our intuitions seem to be very rarely disconfirmed by direct empirical evidence, since the judgements we make on the basis of intuitions are often quite independent of experience and observation. One might of course further challenge that while some intuitions may satisfy the three cs, it is nevertheless clear that not all intuitions fully satisfy them. But perhaps this is acceptable to a moderate rationalist – to use Bealer’s term. After all, if we endorse the fallibility of intuitions, we had better acknowledge that in some cases they fail. It is enough to maintain that intuitions can, at least sometimes, be used as evidence. Bealer himself attempts to make a somewhat stronger case, but here we will set aside his further discussion, largely because it is aimed primarily at the ‘Quinean’

26

Bealer, ‘The Incoherence of Empiricism,’ p. 110.

8.4 Rational intuition

empiricist, whereas we are interested here in a more general picture of the epistemology of metaphysics– one that may be at odds with ‘Quinean’ empiricism to begin with. But it is worth mentioning that this is a topic which continues to be discussed. Chudnoff, who has developed an alternative to the type of understanding-based account of intuitions on which Bealer, Sosa, and others have elaborated, also challenges such empiricism. An important part of his argument, mentioned in passing earlier, is that intuitions concern abstract matters, whereas sensory perception concerns concrete matters. The empiricist, it is suggested, struggles to explain the former: Our abstract knowledge encompasses knowledge about the necessary, normative, infinite, and abstract. Sensory perception gives us knowledge about the contingent, non-normative, finite, and concrete. And it is not clear how memory, testimony, and deductive, inductive, and abductive inference could work up knowledge about the contingent, non-normative, finite, and concrete that sensory perception gives us into knowledge about the necessary, normative, infinite, and abstract.27

This may remind the reader of issues discussed in Chapter 7 regarding the relationship between a priori and a posteriori reasoning. Chudnoff makes a related point: even if the empiricist has trouble accounting for the abstract, the ‘apriorist’ also has trouble explaining how any ‘pure’ form of reasoning alone could form the epistemic basis for any beliefs. It would seem that some pre-existing background knowledge is always required. Chudnoff himself proposes that intuitive reasoning could be the missing link: it ‘injects’ a priori reasoning with some content – this is a crucial element in Chudnoff’s account of intuitions as ‘intellectual perceptions’. 28 However, it will be left up to the reader to decide whether ‘pure’ empiricism can be maintained when it comes to metaphysics. We have certainly seen that the question is considerably more difficult than it may have initially seemed. To finish this section, let us take a brief look at a recent attack on intuition-based philosophy of the type suggested here, by James Ladyman and Don Ross.29 This attack is based on something that we mentioned earlier: the apparent difference between philosophical armchair intuitions 27 28 29

Chudnoff, Intuition, p. 14. Ibid., p. 15. J. Ladyman and D. Ross (with D. Spurrett and J. Collier), Every Thing Must Go (Oxford University Press, 2007), pp. 10–15.

193

194

Intuitions and thought experiments in metaphysics

and those used in science. Ladyman and Ross give a number of examples of metaphysical armchair intuitions which seem to be blatantly incorrect from a scientific point of view. Indeed, it is easy to find such examples, in metaphysics and science alike – just consider the fact that Newton devoted more time to alchemy than any other area of research and apparently thought that this work was highly important. So they emphasize the point that intuitions are clearly not a reliable method of inquiry, but they don’t think that this entails that armchair reasoning is completely worthless. For they admit that it is often said that a particularly good physicist has ‘sound physical intuition’. But here the use of the word ‘intuition’ is supposedly different, as it refers to ‘the experienced practitioner’s trained ability to see at a glance how their abstract theoretical structure probably – in advance of essential careful checking – maps onto a problem space’.30 Ladyman and Ross further distance the intuitions of metaphysicians from the intuitions of scientists by pointing out that the former are often taken as evidence whereas the latter are only heuristically valuable. On this view, intuitions should at best constitute only prima facie evidence: careful study, or in some cases empirical research, is required before they can be accepted. So it may indeed be that intuitions in philosophy and science are used somewhat differently and perhaps this speaks against them as reliable evidence in philosophy. But perhaps we should not put too much weight on this difference until a more comprehensive study of how intuitions are in fact used in science is seen.31 We will now change the topic a bit and look into the method of generating intuitions in the first place, namely thought experiments; it may be illuminating to start by considering exactly how thought experiments are used in science.

8.5 Scientific thought experiments To continue on from the previous section, it is worth noting at the outset that thought experiments are perhaps the most promising example of shared ground between science and philosophy, since in certain areas of natural science, such as physics, thought experiments have a long 30 31

Ibid., p. 15. For what may very well be a first systematic attempt at such a study, see Jonathan Tallant, ‘Intuitions in Physics,’ Synthese 190 (2013), pp. 2959–2980.

8.5 Scientific thought experiments

history. The relationship between science and metaphysics more generally is a topic that we will focus on in Chapter 9, so we can here limit our focus on thought experiments in particular. The view that scientific and philosophical thought experiments are indeed similar is rather popular, at least among philosophers. But rather than attempt to list the potential similarities between them, we might instead consider what could possibly make philosophical and scientific thought experiments different. Some reasons to think that they are different have been put forward

by David Atkinson – a physicist. 32 Atkinson’s main point is that thought experiments which do not lead to real, empirical experiments, are not as valuable as the ones that do. Since philosophical thought experiments typically do not lead to experiments while many scientific ones do, this seems like an obvious difference between them. But note that Atkinson does not consider only philosophical thought experiments to be poor ones, but also a number of scientific thought experiments, such as Galileo’s famous thought experiment about falling bodies in response to Aristotle’s theory. In Galileo’s thought experiment, we are to imagine two objects of difference masses connected via a string. This system of two objects should be dropped from, say, a tower. If we assume that heavier objects fall faster than lighter ones – as Aristotle’s theory of gravity predicts (objects fall at a speed relative to their mass) – then the string connecting the lighter object to the heavier one will quickly tighten, with the heavier object pulling the lighter one down and the lighter one slowing down the heavier object. But, of course, since the system considered as a whole is certainly heavier than either one of the objects considered on their own, the system as a whole should fall even faster. This produces a contradiction and should lead us to conclude that one of our assumptions must be false. Now, this particular thought experiment did supposedly lead to a real one, where Galileo dropped two balls of different masses from the Leaning Tower of Pisa in order to demonstrate that they would fall at the same rate instead of at a rate depending on their relative masses. This was to prove Galileo’s own theory while falsifying Aristotle’s. 32

David Atkinson, ‘Experiments and Thought Experiments in Natural Science,’ in M. C. Galavotti (ed.), Observation and Experiment in the Natural and Social Sciences, Boston Studies in the Philosophy of Science 232 (Dordrecht: Kluwer, 2003), pp. 209–225.

195

196

Intuitions and thought experiments in metaphysics

Without dwelling on the historical details too much, Atkinson’s complaint concerning Galileo’s thought experiment is that there is in fact nothing inconsistent in Aristotle’s srcinal idea, contrary to what Galileo claimed. Aristotelian theory suggests that the time it takes for a body to fall is inversely proportional to its weight and this theory does hold some truth when the body is falling in a liquid, such as water. So Atkinson suggests that Galileo perhaps misread Aristotle and, moreover, presented his thought experiment of the imagined inconsistency as a polemical device. Whether or not this judgement is correct, we do know that even if Aristotle’s reasoning was consistent, his account of gravity is nevertheless unsatisfactory. Furthermore, this hardly tells us anything about the actual process by which Galileo reached his conclusion about falling bodies. Thus, even though Galileo’s thought experiment, as we know it, might not quite establish everything that Galileo thought it did, it is nevertheless able to rule out one possible way to conceive of gravity. In this regard, one might

think that Galileo did indeed produce a fine thought experiment, even if it is based on the assumption that the medium is air rather than, say, water. We might compare the situation with Newtonian mechanics: we know that it breaks down in situations where the speed of light is approached, for instance. But this obviously does not reduce the value of Newtonian mechanics in more limited contexts, for the theory successfully rules out certain possibilities regarding motion (e.g., a body cannot accelerate unless there is a force acting upon it). Consider another famous thought experiment in science, which Atkinson considers to be a good one: the Einstein–Podolsky–Rosen (EPR) thought experiment. The EPR thought experiment attempts to explain away the ‘spooky action at a distance’ phenomenon (as Einstein called it) of quantum entanglement. If we measure, say, the spin of an electron in a quantum system which consists of two electrons travelling to different directions, this apparently has an immediate effect on the other electron in the system, although the two electrons are seemingly independent of each other and could even be miles apart; thus the ‘spooky action at a distance’. Einstein, Podolsky, and Rosen explained the phenomenon by introducing so-called ‘hidden variables’: there must be something more to reality than the standard quantum theory suggests, and this accounts for the strange results. Now, the EPR thought experiment did lead to real experiments, but its conclusion is in fact incorrect. John Bell’s later work is considered to

8.6 Philosophical thought experiments

have corroborated the so-called ‘Copenhagen interpretation’ of quantum mechanics, which does not postulate hidden variables. The details are history, but the value of the EPR thought experiment would seem to be that it produced a testable prediction – something that philosophical thought experiments almost never do. On a related point, Atkinson claims that string theory (or its modern version, ‘supersymmetry’) – a complex theoretical framework attempting to explain the observed elementary particles in terms of the quantum states of one-dimensional ‘strings’ – is an example of a bad thought experiment. The reason for this is the most common complaint against string theory, namely that it seems that we can never achieve the energy required to test the theory empirically, and hence it will not lead to empirical experiments. It’s not entirely clear that this is true, since some attempts to derive testable predictions from string theory have been made, but we can leave this aside here. Let us instead consider some more general indications of a bad thought experiment – in particular, can philosophical thought experiments ever overcome their obvious limitations?

8.6 Philosophical thought experiments Since the previous section began with Atkinson’s analysis of scientific thought experiments, we could begin this section by considering the two indicators which, according to Atkinson and Jeanne Peijnenburg, reveal that a thought experiment is a bad one – their prime examples are famous philosophical thought experiments.33 These indicators are contradictory conclusions and conclusions which beg the question. As an example of the first

one, Peijnenburg and Atkinson mention the Doppelgänger thought experiment which produced a heated debate in the philosophy of mind. Your Doppelgänger is a molecule-for-molecule physical duplicate of you, pro-

duced, say, by a freak chemical reaction generated by a lightning strike and swamp gas.34 The question is of course whether your physical duplicate is also mentally identical to you. Opinions are divided, with one side insisting 33

Jeanne Peijnenburg and David Atkinson, ‘When Are Thought Experiments Poor Ones?’, Journal for General Philosophy of Science 34.2 (2003), pp. 305–322.

34

This ‘swampman’ thought experiment srcinates in Donald Davidson, ‘Knowing One’s Own Mind,’ Proceedings and Addresses of the American Philosophical Association 60.3 (1987), pp. 441–458.

197

198

Intuitions and thought experiments in metaphysics

that of course the duplicate is also mentally identical and the other side claiming that something would be missing – perhaps the duplicate would be a phenomenological ‘zombie’.35 These two very different intuitive reactions are of course untestable, except perhaps by another thought experiment. The upshot is that the Doppelgänger thought experiment produces irredeemably contradictory conclusions, even among experts. In fact, the same thought experiment serves to illustrate the other indication of a bad thought experiment, according to Peijnenburg and Atkinson. They claim that the contradictory conclusions in this case are caused by question-begging premises: the thought experiment is meant to explain our intuitions about the mental and the physical, but these intui-

tions are also the cause of the contradictory conclusions. The worry, then, is that the thought experiment relies on the very intuitions that it is designed to produce and it certainly gives us a limited idea about the methodology behind thought experiments. Take once again the EPR thought experiment, which apparently did not make the correct prediction, although it produced a real experiment (quite a bit after the actual thought experiment was introduced). It seems thus that the EPR thought experiment was good only because of the contingent fact that Bell happened to find a way to test it empirically. One question that emerges is how long we are supposed to wait for a potential empirical experiment before we deem a thought experiment to be a bad one? This is a concern that Daniel Cohnitz has also raised.36 Indeed, unless we can define what constitutes a thought experiment, we might not even recognize the work that thought experiments do (in science and philosophy). Peijnenburg and Atkinson refuse to attempt this: Since we are preoccupied with the difference between good and bad, we do not feel the need to state exactly what thought experiments are; after all one can distinguish good from bad theories, or thoughts, or experiments without being able to define what exactly theories, thoughts or experiments are.37

35 36

37

See David Chalmers, The Conscious Mind (Oxford University Press, 1996), p. 95. Daniel Cohnitz, ‘When are Discussions of Thought Experiments Poor Ones? A Comment on Peijnenburg and Atkinson,’ Journal for General Philosophy of Science 37.2 (2006), pp. 373–392. Peijnenburg and Atkinson, ‘When Are Thought Experiments Poor Ones?’, p. 306.

8.6 Philosophical thought experiments

Yet, one might think that the value of a thought experiment, its ‘goodness’ or ‘badness’, is closely tied to what thought experiments are. If we set aside the purely pragmatic criterion – that a thought experiment is only valuable if it produces a real experiment – then we ought to try to make some progress in this regard. Note that even if we were to look for a purely pragmatic criterion, philosophical thought experiments will clearly require a different criterion than the one proposed by Peijnenburg and Atkinson. As a starting point, we may derive a hint from our earlier discussion: it looks as if a thought experiment can be valuable while failing to correspond with actual reality; that is, thought experiments by themselves do not need to be a reliable guide towards how things are in the actual world. To put this in terms of an example, just take the EPR thought experiment: although the thought experiment is perfectly consistent, it turned out not to correspond with actuality (since it was refuted by Bell’s experiments). Naturally, thought experiments that do not correspond with the actual world in the relevant sense might not be very interesting, at least beyond the purely pragmatic value of urging someone to falsify them experimentally. But the goal of philosophical thought experiments is clearly different – it would seem that it is enough if the thought experiment describes a (metaphysically) possible scenario. Now, it should be immediately noted that one area of debate with regard to many philosophical thought experiments, such as the one involving Döppelgangers, is exactly whether they are possible – or indeed even conceivable . Setting this issue aside and assuming that most philosophical thought experiments do succeed in describing a possible scenario, it remains a separate issue whether this scenario is true, or in other words, corresponds with actuality. Insofar as a thought experiment succeeds in this regard, it is typically left to intuitions or further argument to settle the truth of the matter (but we have of course discussed various problems regarding the use of intuitions as evidence). In any case, the suggestion that thought experiments deal with possibility brings us back towards the discussion of Chapter 7, namely modal epistemology. Since we have discussed modal epistemology in great detail already, we do not need an in-depth discussion here. But it may be worth noting that mere conceivability is unlikely to be enough. The reason for this was discussed in Chapter 7, but can be found already in Roy Sorensen’s classic account of thought experiments: ‘We may open a modal inquiry with a

199

200

Intuitions and thought experiments in metaphysics

casual appeal to what we can imagine, but we cannot close it’.38 In other words, room for error always remains. But even if we sometimes conceive of impossibilities, Sorensen suggests that more often than not we nevertheless manage to conceive of a possible scenario. Conceivability presumably does capture various (metaphysical) possibilities and it is plausible that all possibilities are, at least in principle, conceivable. The sceptic may complain that since there is no clear way to ensure that something we conceive is genuinely possible, this overlap can be regarded as worthless. But as Sorensen notes, drawing an analogy with memory, this type of scepticism may be too extreme to be an interesting thesis (it is near self-refuting).39 If conceivability is not enough, what are our options? An appeal to conceptual knowledge or conceptual analysis might be tried, and often is. This is the approach that Frank Jackson takes. 40 Jackson is specifically interested in philosophical thought experiments, such as Putnam’s Twin Earth thought experiment, but he does consider applying the conceptual framework to scientific thought experiments as well. For instance, he considers Galileo’s thought experiment about falling bodies in comparison to the Twin Earth thought experiment and draws the following conclusion: We should not be too surprised at thought experiments revealing facts about the empirical world. Detective stories make us familiar with the idea that reconstructing ‘in our minds’ what would have been involved in the butler doing it may reveal that he could not have done it. This is surely very different from the Twin Earth thought experiments. They do not lead us to revise our views about what Earth is like, or indeed what Twin Earth is fundamentally like.41

We do not need to return to the Twin Earth thought experiment here. Instead, we could question Jackson’s distinction between philosophical and scientific thought experiments. Consider one of the most famous philosophical thought experiments, from Jackson himself: the case 38 39

40

41

Roy Sorensen, Thought Experiments (Oxford University Press, 1992), p. 41. For further discussion on conceivability, see Marcello Oreste Fiocco, ‘Conceivability, Imagination and Modal Knowledge,’ Philosophy and Phenomenological Research 74 (2007), pp. 364–380. Frank Jackson, From Metaphysics to Ethics: A Defence of Conceptual Analysis (Oxford: Clarendon Press, 1998). Jackson, From Metaphysics to Ethics, pp. 78–79.

8.6 Philosophical thought experiments

of Mary, the colour scientist. 42 As we will recall, Mary is confined into a black-and-white-room and learns every last thing about colour from books. Does Mary learn anything new when she actually sees a new colour, say, red? The expected intuitive reaction is to answer ‘yes’, which is supposed to show that physicalism is false. But how are we to classify this thought experiment, on Jackson’s criteria? According to Peijnenburg and Atkinson, the case of Mary is an example of a bad thought experiment, comparable to the Twin Earth thought experiment.43 However, according to Jackson’s criteria, it would seem that the Mary thought experiment is closer to a scientific thought experiment, as it is supposed to show that physicalism is false and thus it clearly tells us something about the world – an indicator of a scientific thought experiment in Jackson’s terms (never mind the fact that this is a very controversial result!). Or take Sorensen, who considers the Mary thought experiment to be a counterexample to the view that thought experiments are appeals to ordinary language – Jackson’s own thought experiment would then seem to speak against the conceptual approach to philosophical thought experiments, just like his reaction to the Twin Earth thought experiment.44 All this suggests that the distinction between scientific and philosophical thought experiments may break down when we look into the details. Let’s take a step back. Is there anything that is obviously shared between philosophical and scientific thought experiments? If we consider the examples that we have discussed, there is at least one thing that surfaces: thought experiments in both disciplines would seem to make modal commitments. When we construct thought experiments we are, quite generally, interested in different possible scenarios which are somehow constrained. The constraint is typically expressed in terms of what could explain the current empirical data in such a way that it fits a certain theoretical paradigm (such as in the EPR thought experiment) or in terms of the existing theoretical paradigm when we encounter novel empirical data, such as in the Twin Earth thought experiment. The EPR thought experiment attempts to maintain the key elements of the classical framework and the intuitions that support it (such as locality) in the face of quantum entanglement and the 42

43 44

Frank Jackson, ‘What Mary Didn’t Know,’ The Journal of Philosophy 83 (1986), pp. 291–295. Peijnenburg and Atkinson, ‘When Are Thought Experiments Poor Ones?’, pp. 309–310. Sorensen, Thought Experiments, p. 94.

201

202

Intuitions and thought experiments in metaphysics

non-classical characteristics involved in the process of measuring quantum states (such as spin). In contrast, the Twin Earth thought experiment operates within the theoretical framework in which chemical substances are considered to be defined by their microstructure and tests our intuitions in a scenario where novel empirical information challenges this framework (by introducing XYZ, which behaves like water even though it has a different microstructure). There are of course various more complicated elements present in both thought experiments, but both of them clearly require a modal element – indeed one might say that the modal element is the very core of both thought experiments. It is another question entirely how we delimit the space of possible scenarios in such a way that it enables us to focus on interesting thought experiments instead of blatantly silly ones (e.g., those involving magic, as in Disney fairy tales, etc.). It is obviously not enough to insist that the scenario should be internally consistent, since even many fantastical thought experiments manage to do that without having any scientific or philosophical use or plausibility (although the line of demarcation is admittedly vague). These are questions that we have considered, to some extent, in Chapter 7, but more work in this area remains to be done. Unfortunately, we cannot finish the work here. Instead, we will move into a more in-depth discussion regarding the relationship between science and metaphysics, now that we have found at least one interesting area of potential overlap. Another tool that may provide some further insight in this regard is the process of modelling, which we will discuss in Chapter 9.

9

Demarcating metaphysics and science: can metaphysics be naturalized?

Debate concerning the relationship between metaphysics and science has been intense for a long time. It was an especially important topic in the work of Carnap and Quine, given the hostility of the Vienna Circle and logical positivism/empiricism more generally towards metaphysics. 1 However, it should be noted that what the logical positivists meant by ‘metaphysics’ is quite far removed from the type of analytic metaphysics that we have been concerned with in this book. Here’s a succinct example from Carnap: Logical analysis, then, pronounces the verdict of meaninglessness on any alleged knowledge that pretends to reach above or behind experience. This verdict hits, in the first place, any speculative metaphysics, any alleged knowledge by pure thinking or by pure intuition that pretends to be able to do without experience. But the verdict equally applies to the kind of metaphysics which, starting from experience, wants to acquire knowledge about that which transcends experience by means of special inferences […].2

We can see that the type of metaphysics Carnap has in mind does not align very neatly with the difficulty of distinguishing a priori and a posteriori elements discussed in Chapter 7, nor with the specification of intuition in Chapter 8. Indeed, in the paper quoted above, Carnap discusses authors such as Hegel and in the 1957 ‘Remarks by the author’ added to the English translation he explicitly adds that he had in mind the systems of Fichte, 1

2

For an overview, see Richard Creath, ‘Logical Empiricism,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2014 edn); see http://plato.stanford.edu/ archives/spr2014/entries/logical-empiricism/ Rudolf Carnap, ‘The Elimination of Metaphysics Through Logical Analysis of Language,’ trans. by Arthur Pap, in A. J. Ayer (ed.), Logical Positivism (New York: The Free Press, 1959), p. 76; srcinally published in Erkenntnis 2.1 (1931) as ‘Überwindung der Metaphysik durch logische Analyse der Sprache,’ pp. 219–241.

203

204

Demarcating metaphysics and science: can metaphysics be naturalized?

Schelling, Hegel, Bergson, and Heidegger. So, while it’s unlikely that the logical positivists would be entirely happy with contemporary analytic metaphysics, the primary target of their criticism was metaphysics of a quite different type. The background of the movement was the departure of various subdisciplines from the remit of philosophy. Since by the beginning of the twentieth century there were already established subdisciplines of physics, mathematics, biology, psychology, and so on, it was not clear what was left for philosophy to do. One perhaps fairly understandable reaction to this was that all that’s left is ‘metaphysics’, something completely non-empirical, based on a privileged access to some quite different type of evidence than that to which the empirical sciences appeal. This led to the mystification of philosophy, which now seemed quite distant from the various empirical sciences that went from triumph to triumph. The logical positivists, then, were extremely suspicious of the type of a priori reasoning that would clearly have to be the basis of anything that was left for philosophy to study. Of course, as we have seen in the previous chapters, the situation is not quite as simple as that, even if the strict demarcation between metaphysics and science were correct: there are still ways to accommodate a posteriori methods in metaphysics. The question about the demarcation between metaphysics and science may not be quite as pressing as it was when the logical empiricists were active, but this doesn’t mean that the problem has been solved. Far from it: the development of modern physics and especially quantum mechanics have made it clear that the world may in fact be stranger than even the wildest metaphysical speculations have dared to suggest. As Tim Maudlin puts it: Empirical science has produced more astonishing suggestions about the fundamental structure of the world than philosophers have been able to invent, and we must attend to those suggestions.3

This generates a novel worry about the status of metaphysics: if the world is truly as strange as science suggests, then how could metaphysics ever be able to say anything informative about it independently of the sciences? The likely answer is that it cannot, but this is already something that 3

Tim Maudlin, The Metaphysics Within Physics (Oxford University Press, 2007), pp. 78–79.

Demarcating metaphysics and science: can metaphysics be naturalized?

all or at least most contemporary metaphysicians would accept. Indeed, it would be a mistake comparable to logical empiricism to pursue metaphysics completely in isolation of empirical science. But some would argue, especially in view of all the strange puzzles in physics, that this goes both ways: science needs metaphysics as well. In this chapter we will examine this delicate relationship between metaphysics and science in more detail. We will start from a more classical picture: metaphysics is autonomous and able to tell us something about the world on its own, despite the complications introduced by modern physics. On this view, metaphysical inquiry into the fundamental structure of reality can uphold scientific realism: the categories, natural kinds, and other classifications introduced in metaphysics do, by and large, ‘carve at the joints’. In the second section we will move to the other extreme: fully naturalistic metaphysics. According to this view, metaphysics cannot hope to say anything about reality independently of science. We will lay out this view in its most extreme form, which effectively becomes eliminativist: the only possible metaphysics is really just science. This type of view is perhaps closest to logical empiricism, but attempts to avoid some of its caveats. The third section continues in the same vein, but introduces further qualifications regarding the relationship between science and metaphysics. Here, the focus is especially on fundamental physics. We will examine two principles, which have recently been broadly discussed in the literature: the Principle of Naturalistic Closure and the Primacy of Physics.4 The upshot of these principles is that metaphysics still has a role to play, but it is severely limited, primarily unificatory. We will also critically assess these principles. The fourth section examines the possibility of building a methodological bridge between science and metaphysics: even if the subject matter of the two disciplines is distinct, perhaps there are some similarities in their method? Finally, in the fifth section, a more modest, reconciliatory view of the relationship between science and metaphysics will be proposed: ‘moderately naturalistic metaphysics’. According to this view, metaphysics does have a fairly autonomous position, but it must take heed of the latest results in science. Similarly, science can benefit from metaphysical analysis, since

4

These principles srcinate in J. Ladyman and D. Ross (with D. Spurrett and J. Collier), Every Thing Must Go (Oxford University Press, 2007).

205

206

Demarcating metaphysics and science: can metaphysics be naturalized?

scientific theories must be interpreted, and this interpretation can be seen partly as the task of metaphysics. 9.1 Autonomous metaphysics

There are plenty of metaphysicians who still pursue metaphysical topics with little or no reference to scientific work. This may be more justifiable in some areas of metaphysics than in others, but it does no doubt require some justification, given that metaphysics supposedly deals with the structure of reality and science arguably gives us the best tools to gain accurate knowledge about reality. One strategy is to focus on the role of metaphysics as conceptual analysis. On this view, metaphysics is concerned with clarifying certain important notions and primarily based on linguistic considerations. Metaphysics would thus be, by and large, an a priori discipline employing armchair methods, and hence largely independent of a posteriori elements. There is more to be said about the relationship between conceptual analysis and science – for instance, if metaphysics analyses scientific concepts, then there would seem to be an important area of overlap between the disciplines. It could also be argued that, rather than being autonomous, metaphysics conceived as conceptual analysis is derivative: science defines the relevant concepts, metaphysics simply clarifies them. We have discussed this type of approach in passing, in Chapters 7 and 8, but we will leave it aside here. Instead, we will examine a more traditional approach attempting to secure the autonomy of metaphysics from science – perhaps the most traditional. The roots of this approach go back to Aristotle: There is a science which investigates being as being and the attributes which belong to this in virtue of its own nature. Now this is not the same as any of the so-called special sciences; for none of these others deals generally with being as being. They cut off a part of being and investigate the attributes of this part – this is what mathematical sciences for instance do. Now since we are seeking the first principles and the highest causes, clearly there must be some thing to which these belong in virtue of its own nature.5

5

Aristotle, Metaphysics, trans. W. D. Ross, revised by J. Barnes (Princeton University Press, 1984), Bekker page numbers 1003a22–28.

9.1 Autonomous metaphysics

In Aristotle’s time, mathematics had already been demarcated from philosophy, at least to a certain extent. Aristotle himself was of course a ‘natural philosopher’ in its true sense; he also engaged in some empirical research (although it is debatable as to how much) and wrote about biology, physics, and many other ‘special sciences’. But it is metaphysics that Aristotle considered to be the most general science. Thus, Aristotelian metaphysics is the study of being as it is in itself (being qua being), whereas the special sciences investigate only a part of that being. It may seem strange to start a discussion about the relationship between metaphysics and science from Aristotle, but it is useful to keep in mind that philosophy and science were not always as obviously distinct as they may now appear to be. To a certain extent, the contemporary defence of various forms of ‘naturalism’ is an attempt to get back towards a situation not unlike the one seen in ancient philosophy, where natural science was still more or less a part of philosophy. No one, or hardly anyone, wants to defend a conception of philosophy or metaphysics that is non-naturalistic – that would almost be like saying that it is magic! For instance, even many contemporary proponents of dualism might consider their view to be perfectly compatible with (some forms of) naturalism. One could say there has been a certain inflation of the term ‘naturalism’. Accordingly, we will avoid using the term to define views, except when it is qualified further. 6 The closest that we can get to a definition of ‘naturalism’ that takes into account of all the different uses is perhaps something like ‘scientifically minded’, or ‘scientifically respectful’. Or even just a negative thesis: ‘not directly incompatible with contemporary science’. It’s clear that something as vague as this is not particularly helpful in demarcating different approaches. But let us return to ‘autonomous metaphysics’. Note that we can now say that if metaphysics is autonomous, this does not entail that it is non-naturalistic. What it does entail is that the subject matter of metaphysics must be, to a certain extent, different from that of natural science. But different in what sense? If we follow Aristotle’s cue, it would appear that metaphysics is the science that examines being ‘in virtue of its own nature’. So rather than focusing on any single detail of ‘being’ – or 6

For discussion of various forms of naturalism in philosophy, see Jack Ritchie, Understanding Naturalism (Stocksfield: Acumen, 2008).

207

208

Demarcating metaphysics and science: can metaphysics be naturalized?

reality – autonomous metaphysics studies it as a whole. How is this different from fundamental physics? Take the Standard Model of particle physics, which lists the particles that are currently thought to be fundamental, the ‘building blocks’ of reality. Fundamental physics studies the properties of these particles, the relations between them, and the fundamental forces that interact with them. But this is not, of course, quite what metaphysicians are doing – certainly not if metaphysics is taken to be completely autonomous from science. Instead, metaphysics in the present sense may be conceived as studying reality in an even more fundamental sense than fundamental physics does. This type of metaphysics is not interested in listing the various fundamental particles: fermions, bosons, … Rather, it is interested in listing the most basic categories that the fundamental ‘building blocks’ belong to. This is metaphysics of a ‘neo-Aristotelian’ ilk, the study of the fundamental ontological categories: particulars, universals, tropes/modes, and so on. An important part of autonomous metaphysics is to determine how many of these fundamental ontological categories there are.7 Although the study of ontological categories is a good example of how autonomous metaphysics can proceed, this doesn’t mean that one couldn’t pursue such research in connection with more ‘scientific’ metaphysics. For instance, many metaphysicians working on these questions would readily revise their ontological picture in response to new scientific discoveries. One example of such a revision could be the attention that monism – the view that the world is, in some sense ‘one’ – has received in recent years. Jonathan Schaffer, who is an industrious defender of a type of monism, takes ‘quantum emergence’ to be a strong argument in favour of monism.8 The argument runs as follows: (1) The whole possesses emergentproperties (due toquantum entanglement). (2) If the whole possesses emergent properties, then the whole is prior to its parts. (3) Therefore, the whole is prior to its parts. 7

8

For instance, a defence of the view that there are four such categories can be found in E. J. Lowe, The Four-Category Ontology (Oxford: Clarendon Press, 2006). For various arguments to the effect that there should be fewer or more categories than four, see the articles in Tuomas E. Tahko (ed.), Contemporary Aristotelian Metaphysics (Cambridge University Press, 2012). Jonathan Schaffer, ‘Monism: The Priority of the Whole,’ Philosophical Review 119 (2010), pp. 31–76.

9.1 Autonomous metaphysics

This results in priority monism, which we discussed in Chapter 6. An assessment of the argument would require looking into the relevant physics, but all that matters for us now is that, as a matter of fact, much of what could initially be labelled as ‘autonomous metaphysics’ is not quite as autonomous as may first seem, since input from contemporary science is taken into account. But could this relation be symmetric? That is, could ‘autonomous’ metaphysics also influence science? This certainly seems to be a hope that some metaphysicians have. The influence may not be seen as direct, though, but instead as laying the foundations of scientific inquiry more generally or in terms of structuring scientific thought in some way. The idea is to develop a metaphysical foundation for natural science – to explain why science is possible and indeed successful in the first place. 9 However, at this point, more ‘scientifically-minded’ philosophers tend to get outraged: surely it is not our place as philosophers to tell scientists how their discipline should be, or even can be pursued! But that is not of course what even the more ambitious metaphysicians have in mind, exactly. A more charitable reconstruction would be to say that metaphysics may be considered to open up novel areas of research, perhaps by giving birth to new special sciences. This approaches the idea expressed in the earlier Aristotle quote: metaphysics studies being in itself and in the process of this study, new areas – parts of being – may be discovered. The special science that studies that new part of being is no longer metaphysics, but since metaphysics encompasses all of being, there is no part of it that wouldn’t be of some relevance for metaphysics as well. There are plenty of potential examples of philosophy giving birth to new special sciences. Besides Aristotle’s work, which paved the way for physics, biology, and other special sciences that have been established for a long time, we could mention psychology and cognitive science as more recent examples. The exact process by which these special sciences came to be is of course quite complicated, but it’s quite clear that many questions being pursued in various special sciences today were once questions that were pursued by philosophers (or, perhaps more accurately, philosopher-scientists). Let us conclude this section with a more specific example of how metaphysics may be understood to be pursued autonomously of science. The 9

Again, see Lowe, The Four-Category Ontology; the subtitle of the book is A Metaphysical Foundation for Natural Science.

209

210

Demarcating metaphysics and science: can metaphysics be naturalized?

case of material composition, which we have encountered repeatedly in this book, serves as a fine example. Recall the case of a statue and the lump of clay that composes it. According to a popular argument, even though the statue depends on the lump of clay for its existence, the statue and the lump are not numerically identical because they have different persistence conditions.10 The lump of clay cannot survive the removal of a chunk of its constituent matter, while the statue can survive the loss of some chunk of it, such as an arm – we still regard the ancient Greek statues as statues even when they are evidently missing some pieces, like arms and noses. So the statue can depend for its existence – by metaphysical necessity – on the lump, without the statue and the lump being identical. A metaphysician might analyse the situation as follows: A statue is essentially statue-shaped because it falls under the sortal statue. So the statue cannot survive, say, being remoulded into a pot. This suggests that objects such as a lump of clay and a statue can share all their material properties and their space-time region. But the ‘common sense’ intuition is that two things with the same properties in the same place are surely identical. The metaphysician is then faced with the task of reconciling common sense – which a great many philosophers wish to vindicate – with the metaphysical upshot of the above analysis. None of this, of course, has much to do with science, but is rather based on common-sense intuitions and a priori reasoning (cf. Chapter 8 for further discussion of intuitions and a priori inquiry). In the teeth of those philosophers who are more sceptical about autonomous metaphysics, this type of philosophizing struggles to earn merit: [Autonomous] metaphysics proceeds by attempts to construct theories that are intuitive, common-sensical, palatable, and philosophically respectable. The criteria of adequacy for metaphysical systems have clearly come apart from anything to do with the truth. Rather they are internal and peculiar to philosophy, they are semi-aesthetic, and they have more in common with the virtues of story-writing than with science.11

This statement is obviously polemical, but it does raise a valid concern: without a clearer picture of what counts as a good metaphysical 10

11

See for instance L. A. Paul, ‘The Context of Essence,’ Australasian Journal of Philosophy 82 (2004), pp. 170–184. Ladyman and Ross, Every Thing Must Go, p. 13.

9.2 Fully naturalistic metaphysics

system, autonomous metaphysics may lose its claim to truth. Attempts to establish better criteria for assessing metaphysical systems have of course been made, and we have already discussed some of these attempts, such as the nature of a priori inquiry. But it might seem easier and more sensible to just give up on a completely autonomous metaphysics and instead attempt to ‘naturalize’ metaphysics. Let us now give the stage to those who think that this is the only way to salvage anything of metaphysics – if the resulting ‘scientism’ is to be even called ‘metaphysics’. 9.2 Fully naturalistic metaphysics

The view that we will now consider could be considered to go as far as rejecting metaphysics altogether. The most influential defence of this type of ‘scientism’ is no doubt Bas van Fraassen’s, so we will take his view as a prominent example. 12 Van Fraassen calls his view ‘the empirical stance’. The term ‘scientism’ is often used in a pejorative sense, but in recent literature it has been adopted by proponents of a more scientifically-minded metaphysics. 13 The core of van Fraassen’s approach can be summarized with the following three challenges to traditional metaphysics. First, van Fraassen argues that th e remoteness of metaphysical questions from empirical considerations makes them useless, even if not quite meaningless in Carnap’s sense. Science und ergoes a constant testing and even falsification, but metaphysics claims to seek the truth. However, metaphysics is arguably never able to reach the position of establishing whether anything it says is actually true or false. Empirical science can justify and ground its own method in virtue of its practical relevance. Metaphysics, instead, turns out to be a merely formal or conceptual exercise, without practical or scientific relevance. This is the upshot of van Fraassen’s first critical point. Secondly, van Fraassen argues that metaphysical questions are irredeemably context-dependent and hence lack well-defined ‘answering strategies’. One example of this is the question ‘Does the world exist?’, which van Fraassen dismisses as extremely simple, pointing out that

12 13

Bas C. van Fraassen, The Empirical Stance (New Haven, CT: Yale University Press, 2002). Ladyman and Ross, Every Thing Must Go, Ch. 1.

211

212

Demarcating metaphysics and science: can metaphysics be naturalized?

metaphysicians have nevertheless spent a lot of time on it. This type of epistemic challenge is more or less what the previous chapter was attempting to address. Thirdly, van Fraassen objects that metaphysics accounts for ‘what we initially understand [in terms of …] something hardly anyone understands’. 14 This undermines the explanatory value of metaphysics; the contrast is especially striking when compared to science. The upshot is thoroughly eliminative : we really ought to eliminate it from the discourse altogether. Note that all this is coming from a scientific anti-realist . Van Fraassen’s position is that while science may seek the truth, its primary goal is to establish results of some practical or instrumental importance – something that will be of use. In other wor ds, in science it is not a catastrophic result if a theory turns out to be false, as long as it was of some use. And simply the fact that a scientific theory survived to the stage where it got to be falsified suggests that it has some value, some heuristic use. Consider Newton’s impor tant work in physics. We know that Newtonian physics is not exactly true, and it produced (and in fact continues to produce!) many false beliefs. But there seems to be no question about the fact that Newton’s work (in physics, but presumably not in alchemy) was extremely useful if not necessary for scientific progress. By contrast, many metaphysical theories do not even have the merit of being falsifiable , so it turns out that metaphysics is completely detached from the scientific model. A metaphysician, it seems, typically cannot claim any pragmatic or cognitive utility for her falsified theory; although van Fraassen himself emphasizes empirical adequacyrather than pragmatic utility:this is a requirement for a certain type of isomorphism between what a theory says about observable phenomena and the model (a family of structures) that the theory describes. Van Fraassen points out that the failure of metaphysics to be properly falsifiable is particularly problematic for those contemporary metaphysicians who strive to uphold some version of naturalism (which is most of them). But does van Fraassen think that there is anything left for philosophy to do after this devastating deconstruction? If there is, one thing is clear:philosophy should not be based on dogma. Instead, van Fraassen thinks that philosophy, or 14

Van Fraassen, The Empirical Stance, p. 3.

9.2 Fully naturalistic metaphysics

philosophical views, should be conceived asstances: ‘A philosophical position can consist in a stance (attitude, commitment, approach, a cluster of such– possibly including some propositional attitudes suchas beliefs as well)’.15 We can see that to conceive of philosophical views as stances is not easily combined with the type of metaphysical realism that we have assumed to be the default position throughout much of this book. Or, to put it another way, van Fraassen’s proposal would lead us to reject what James Ladyman and Don Ross call ‘strong metaphysics’, which is more or less the type of metaphysics that all non-deflationary metaphysicians pursue.16 This is because a stance appears to have a strongly subjective element; it is an attitude rather than a doctrine. We will look into the ‘weak metaphysics’ that Ladyman and Ross propose themselves in a moment, but let us first give van Fraassen a chance to persuade us to ‘convert’ to his preferred stance, the empirical stance. One might think that ‘convert’ is inappropriate here, suggesting that the view is more like a religion. But in fact this is exactly how van Fraassen puts it: ‘Being or becoming an empiricist will then be similar or analogous to conversion to a cause, a religion, an ideology, to capitalism or to socialism, to a worldview’.17 Given this, it’s clear that van Fraassen does not aim to lay out the empirical stance in terms of philosophical arguments like the ones we have been discussing so far. Instead, the empirical stance is primarily defined by a negative attitude: rule out a dogmatic attitude, reject explanation for the sake of explanation – that is, where no testable, (empirical) evidence-based explanation is forthcoming – and do not postulate explanations just for the sake of psychological satisfaction. While this may seem like a healthy attitude, it does bear a risk: if we do not postulate anything, then we may lack the motivation to develop new empirical methods as well. In other words, sometimes postulating even a false explanation could have a theoretical utility. The positive part of the empirical stance is a simple endorsement of scientific virtues, namely that we should take a view only as seriously as evidence warrants. Let us pause here for a moment, for many readers will no doubt be familiar with examples from science which are not fully compatible with the empirical stance. A good example is string theory:

15 16 17

Ibid., pp. 47–48. Ladyman and Ross, Every Thing Must Go, p. 60. Van Fraassen, The Empirical Stance, p. 61.

213

214

Demarcating metaphysics and science: can metaphysics be naturalized?

Depending on who one listens to, string theory has either already led us a considerable distance down the road to a complete theory of quantum gravity, or it has achieved absolutely nothing that counts as physics rather than mathematics.18

The basic problem with string theory is that it has made little in terms of testable predictions – exactly the same issue that people like van Fraassen find problematic when it comes metaphysics. Now, one reaction to this is to consider string theory just as ‘bad’ as ‘strong metaphysics’. Indeed, the physicist Lee Smolin reacts against string theory as strongly as van Fraassen does against metaphysics; Smolin is especially troubled by the fact that string theory has received vast amounts of funding in the last few decades, at the cost of leav19 We do not need to ing research on alternative theories largely unfunded. dwell on the politics of thesituation, but it is worth noting that thecelebrated empirical sciences are apparently at least sometimes subject to problems similar to those encountered in metaphysics, according to van Fraassen. How could a metaphysician with less eliminative tendencies reply to van Fraassen’s challenge? Recall that one of van Fraassen’s points is that scientific theories are ‘initially understood’. This flies in the face of a point that was briefly entertained in Chapter 7: a scientific theory cannot be ‘understood’ unless it is interpreted, and the metaphysician may insist that interpretation requires tools coming from outside of science: the interpretation of any theory T cannot be provided within T, for otherwise it should itself be interpreted, giving rise to a vicious regress (here T can be considered science as a whole, on a par to the empirical stance). When it comes to areas like quantum mechanics, even most physicists would acknowledge that philosophical questions cannot be avoided in the interpretation 20

of the relevant theories. Before those tools are applied, at most one has the sort of instrumental ability that we saw van Fraassen to push earlier. Indeed, these instrumental abilities may be satisfactory to scientists insofar as their immediate goals are considered (to produce practical applications,

18 19

20

Ladyman and Ross, Every Thing Must Go, p. 168. For a popular account, see Lee Smolin, The Trouble with Physics (London: Penguin Books, 2006). For further discussion, see Craig Callender, ‘Philosophy of Science and Metaphysics,’ in S. French and J. Saatsi (eds.), The Continuum Companion to the Philosophy of Science (London: Continuum, 2011), pp. 33–54. One good example might be the ’many worlds’

9.2 Fully naturalistic metaphysics

to formulate and test theories), but cannot be satisfactory insofar as science must be understood as mapping the way the world is. Now, if metaphysics turns out to be necessary for interpreting scientific theories, it also appears plausible that the concepts and categories typical of metaphysics – the notions contained in the metaphysical vocabulary that we use for interpreting scientific theories – are not (necessarily) obscure. As a matter of fact, since metaphysical vocabulary has been and is used to interpret scientific theories quite successfully, the argument from obscurity could be resisted with reference to examples, such as, say, the significant philosophical work that has gone into interpreting general relativity.21 Moreover, metaphysical analysis appears in this respect closer to the common-sense approach than its scientific counterpart. For instance, is the notion of a universal, say, any more obscure than that of a Higgs boson – taking into account the fact that the recent ‘discovery’ of the boson in question relies on certain somewhat arbitrary criteria for ‘discovery’, commonly used in particle physics?22 The answer is by no means obviously affirmative unless one equates clarity with measurability – measurability to a certain degree, that is. But, again, this is not what one normally intends by ‘understandability’ and ‘clarity’ – not even in an openly anti-realist context such as van Fraassen’s. If concepts such as substance, relation, property, and the like are at least no less understandable than scientific concepts – and are in fact employed to interpret scientific theories – then van Fraassen’s challenge could perhaps be addressed.

interpretation of quantum mechanics, see Simon Saunders, Jonathan Barrett, Adrian Kent, and David Wallace (eds.), Many Worlds? Everett, Quantum Theory, and Reality (Oxford 21

22

University Press, 2010). For a survey, see Thomas A. Ryckman, ‘Early Philosophical Interpretations of General Relativity,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2014 edn); see http://plato.stanford.edu/archives/spr2014/entries/genrel-early/ Particle physics has an accepted definition for what is sufficient to claim a discovery: 5.0 sigma significance, i.e. a level of certainty up to five standard deviations. In statistical terms, this means a probability of less than one in a million that the observed phenomenon was produced by something other than the postulated Higgs, namely statistical fluctuation. The reason for this type of talk is of course that the Higgs cannot be observed directly. Rather, we observe decay products, such as photons, which could be produced by a number of phenomena that have to be ruled out. This also makes it quite clear that there is always an aspect of fallibility in these types of results.

215

216

Demarcating metaphysics and science: can metaphysics be naturalized?

Note that the type of metaphysics that van Fraassen is primarily reacting against is ‘Quinean’, or at least that is how he sees the project: ‘The genuflection toward science among Quine’s heirs has all too often been toward a naive caricature purveyed by the past generation of philosophers at whose knees Quine himself learned it’.23 Accordingly, he does not discuss or analyse the various more fine-grained formulations of metaphysics and ontology that we have been pursuing in this book. In fact, a proponent of ‘strong metaphysics’ might even concede van Fraassen’s point if it is applied to those who attempt to strengthen the position of metaphysics with an appeal to its continuity with science – E. J. Lowe for one certainly does concede the point. 24 Indeed, Lowe suggests that van Fraassen is absolutely right when it comes to the assertion that metaphysicians claim continuity with science by appealing to a shared met hodology, namely inference to the best explanation. Van Fraassen’s point, as we have seen, is that metaphysics fails, catastrophically, to uphold the standards of science in this regard. But this critique only goes through if metaphysics is indeed conceived as continuous with science in this sense, something that Lowe ascribes to ‘false friends’ of metaphysics: [T]he fault lies here not with metaphysics as such, properly conceived, but only with those of its false friends who mistakenly seek to enhance its credit by assimilating its task to that of empirical science. Metaphysics and empirical science are not ‘continuous’ with each other in any sense which implies that they have the same goals and methods, or that metaphysics is just the extension of empirical science to questions of greater generality than any that are addressed by the so-called ‘special’ sciences. Rather, when both are conducted fruitfully, metaphysics and empirical science exist in a symbiotic relationship, in which each complements the other.25

The somewhat surprising upshot, then, is that metaphysics, ‘properly conceived’, as Lowe puts it, can not only withstand van Fraassen’s challenge, but can also agree that it has correctly identified a fatal flaw in one popular approach to metaphysics. Certainly, Lowe and other proponents of ‘serious metaphysics’ owe us a story about how metaphysics and science are related, if they’re not continuous, but in the above passage we already have 23 24 25

Van Fraassen, The Empirical Stance, p. 11. E. J. Lowe, ‘The Rationality of Metaphysics,’ Synthese 178 (2011), p. 101. Lowe, ‘The Rationality of Metaphysics,’ pp. 101–102.

9.3 The Principle of Naturalistic Closure and the Primacy of Physics

a hint about what the answer might be. We will return to this issue in the fourth section of this chapter, but first we will consider another scientifically minded view, which attempts to provide a ‘naturalized’ conception of metaphysics. 9.3 The Principle of Naturalistic Closure and the Primacy of Physics

The two principles in the title of this section come from Ladyman and Ross. 26 They have been discussed extensively in recent years and it would be impossible to do full justice to this discussion here. We will instead focus on certain important aspects of the debate, relevant especially from the point of view of metametaphysics. It has already become clear that Ladyman and Ross share some aspects of van Fraassen’s criticism of metaphysics; the polemical first chapter of their book has spurred a remarkable reaction not only from metaphysicians, but also from philosophers of science. Since we ended the previous section with a quote from Lowe, we might start this one by considering a challenge that Ladyman and Ross pose for him and anyone else who pursues what has been pejoratively labelled ‘neo-scholastic metaphysics’. Crucially, Ladyman and Ross attempt to generalize their critique of contemporary metaphysics beyond the broadly ‘Quinean’ picture, with which van Fraassen seems to be primarily concerned. For instance, they attempt to challenge metaphysicians’ use of intuitions and a priori inquiry. However, as we saw in Chapter 8, these issues are fairly complicated. Ladyman and Ross target one admittedly popular approach, according to which metaphysical a priori inquiry is more or less just conceptual analysis. They appear to suggest that Lowe and indeed most contemporary metaphysicians adopt ‘the familiar methodology of reflecting on our concepts (conceptual analysis)’ for doing metaphysics. 27 Ladyman and Ross are quite correct to ask how conceptual analysis could possibly reveal anything about the structure of reality, but this is (again!) a concern that many metaphysicians share – Lowe among them. So perhaps this critiquing aspect of the

26 27

Ladyman and Ross, Every Thing Must Go, Ch. 1. Ibid., p. 16.

217

218

Demarcating metaphysics and science: can metaphysics be naturalized?

project to naturalize metaphysics is not as fine-grained and charitable as one might hope. For Lowe, metaphysics is the study of possibility – the realm of a priori inquiry is metaphysical modality. On this view, metaphysics studies what sorts of things there are in the world (fundamentally) and how they are related, how reality is structured. The relevant method is a priori inquiry into what sort of things are possible, followed by a partly empirical study regarding which of those things are possible together. Only on this basis, argues Lowe, can we determine what is actually the case. Ladyman and Ross are aware of this more fine-grained type of approach, but they are suspicious of the modal element: they deny that a priori inquiry can reveal what is metaphysically possible. Indeed, we have seen in earlier chapters that much work remains to be done in the area of modal epistemology, which appears to be absolutely central for the type of view under consideration now. But evidently some promising work in this area does exist. We have discussed it extensively, but Ladyman and Ross make no real effort to engage with it. They do however point out that metaphysicians have often erred in their modal judgements, for instance with regard to non-Euclidean geometry, deterministic causation, non-absolute time, and so on. This is of course true, but hardly surprising: no one has suggested that modal inquiry is easy or infallible. While the infallibility of a priori inquiry may have been a doctrine of Cartesian metaphysics, it is most certainly not a doctrine of contemporary analytic metaphysics, as we clearly observed in the previous chapter . Hence, it is indeed true that metaphysicians, like scientists, make mistakes. Kant may have held that non-Euclidean geometry is impossible, but physics soon showed that not only is it possible, but actual. The lesson that we should take from this is that a more charitable picture of contemporary metaphysics ought to acknowledge the possibility of revision – not unlike in the case of empirical results, which can be revised in the light of new empirical data. But regardless of the negative project put forward by Ladyman and Ross (as well as many other ‘scientifically-minded’ philosophers), they also propose an extremely interesting and influential methodological framework – we may call it the ‘scientistic stance’, which is a label they adopt with reference to van Fraassen’s idea of stances. 28 28

Ibid., p. 64.

9.3 The Principle of Naturalistic Closure and the Primacy of Physics

The core of the scientistic stance is the search for unification among scientific theories on the basis of physics. In order to achieve such unification, we should finally outline the two principles mentioned in the title of this section: the Principle of Naturalistic Closure (PNC) and the Primary of Physics Constraint (PPC). Here are short formulations of both principles: (PNC): Any new metaphysical claim that is to be taken seriously at time t should be motivated by, and only by, the service it would perform, if true, in showing how two or more specific scientific hypotheses, at least one of which is drawn from fundamental physics, jointly explain more than the sum of what is explained by the two hypotheses taken separately. 29 (PPC): Special science hypotheses that conflict with fundamental physics, or such consensus as there is in fundamental physics, should be rejected for that reason alone. Fundamental physical hypotheses are not symmetrically hostage to the conclusions of the special sciences. This, we claim, is a regulative principle in current science, and it should be 30

respected by naturalistic metaphysicians. These principles are meant to capture the negative project – to (partly) justify the critical remarks we’ve discussed above – as well as ground the positive project. It is the first principle especially that has received critical attention, but both principles have certainly been challenged. Ladyman and Ross are careful to note that they do not wish to go as far as the logical positivists did with their verificationism – the view that only empirically verifiable statements are meaningful. The logical positivists cast metaphysics (among other things) aside altogether and did not even attempt to apply their principles to metaphysics, in order to salvage something from it. In contrast, Ladyman and Ross believe that the above principles can be applied to metaphysics, to produce a fully naturalistic picture. There is certainly some plausibility in prioritizing science like this. This much is made clear by the various examples involving ‘pseudo-science’ discussed by Ladyman and Ross. Much of the strength of the ‘pseudo-science’ critique relies on the claim that from the point of view of the layman (and perhaps the typical philosopher) fundamental physics is commonly viewed as a network of ‘microbangings’. Indeed, talk of ‘particle’ physics may suggest that fundamental physics deals with solid particles and applies Newtonian mechanics 29

Ibid., p. 37.

30

Ibid., p. 44.

219

220

Demarcating metaphysics and science: can metaphysics be naturalized?

to them, but this is quite far from reality. Nature does not play billiards! But the question is whether this is enough to fully motivate principles like (PNC) and (PPC). Take (PNC) first. It would seem to reduce metaphysics merely to the task of unifying scientific hypotheses. Many metaphysicians would think (and have thought) that this misconstrues the methodology of metaphysics. For instance, in a review of Ladyman and Ross’s book, Cian Dorr puts it as follows: The whole approach of Every Thing Must Go reflects an exaggerated sense of the importance of argument in metaphysics, and a corresponding underestimation of the difficulty of merely crafting a view coherent and explicit enough for arguments to get any grip.31

The question arises: is it appropriate to evaluate metaphysical argumentation with the standards of natural science, as Ladyman and Ross appear to do? Insofar as the methodologies of science and metaphysics are different, it would seem to be inappropriate to apply the same standards to them. Moreover, there are specific metaphysical questions which do not seem to be hostage to science in the manner suggested by (PNC). For instance, Katherine Hawley has argued that questions regarding the philosophy of time can hardly be settled purely on the basis of physics; her case study is the relationship between the special theory of relativity (STR) and presentism, the view that only what is present exists. 32 According to STR, there are different frames of reference, which influence our answer when faced with questions about the simultaneity of two distant events. But because STR does not privilege any one frame of reference over another, it also does not dictate whether or not two events occur simultaneously. Since those philosophers of time who defend presentism claim that only what is present exists, they must say that only those events that are simultaneous with the present exist. Yet, STR suggests that this question is open in a way in which the question of what exists is surely not open. And here we would seem to have an automatic refutation of the metaphysical theory of presentism based on science – a nail in the coffin of

31

32

Cian Dorr, ‘Review of Every Thing Must Go,’ Notre Dame Philosophical Reviews (2010); see https://ndpr.nd.edu/news/24377-every-thing-must-go-metaphysics-naturalized/ Katherine Hawley, ‘Science as a Guide to Metaphysics?’, Synthese 149.3 (2006): 451–470.

9.3 The Principle of Naturalistic Closure and the Primacy of Physics

autonomous metaphysics. But there are various ways in which presentists could attempt to respond (and have responded). Some of them concern the interpretation of STR, but there are other strategies as well. Hawley mentions three possible reactions: First, we might accept that presentness is frame-dependent, accept that existence cannot be frame-dependent, and thus reject presentism. Second, we might accept that presentness is frame-dependent, insist that only what is present exists, and thus conclude that existence is frame-dependent. This second option is too relativistic (in a bad way) for almost everyone, and may undermine the intuitions that lead initially to presentism. Third, we might accept that existence cannot be frame-dependent, insist on the truth of presentism, and thus conclude that there is a privileged frame of reference, one which escapes notice in STR. Simultaneity in that privileged frame is absolute simultaneity, and events absolutely simultaneous with my typing now are absolutely present. Positing a privileged frame of reference does not compel us to adopt presentism, for we might argue that what is absolutely past and future is also real. But the third option permits us to be presentists without conceding that existence is frame-dependent. 33

So it appears that depending on other theoretical choices, STR is in fact compatible with various views about the nature of time, including presentism. However, it should be noted that Einstein specifically developed STR to account for the absence of a privileged frame, so the onus would appear to be on the presentists. Moreover, STR does, of course, entail some constraints on what we can do in the metaphysics of time, but it does not appear to immediately solve the metaphysical puzzle about time – there is some room for interpretation regarding frame-dependence. Take another example: in Chapter 6 we saw that input from physics is surely needed if we hope to make progress regarding the debate about a fundamental level, but what if the level is beyond the reach of all physically possible measuring devices? For instance, if some of the fundamental particles suggested by the Standard Model do in fact have internal structure, the amount of energy needed might be so vast that it would be physically impossible to build a particle accelerator large enough to ever find out about the constituents of these particles. Ladyman and Ross make 33

Hawley, ‘Science as a Guide to Metaphysics?’, pp. 465–466.

221

222

Demarcating metaphysics and science: can metaphysics be naturalized?

several claims that have the appearance of these types of metaphysical issues, which we have been discussing throughout this book. For example, they argue that ‘we should not interpret science – either fundamental physics or special sciences – as metaphysically committed to the existence of self-subsistent individuals’.34 Now, this is an interesting, substantive claim, but without resorting to the full metaphysical ‘tool box’ that we partially laid out in Chapters 5 and 6, it is quite difficult to make sense of it. Indeed, it is not even clear that this claim can be reconciled with (PNC), for without further analysis its unificatory power remains questionable. Hence, one might ask, how different is this from ‘neo-scholastic’ metaphysics after all? A final example, which we will not consider in detail: quantum mechanics seems to have all sorts of implications towards the metaphysics of identity and individuation, but it may not help us to decide which metaphysical theory of identity is ultimately the correct one.35 This is because of the possibility of underdetermination of metaphysics by physics; that is, the empirical evidence we have is compatible with a range of metaphysical views and hence does not help in deciding between them. 36 This, according to Hawley, appears to be what has happened in the case of presentism and the special theory of relativity. With this spirited reaction in defence of metaphysics out of the way, it must be granted that Ladyman and Ross have done important work to deconstruct arguments in metaphysics that seem to be backed by little more than ‘pseudoscience’. In this regard, their second principle, (PPC), is initially easier to accept. Note however that the requirement of ‘consensus’ in the second principle is likely to cause controversy and many would insist that the principle should be supplemented with a fallibilist attitude towards such a ‘consensus’. Obviously, we need to be careful with examples drawing from science and even more careful if input from metaphysics to

34 35

36

Ladyman and Ross, Every Thing Must Go, p. 119. For details, see for instance Mauro Dorato and Matteo Morganti, ‘Grades of Individuality. A Pluralistic View of Identity in Quantum Mechanics and in the Sciences,’ Philosophical Studies 163 (2013), pp. 591–610. For further discussion on underdetermination, see Steven French, ‘Metaphysical Underdetermination: Why Worry?’, Synthese (2011) 180, pp. 205–221; see also Steven French, The Structure of the World:Metaphysics and Representation (Oxford University Press, 2014), especially Ch. 2.

9.3 The Principle of Naturalistic Closure and the Primacy of Physics

science is proposed. But the cost of accepting both of the principles proposed is that metaphysics becomes simply a ‘handmaiden’ of science. Note also that we have omitted entirely a discussion of the positive proposal put forward by Ladyman and Ross, a view they call ‘ontic structural realism’ (OSR). We encountered the structuralist view briefly in Chapter 6, but have not properly engaged with it. This is partly because the details of the theory are less important from a metametaphysical point of view than from the point of view of first-order metaphysics or philosophy of science, where (OSR) is an important contender, and partly because we could not do justice to (OSR) here.37 We might note, though, that this ontological picture has been challenged, among other things, on the basis of an oversimplistic picture regarding ontological dependence: one version of (OSR) suggests that objects depend for their identities on the structures to which they belong, so the structure is more fundamental than the objects – hence the title of Ladyman and Ross’s book, Every Thing Must Go. Now, this looks like a claim of identity-dependence, but as we’ve seen in the discussion in Chapters 5 and 6, there are several difficult questions to address before we can accurately formulate and use this notion of dependence – a task which Ladyman and Ross do not undertake.38 Setting the details of the ontological picture aside, there is another important methodological challenge that can be derived from the two principles, (PNC) and (PPC). This concerns a core area of metaphysics – one that we discussed in considerable detail in Chapter 7 and also earlier – namely modality. We already mentioned the challenge regarding modal epistemology, namely that philosophers have often wrongly regarded something as possible or necessary. More generally, one might worry that it is unclear what the conceptual space that metaphysics is supposedly concerned with is like, and it is sensible to think that it is ultimately physical modality that determines what we regard as metaphysically possible, necessary, or impossible. In other words, following (PPC), a scientifically minded philosopher could argue that all modality reduces to physical modality and hence any judgements that we make about what is possible or necessary should have physics – or science more generally – as their source. 37

38

For a classic overview of structural realism, see James Ladyman, ‘What is Structural Realism?’, Studies in History and Philosophy of Science Part A 29 (1998), pp. 409–424. For further discussion, see, e.g., Donnchadh O’Conaill, ‘Ontic Structural Realism and Concrete Objects,’ The Philosophical Quarterly 64.255 (2014), pp. 284–300.

223

224

Demarcating metaphysics and science: can metaphysics be naturalized?

Craig Callender has raised a challenge not unlike this one, pointing out that the focus on metaphysical modality is suspicious precisely because metaphysicians have claimed that this is an area on which science can say very little.39 Callender thinks that the focus on metaphysical modality is one of the primary reasons for the disconnect (or even hostility!) between contemporary metaphysics and philosophy of science. The rift could be put as follows. Science concerns the actual world and makes claims about what is physically possible and necessary, but metaphysicians want something stronger: they want to know what is metaphysically necessary, say, whether the actual laws of physics hold across all possible worlds. Indeed, the debate about the laws of nature is one of the central topics of contemporary analytic metaphysics and it is perhaps not surprising that those philosophers who have scientific training or interests find this debate frustrating. Who cares about whether salt dissolves in water in some remote possible world!40 Well, as Callender points out, this is not really the right way to think about the matter: I submit that the basic problem with some metaphysics today is the idea that the philosopher and scientist doing ontology are performing fundamentally different and separate jobs. The metaphysician’s picture that the scientist works in the lab, discovering the actual world’s features, while the metaphysician discerns the wider universe of possible, isn’t right. The error is thinking that the science of the actual world doesn’t affect what one thinks is possible or impossible. The history of science and philosophy amply displays that what we think is possible or impossible hangs on science. Or going in the other direction, the error is thinking that modal intuitions are reliable if they are not connected to a systematic theory of a large domain, one possessing many theoretical and empirical virtues.41

39

40

41

Callender, ‘Philosophy of Science and Metaphysics,’ p. 40. Similar concerns are raised in Maudlin, The Metaphysics Within Physics. For discussion regarding laws of nature and this very example, see for instance Alexander Bird, Nature’s Metaphysics (Oxford University Press, 2007). See also Tuomas E. Tahko, ‘The Modal Status of Laws: In Defence of a Hybrid View,’ The Philosophical Quarterly (2015 [online], forthcoming). Callender, ‘Philosophy of Science and Metaphysics,’ p. 43–44.

9.4 Methodological similarities

This is, by all accounts, a fair assessment, one that the present author would readily agree with. But there is one more issue worth mentioning before we move on, having to do with the relevant interpretation of modal claims. While Callender is correct in that the exclusive focus on irreducible metaphysical modality is suspicious, the thought that the modal space that scientists deal with is somehow absolutely well defined does not follow. Even a proponent of ‘strong metaphysics’ need not commit to the irreducibility of metaphysical modality. Indeed, as we saw in Chapter 7, on the essentialist framework of Fine and Lowe metaphysical modality reduces to essence, albeit it is unlikely that a scientifically minded philosopher would find the appeal to irreducible essence any more satisfying. Nevertheless, the point is that (semi-)autonomous metaphysics does not need to commit to any single piece of doctrine, such as irreducible metaphysical modality. The scientifically minded philosopher only needs to insist on the necessity of the conceptual and methodological ‘toolbox’ of metaphysics for performing certain tasks, such as the interpretive task we have now encountered several times. Note that this immediately makes room for a more moderate form of naturalism about metaphysic s – something that we will discuss in more detail in the last section of this chapter. But first, let us take a look at the potential methodological similarities between science and metaphysics, for it seems that Callender is calling exactly for a closer study of such similarities, instead of focusing on the evident differences. 9.4 Methodological similarities

Whilst rushing through the various comments and critical remarks concerning the relationship between science and metaphysics we have, hopefully, made at least some progress. There is clearly quite a bit of chatter between scientists (or rather, ‘scientifically-minded’ philosophers) and metaphysicians and despite some extremists, there is a rough consensus that the two cannot be entirely independent of each other. But if there is some overlap between the disciplines, we still ought to ask: does this overlap concern the methods or the subject matter of science and metaphysics? One recent and potentially helpful discussion on just this issue is due to L. A. Paul, who suggests that metaphysics and science have effectively the

225

226

Demarcating metaphysics and science: can metaphysics be naturalized?

same methodology, but distinct subject matters.42 The core of Paul’s proposal can be summarized with the following passages: The ontological account describes the metaphysically prior categories and constituents of the physically fundamental entities, and in this sense describes features of the world that are more fundamental than those of natural science.43

This proposal takes advantage of the metaphysical ‘toolbox’ that we are by now quite familiar with: Paul suggests that the subject matter of metaphysics is ontologically prior to that of science. She thinks that this ontological priority is reflected by conceptual priority: The fact that the subject matter of metaphysics can be ontologically prior to the subject matter of science is reflected in the fact that many concepts of metaphysics are conceptually prior to the concepts of science. […] There is no way to make sense of the central concepts of classical field theory or quantum chromodynamics without using a concept of property.44

Note that Paul is making a strong claim here: she appears to suggest that we cannot even understand or interpret science without resorting to notions which are distinctly philosophical, or require a philosophical analysis. It is indeed relatively uncontroversial that the notion of a property is prior to that of, say, electric charge: electric charge is a specific case of property. In this sense, metaphysical concepts do appear more fundamental than scientific concepts. But why should we think that this priority holds quite generally? Moreover, this in itself may not be sufficient to establish that metaphysical inquiry has a ‘privileged’ role to play when it comes to studying the fundamental structure of reality, as one could interpret Paul to be suggesting. In general, one might have doubts about whether conceptual priority is a convincing criterion or indication of ontological priority. There are at least two reasons to doubt that there is such a strong link between them. First, it is often very difficult to even determine when one concept is prior to another – the order of the acquisition of concepts being of little help. Secondly, even if a specific sense of priority and dependence among 42

43 44

L. A. Paul, ‘Metaphysics as Modeling: The Handmaiden’s Tale,’Philosophical Studies 160 (2012), pp. 1–29. Paul, ‘Metaphysics as Modeling: The Handmaiden’s Tale,’p. 5. Ibid., p. 6.

9.4 Methodological similarities

concepts is granted, work is required to establish which of the seemingly more fundamental linguistic or conceptual categories latch onto objective structures of reality. Paul does however have something more to add in this regard: she seems to have in mind an explanatory link, broadly understood. Given that we can formulate this type of link more precisely by using the ‘toolbox’ of metaphysics, this type of explanatory link could perhaps be described in terms of grounding or explanatory dependence, producing the type of ontological priority that Paul was looking for. However, it appears that we would need something more than this. The mere fact that we use certain concepts in scientific theories and explanations does not directly entail that we should be ontologically committed to their existence in anything more than a heuristic sense. The reason for this is that certain presuppositions concerning the ontological structure of reality are likely to be at work in every explanation; and different presuppositions might well lead to different explanations and different ontological commitments . This clearly blocks the inference at hand, from the use of certain concepts in explanation to the objective existence of those ontological elements. In other words, what guarantees that the relevant concepts accurately reflect the structure of reality? This is a more difficult question to answer, so let us set it aside here in order to focus on another important aspect of Paul’s proposal, having to do with the methodology of metaphysics. Recall that Paul considers metaphysics and science to have, by and large, the same methodology, but distinct subject matters. We have now seen that the case concerning subject matter is far from settled, but what about the case of methodology? At the core of Paul’s account is the claim that both science and metaphysics rely on modelling. Specifically, Paul suggests that – as they would appear to be for scientists – models are the metaphysician’s primary tool of theory-forming. The categories of entities involved in metaphysical models are different from those used in science (e.g., one would talk of properties or substances rather than of particles or genes), but metaphysics and science both model parts of reality nonetheless. Both science and metaphysics use a priori reasoning to infer to the best explanation, which helps us choose between empirically equivalent models. Now, we’ve already noted that the method of inference to the best explanation may be problematic, but let us set aside those worries for the moment. Additionally, Paul suggests that both in science and in metaphysics the

227

228

Demarcating metaphysics and science: can metaphysics be naturalized?

usual theoretical virtues of simplicity, ontological parsimony, elegance, explanatory power, and fertility may be used to evaluate models. In Paul’s view, then, metaphysics and science are primarily demarcated by their subject matter, not their methodology. The idea of metaphysics as modelling is initially attractive, but in order for it to work in the area of metaphysics, we need some account of how to choose between empirically equivalent models. In scientific contexts, models can typically be tested in a relatively straightforward manner. Not so in metaphysics. So how do we assess models? How do we decide which of two models is better if they differ in no way with regard to the empirical predictions that they make – if they even make any? Paul’s strategy appeals to inference to the best explanation based on quantification over the best available theories – somewhat reminiscent of Quine (although Paul does not mention Quine). This viewpoint is, as we have seen, problematic: Does it provide a reliable link between one’s metaphysical theory and the actual structure of reality, or merely a pragmatic criterion? In general, one could attempt to apply the above-mentioned theoretical virtues to justify model selection. However, it is not obvious that, say, simpler or more elegant scientific theories are more likely to be true, even if it were the case that they are more explanatory, easier to understand, or pragmatically valuable. These are all issues of intense dispute in the philosophy of science. In science, however, empirical data can be brought to bear on the evaluation of one’s conjectures – something to which metaphysics cannot or at any rate typically does not have access in its own right. This naturally represents a potential point of fundamental demarcation between science and metaphysics. Indeed, Ladyman immediately replies, re-emphasizing something that we have already come across: since in metaphysics truth is all that matters (metaphysics has no practical applications) but theoretical virtues are not obviously truth-conducive, the fact that metaphysical models can only be assessed on the basis of their theoretical virtues suggests that the proposed methodology is under threat.45 Even if this issue concerning empirically equivalent models could be settled, there are aspects of the ‘metaphysics as modelling’ suggestion that we ought to discuss in more detail. For this purpose, Paul has helpfully 45

James Ladyman, ‘Science, Metaphysics and Method,’ Philosophical Studies 160.1 (2012), pp. 31–51.

9.4 Methodological similarities

provided us with a number of examples on how exactly this method can be used and applied. Thought experiments are one particularly promising avenue in this regard. For the process of modelling, two further tools are of importance: abstraction and idealization. For instance, when examining the connection between two distinct events, Socrates’s death and his drinking hemlock, we may abstract away from the complex details of the actual world in order to identify a counterfactual dependence relation between these events. Paul suggests that by abstracting away superfluous details, we can conclude that such counterfactual dependence is sufficient for causation. Hence, Socrates’s death was caused by his drinking the hemlock. The process of abstraction is important, Paul says, because in actuality superfluous features may ‘muck up’ the real-world facts relating to dependence. However, the model we get by abstraction is sufficiently isomorphic to the real-world case for a reliable connection, or so the story goes. If this were correct, the suggested parallel between science and metaphysics would indeed hold: we get to identify the fundamental features of reality in both cases via abstraction and idealization. This account seems promising, but it can certainly be challenged. For it relies on an underlying account regarding the role of counterfactuals in metaphysical modelling. In Chapter 7 we have already pointed out certain problematic aspects of this account, with regard to Timothy Williamson’s theory of modal epistemology.46 And Paul makes use of Williamson’s account explicitly.47 Here is a brief recap of the problem we discussed earlier. The account does not give us reliable means to explain the background knowledge, Williamson’s ‘constitutive facts’, needed to get the account started. Constitutive facts are things that we must hold fixed across metaphysically possible counterfactual scenarios, and knowledge of such facts is needed to secure traditional examples of metaphysical necessities. But how is one to determine which facts are constitutive? If one’s evaluation of counterfactual scenarios depends on what is held fixed, it obviously follows that there is more to metaphysical model-selection than counterfactual reasoning: at the least, it also includes a non-counterfactual determination of the fixed background. The upshot is that counterfactual supposition – at least on its own – does

46

47

Timothy Williamson, The Philosophy of Philosophy (Blackwell Publishing, 2007), especially Ch. 5. Paul, ‘Metaphysics as Modeling: The Handmaiden’s Tale,’p. 23.

229

230

Demarcating metaphysics and science: can metaphysics be naturalized?

not provide a complete story about our epistemic access to metaphysically necessary facts, as knowledge of constitutive facts must precede the counterfactual account itself.48 Consider the following example. Paul argues that a counterexample to the claim that causation is necessarily a relation of counterfactual dependence between events would require finding a metaphysically possible world with a case of causation between events not exhibiting counterfactual dependence. But how can we (fail to) find such a world – thus vindicating the counterfactual account of causation – if whether it exists at all depends on what we take to remain fixed across possible worlds? Accordingly, the problem that immediately follows is that if causation is indeed necessarily a relation of counterfactual dependence between events, then it is exactly one of those precious metaphysical necessities that should be held fixed across metaphysically possible counterfactual scenarios. According to a Williamson-style analysis, failing to hold a constitutive fact fixed in a counterfactual supposition entails a contradiction. But the question at hand concerns whether it is a (metaphysically necessary) constitutive fact that causation is a relation of counterfactual dependence between events. In either case we are going to get unexpected results in our counterfactual suppositions. Specifically, on the one hand we will get a contradiction where there should be none if we mistakenly hold it fixed that causation is a relation of counterfactual dependence between events, and on the other hand we will not get a contradiction where there should be one if it is a constitutive fact and we fail to keep it fixed. In other words, the method can produce reliable results only if we know which facts to hold fixed in the first place. Summarizing, Paul clearly identifies an important parallelism between scientific and metaphysical methodology: they both employ inference to the best explanation and criteria for theory choice based on theoretical virtues; and they both work with abstract and idealized models of reality. Her suggestion that metaphysical models are in some sense more fundamental than scientific models is also intriguing. However, it remains open to debate how exactly metaphysical models connect to scientific models, and how (if at all) theoretical virtues can be reliable guides for theory choice in metaphysics. 48

On this, see Tuomas E. Tahko, ‘Counterfactuals and Modal Epistemology,’ Grazer Philosophische Studien 86 (2012), pp. 93–115.

9.5 Moderately naturalistic metaphysics

9.5 Moderately naturalistic metaphysics

In what remains of this chapter, we will examine a more moderate position, attempting a reconciliation between science and metaphysics. This position has important similarities with the one proposed by Paul, but 49

also with Lowe’s more ambitious metametaphysical position. Recall that we discussed Lowe’s position in connection to autonomous metaphysics – Lowe considers metaphysics and science to differ both in terms of their subject matter and with respect to methodology. However, by combining elements from Paul’s and Lowe’s proposals, we can outline a position that may improve on both of them. We do not need to rehearse Lowe’s position in detail here, for the relevant aspects have been covered earlier in this chapter and in previous chapters. The key aspect that we need to recall from Lowe’s work is that, as he puts it, metaphysics can be characterized as ‘the science of the possible’.50 Moreover, we should recall the remarks made in Chapter 7 regarding the relationship between a priori and a posteriori inquiry, namely that there are reasons to think that both of these are needed when pursuing metaphysics, and that they can be considered to proceed in a manner of ‘bootstrapping’ – that is, by alternating in a way that renders talk of ‘pure’ a priori or a posteriori knowledge inert. We will also consider whether it is possible to address one of the shortcomings we identified in Paul’s account, namely finding a way to choose between models that goes beyond a comparison of theoretical virtues. The suggestion, which is perhaps rather surprising, is that it is at least in some cases possible to test metaphysical hypotheses empirically, albeit in an indirect sense. This can be done by applying such hypotheses to the interpretation of our best scientific theories, as has been suggested already. Lowe thinks that metaphysical a priori arguments are the source of the type of constitutive or essentialist knowledge that the counterfactual approach familiar from Williamson – which we encountered in the previous section – seems to struggle with. Further, Lowe holds that metaphysical modality is grounded in essence. But what does this have to do with scientific knowledge? Well, Lowe’s view is that we have a priori access to modal knowledge concerning ontological categories – this epistemic assumption 49

50

For a recent example of a view on these lines, see Matteo Morganti, Combining Science and Metaphysics (New York: Palgrave Macmillan, 2013). Lowe, ‘The Rationality of Metaphysics,’ p. 100.

231

232

Demarcating metaphysics and science: can metaphysics be naturalized?

is the core of the methodology of Lowe’s metaphysics – and he specifies that the relevant epistemic process is not based on intuitions or thought experiments, but rather on direct a priori access to essentialist facts which ground the modal truths. It is central to this account that such essentialist knowledge precedes empirical knowledge about which ontological categories are actual. However, and this is a crucial point, Lowe maintains that our epistemic access to essentialist knowledge is relatively unproblematic and doesn’t in fact require any specific philosophical treatment. An illustrative example that Lowe provides concerns the case of transuranic elements: many of them were only synthesized after their possible existence was determined by non-empirical means. With the help of Mendeleev’s periodic table, chemists have been able to predict the existence of a number of yet-to-be-discovered elements and to make highly accurate predictions about their properties. Later on, we were able to synthesize these elements and verify that they indeed had the predicted properties. Lowe proposes that this process would not have been possible without a prior grasp of the essences of these transuranic elements.51 But simply understanding what would qualify as a transuranic element of a certain type was sufficient for defining the relevant categories. This easily generalizes, seemingly lending support to Lowe’s view of metaphysical knowledge of essences. If Lowe is correct, it seems that a priori reflection was sufficient to predict the characteristics of the yet-to-be-seen elements; and that these predicted characteristics were essential features of those elements. But it must be acknowledged that this prediction reflected (what we take to be) objective features of the world only because, for the known elements of the table, we had prior empirical acquaintance with them. Lowe’s idea, then, is that some a priori idea of what is possible and what is not is certainly needed to even start to structure empirical data conceptually. The type of modal rationalism that Lowe puts forward, as we’ve seen, infers from this that our grasp of essences is prior to the gathering of empirical data and input. This may be plausible enough in some cases – Lowe’s most forceful examples do not concern natural kinds, but abstract objects such as sets. In these cases, Lowe’s take on essences might well be a defensible approach.But since knowledge of natural kinds (understood as explaining what is shared by the concrete members of 51

See E. J. Lowe, ‘Two Notions of Being: Entity and Essence,’ Royal Institute of Philosophy Supplements 83.62 (2008), pp: 23–48.

9.5 Moderately naturalistic metaphysics

a kind rather than as pointing to abstract universals) requires input from the empirical sciences, it appears that the process of coming to know essences cannot be completely a priori, at least not in all cases. In fact, unless some account is given of this empirical element Lowe’s point runs the risk of being reduced to an uncontroversial but relatively unimportant claim about our understanding of concepts (rather than grasp of ‘real’ essences). What remains of Lowe’s proposal, then? Can his suggestion that metaphysics is an exploration of asui generis, fundamental space of possibilities be saved in spite of the shortcomings just pointed out? To maintain that we have a priori access to essence even in the case of material objects such as natural kinds, Lowe would have to adopt the view that the essential features of material objects are, by and large, accessible to us via a simple reflective process. This does not compel Lowe to claim that the full essence is to be reached so easily, but we ought to be able to access enough of the essence to be able to separate one object from another. Hence, if we try to separate water from other superficially similar liquids, it would not be enough to say that, for instance, water is a transparent liquid (in ‘normal’ conditions, i.e., in room temperature etc.) because there are several other chemical substances, such as hydrogen peroxide, that have the same (essential) features. But consider where Lowe ends up: [O]ur natural classifications do not need to be, and in factshould not be, forced into a single, all-embracing taxonomic scheme. Realdivisions in nature are reflected by our natural classificatory schemes, but they are often divisions at different levels, allowing for agood deal of cross-classification. So, to revert to an earlier example, thereis nothing wrong in saying that, for some purposes, diamond, graphite and charcoal may be regarded as different kinds of substance[.]52

So diamond, graphite, and charcoal may be regarded as different kinds of substance for some purposes, even though they can be regarded as the same substance, namely carbon, from a microstructural point of view. On the face of it, the passage above may even seem to invite conventionalism, but this is surely not what Lowe has in mind. Nevertheless, it does leave his view open to a conventionalist challenge, for who is to say which of our classificatory schemes is supposed to reflect the real divisions in nature, if any? 52

E. J. Lowe, ‘Locke on Real Essence and Water as a Natural Kind: A Qualified Defence,’ Aristotelian Society Supplementary Volume 85 (2011), pp. 16–17.

233

234

Demarcating metaphysics and science: can metaphysics be naturalized?

Let us see where we can get with this. What we are effectively looking for is a model, more or less in Paul’s sense, which would give us the tools to assess cases like elements, explaining our ability to make predictions about them in advance of observations. So consider carbon. Diamond and graphite, which Lowe mentions, are two of the several allotropes of carbon – pure forms of the same element that differ in structure. But there is of course a reason behind the fact that we typically classify different allotropes as members of the same kind, despite the significant differences in their chemical properties. The reason has its source in the fact that despite their many differences, the allotropes of carbon can be subjected to the same type of chemical reaction with the same type of result. This reaction is oxidation: If one burns diamond, graphite, or charcoal, the result is always the same, namely pure carbon dioxide. If any of these allotropes were compounds of different elements instead, the burning would result in some impurities. As it happens, this is exactly how Lavoisier, the famous eighteenth-century chemist, discovered that diamond is indeed an allotrope of carbon. Lavoisier heated a diamond in a glass jar until it disappeared and observed that the weight of the jar had not changed, hence concluding that the diamond must have been made of carbon to produce the carbon dioxide gas present in the jar after heating. 53 We can conclude that when modelling elements, we ought to take into account at least two things. The first is the ability of elements to form allotropes and their ability to form compounds with other elements (which can also take several forms; that is, polymorphism in general must be taken into account). The second is the survival of something essential to the element in all of the different forms that it can take. More precisely, it is in virtue of the essence of carbon that carbon atoms are capable of forming allotropes with varying crystalline structures, and it is also in virtue of the essence of carbon that the vast range of forms that carbon can take in a variety of allotropes and compounds still share some aspect of the elemental form of carbon. What these aspects are is a question of chemistry and we need not go into a lot of detail here. 54 But we could perhaps understand the 53

54

For further discussion, see Tuomas E. Tahko, ‘Empirically-Informed Modal Rationalism,’ in R. W. Fischer and F. Leon (eds.), Modal Epistemology After Rationalism, Synthese Library (Dordrecht: Springer, forthcoming). But see Tuomas E. Tahko, ‘Natural Kind Essentialism Revisited,’ Mind 124.915 (2015), pp. 795–822.

9.5 Moderately naturalistic metaphysics

chemical properties of compounds to be a result of the interaction of the elements present in those compounds, which would point towards microstructural essentialism. If this were correct, it would have to be the case that something survives in the causal processes that elements undergo when forming compounds – and also allotropes. This all-too-brief example gives us some clues towards combining elements from Paul’s and Lowe’s proposals into an empirically informed analysis of metaphysical-cum-scientific inquiry. Since the model we just examined makes some predictions about how future scientific inquiry could proceed, it is in a sense at least indirectly testable: that is, it is not entirely immune to, and indifferent towards, empirical input. This type of indirect testability could be used to build a bridge between scientific and metaphysical models and steer clear of the objection that, since metaphysics is not empirically relevant and yet aims at the truth, metaphysics should be dismissed. Relatedly, the systematic application of metaphysical concepts and hypotheses (taken from the space of possibilities in accordance to Lowe’s suggestion) with a view to interpreting our best current science seems to provide at least one criterion for going about answering metaphysical questions and selecting between metaphysical models. What gets selected, in particular, are those among the latter that are not only the best in terms of theoretical virtues but also most suited, all things considered, for making sense of the best available science. This, together with the strategy based on the modal constraints for scientific theorizing emerging from Lowe’s proposal – the idea that a preliminary exploration of possibility space is necessary for scientific inquiry to even get started – is tantamount to saying that, in spite of its (relative) autonomy, metaphysics is not a merely conceptual or linguistic activity, fully distinct from science. The upshot would be what might be labelled ‘moderately naturalistic metaphysics’. It will be left to the reader to assess these options – and others, for our discussion has hardly covered all the options. But perhaps it is reasonable to conclude this book by noting that contemporary analytic metaphysics can go at least some way towards meeting the challenge of combining science and metaphysics, and indeed demarcating them.

235

Glossary

allism and noneism: Competing views regarding controversial entities such as past and future things, merely possible things, universals, numbers, and so on. Allism is the view that all such things exist, whereas noneism states that none of them exist. application conditions: The conditions under which a term may be properly applied. atomism: The theory, srcinating from Democritus, that physical objects consist of simples, indivisible particles. See also mereological atomism. conceptual analysis: A methodological position emphasizing the linguistic or conceptual clarification of the notions we use in philosophy and/or science; typically a priori. conceptual necessity: Necessary in virtue of the definitions of concepts (plus the laws of logic), e.g., ‘All bachelors are unmarried’. constitutive facts: Elements of background knowledge that are held fixed in counterfactual suppositions. A plausible example of constitutive facts are metaphysically necessary facts concerning the essences of entities. contingent: In contrast to necessary; could have been otherwise. For instance, a proposition is metaphysically contingent as opposed to metaphysically necessary if it is true in some metaphysically possible worlds and false in some other metaphysically possible worlds. conventionalism: The view that some or all aspects of metaphysical inquiry are based on linguistic, psychological, or societal conventions rather than genuine distinctions in reality. counterfactual supposition: Counterfactual suppositions concern non-actual matters. If p holds in the actual world, what would follow if not-p? deflationary, deflationism: The view that ontological questions (universally or in some specific area) are not substantive, but are rather based on confusion, linguistic choice, or something else.

236

Glossary

emergence, emergent: The idea that complex systems and patterns can arise or emerge out of simple interactions; the emergent properties of a complex system may not be reducible to the parts of the system. endurantism and perdurantism: Competing views regarding temporal parts. Perdurantists hold that material objects have temporal parts in addition to spatial parts and hence that objects ‘perdure’; objects persist by having temporal parts. Endurantists hold that material objects do not have temporal parts, which leads them to think that objects are wholly present at any time that they exist; they ‘endure’. See also three- and four-dimensionalism. entity: Something existing, an object, although notnecessarily a material object. essence: The essence of an entity can be expressed through its identity and existence conditions; the properties which make it the very thing it is. These properties are necessary for the entity: it couldn’t exist without them. Sometimes essence is explicated in terms of real definition, sometimes in purely modal terms. Typically, essences are not considered to be entities themselves. essentialism: The view that (at least some) things have essences, which make them the very things or kinds of things that they are. One version of this view is natural-kind essentialism, which postulates essences for natural kinds, such as water. external questions: Originating from Rudolf Carnap. External questions are questions that go beyond established frameworks; they concern the applicability of a framework to reality. According to Carnap, these are typically questions that philosophers (metaphysicians) ask, but Carnap regards them as meaningless. See also internal questions. essential dependence: A form of ontological dependence whereby an entity x depends essentially for its existence upon an entity y, in the sense that it is part of the essence of x that x exists only if y exists. existential dependence: A form of ontological dependence whereby an entity x depends upon an entity y for its existence, in the sense that x couldn’t exist if y didn’t exist. experience-based intuition: Intuition derived from experience; concerning the phenomenology of the subject matter of the intuition rather than conceptual analysis. fallibilism, fallible: The view that a given position is subject to revision; the position is fallible. four-dimensionalism: See three- and four-dimensionalism.

237

238

Glossary

fundamental, fundamentality: Something ontologically independent , more basic than, or prior to everything else. It is often thought that all non-fundamental things must ontologically dependon or be grounded in some fundamental things. ground, grounding: A metaphysical priority relation distinct from causation. For instance, the members of a set are prior to the set itself; the existence of the set is grounded in its members. Grounding is often considered to obtain between facts, i.e., if fact x is grounded in fact y, then y metaphysically explains fact x. gunk: The idea of atomless gunk in mereology is that everything has a proper part; there are no mereological atoms. identity-dependence: A form of ontological dependence whereby an entity x depends upon an entity y for its identity, in the sense that x wouldn’t be the entity that it is or the kind of entity that it is if y didn’t exist. infinite regress: A relationship transmitted through an infinite number of terms in a series or chain, where there is no last term that terminates the chain. See also metaphysical infinitism. internal questions: Originating from Rudolf Carnap. Internal questions are questions within a certain framework. They make sense once the framework is established and adopted. For instance, internal questions can be questions of varying difficulty in a scientific framework. But general questions regarding what exists in a framework are trivial, so when philosophers (metaphysicians) ask general questions, they presumably do not have internal questions in mind, or so Carnap thought. See also external questions. joint-carving: When a predicate carves at the joints, it is considered to correspond with some structure in reality; it picks out a distinction which is genuine rather than conventional. If a predicate carves perfectly, then it picks out something fundamental. language pluralism: The view, of Carnapian srcins, that we can choose our (linguistic) framework liberally. logical necessity: Necessary in virtue of the laws of logic, e.g., ‘It either rains or does not rain’. mereology: The study of parts and wholes, or parthood, and closely related concepts. mereological atomism: The view that there are mereological atoms, simples that have no further proper parts and that everything is ultimately composed of. metaphysical foundationalism: The view that there is a foundation to reality, a fundamental level that everything else ultimately depends on. On this view, chains of dependence must terminate.

Glossary

metaphysical infinitism: The view that there is no foundation to reality, no fundamental level that everything else ultimately depends on. On this view, chains of dependence could go on ad infinitum. See also infinite regress. metaphysical necessity: Necessary in virtue of the essences of entities, e.g., ‘Water is H2O’. metaphysical realism: The doctrine that there is a mind-independent external world. modal, modality: Modality, or modal terms, describe what could be or must be the case, i.e., they deal with possibility and necessity. modal empiricism: A family of views emphasizing empirical elements in modal epistemology. modal epistemology: The study of how we come to know modal truths. modal rationalism: A family of views emphasizing a priori, non-empirical elements in modal epistemology. monism: The view, famously associated with Spinoza, that there is only one independent thing or substance. A more moderate version popularized by Jonathan Schaffer, priority monism, suggests that there is only one fundamental thing. natural kind: When science or philosophy picks out a grouping that reflects the structure of the world that is independent of the interests and actions of humans, it has picked out a natural rather than an artificial kind. natural necessity: Also known as nomological necessity and physical necessity. Necessary in virtue of the laws of physics or laws of nature. naturalism: A group of views prioritizing natural science; the idea that natural sciences exhausts our area of inquiry and natural phenomena are all there is. Various different, weaker, and stronger versions of the view exist. nihilism: The view that a collection of material objects never composes a further object; composition never occurs. A possible answer to the Special Composition Question. nomological necessity: Also known as natural necessity and physical necessity. Necessary in virtue of the laws of physics. non-experimental intuition: A specific type of intuition associated with experimental philosophy. Refers to an intuition to which the designer of an experiment studying the intuitions of two populations directly appeals or which must eventually be appealed to in order to establish a criterion for determining whether two populations share a common concept. noneism: See allism and noneism. Ontologese: The special language of ontology, more suitable for ontological debates than ordinary English.

239

240

Glossary

ontological dependence: A family of relations referring to various forms of metaphysical, non-causal dependence relations between entities. Identity-dependence, essential dependence, and existential dependence are forms of ontological dependence. ontological realism: In this book, the notion is used in a very general sense, referring to the view that ontological facts are objective or mind-independent. But the reader should be aware that the notion is used in various more precise (and controversial) senses in the literature. perdurantism: See endurantism and perdurantism. physical necessity: Also known as natural necessity and nomological necessity. Necessary in virtue of the laws of physics. possible worlds: Often thought to be just a façon de parler – a way of speaking about modal terms, a heuristic tool – or according to modal realists, like David Lewis, to refer to real alternative worlds (although not accessible from the actual world). principle of charity: The principle according to which a debate can fail to be merely verbal only if each side in a debate is genuinely charitable and still fails to interpret the assertions of the other side as being true in their respective language. principle of independence: A principle introduced by Alexius Meinong, which states that an entity may have any properties whatsoever, independently of its existence. quantifier variance: The view that there is no uniquely best ontological language for describing the world; suggests that many (but not necessarily all) of our ontological disputes are merely verbal or conventional. rational intuition: A conception of intuition as a special rational faculty, emphasizing its role in a priori judgements; our belief in certain propositions is justified by rational intuition rather than by the senses, introspection, or memory. real definition: An Aristotelian notion associated with the nature or essence of a thing. To give a real definition of some entity is to state its existence and identity conditions. reduction, reductionism: When one thing x reduces to another thing y, then everything about x is in some sense already contained in y and can be expressed by talking about y. Reductionism applied to two fields of study more generally is often thought to undermine the distinction between them, e.g., the attempt to reduce consciousness to neurophysiology. reflexivity: A relation R is reflexive if and only if R is self-relating, i.e., everything bears relation R to itself, e.g., ‘being self-identical’.

Glossary

relativism: May be applied to various areas. Relativism concerning a given subject matter suggests that judgements made about the subject matter are relative in some sense, e.g., relative to linguistic decisions or psychological biases. Special Composition Question (SCQ): The question: Under what conditions do a number of objects compose a further object? Introduced by Peter van Inwagen, according to whom the answer to the question must be formulated in non-mereological terms. structuralism: A group of views favouring structure-oriented ontology in contrast to object-oriented ontology, e.g., we focus on ‘fields’ rather than ‘particles’. One version of this view is ontic structural realism. substance: An Aristotelian notion referring to something independently existing, fundamental, or basic. symmetry: A relation R is symmetric if and only if: if x is related by R to y, then y is related by R to x, e.g., ‘being siblings’. theoretical virtues: A group of features that a scientific or philosophical theory may have, usually assumed to make the theory preferable because it is more likely to be true or more likely to be useful. Theoretical virtues include simplicity, explanatory power, internal consistency, empirical adequacy, and so on. three- and four-dimensionalism: Competing views regarding temporal parts. On the four-dimensional view, a material object can be thought of as a type of ‘spacetime worm’, since all the different temporal parts are also, in some sense, parts of the object. In contrast, the three-dimensional view attaches no special significance to past or future temporal parts of an object – the object is there at a given time and that’s it. See also endurantism and perdurantism. transitivity: A relation R is transitive if and only if: if x is related by R to y, and y is related by R to z, then x is related by R to z, e.g., ‘being taller than’. truthmaking, truthmakers: Truthmaker theory concerns the idea that when something x makes it true that p, then x is a truthmaker for p. In other words, p is true in virtue of x. A variety of different views about the truthmaker relation, the nature of truthmakers such as x as well as the nature of truthbearers such as p have been entertained. universalism: The view that a collection of material objects always composes a further object. A possible answer to the Special Composition Question. vague, vagueness: Something that lacks sharp boundaries. A term or predicate is vague when there is no sharp cut-off between cases where the term applies and cases where it doesn’t apply, e.g., ‘is bald’. Some consider vagueness to

241

242

Glossary

always be merely linguistic, some think that there could also be vagueness in the world, that is, metaphysical vagueness. well-foundedness: An order is said to be well-founded if every non-fundamental element of the order is fully grounded by some fundamental element(s). The notion of well-foundedness may be used to explicate metaphysical foundationalism.

Bibliography

Aristotle. 1984. Metaphysics (W. D. Ross, trans.; revised by J. Barnes). Princeton University Press. Armstrong, D. M. 2004. Truth and Truthmakers, Cambridge University Press. Atkinson, D. 2003. ‘Experiments and Thought Experiments in Natural Science,’ in M. C. Galavotti (ed.), Observation and Experiment in the Natural and Social Sciences, Boston Studies in the Philosophy of Science 232. Dordrecht: Kluwer, pp. 209–225. Audi, P. 2012. ‘A Clarification and Defense of the Notion of Grounding,’ in F. Correia and B. Schnieder (eds.), Metaphysical grounding. Cambridge University Press, pp. 101–121. Barrow, J. D. 2001. ‘Cosmology, Life, and the Anthropic Principle,’ Annals of the New York Academy of Sciences950, 139–153. Bealer, G. 1992. ‘The Incoherence of Empiricism,’ Aristotelian Society Supplementary Volume 66, 99–138. 1998. ‘Intuition and the Autonomy of Philosophy,’ in M. DePaul and W. Ramsey (eds.), Rethinking Intuition: The Psychology of Intuition and Its Role in Philosophical Inquiry. Lanham, MD: Rowman and Littlefield, pp. 201–240. 2002. ‘Modal Epistemology and the Rationalist Renaissance,’ in T. S. Gendler and J. Hawthorne (eds.), Conceivability and Possibility. Oxford University Press, pp. 71–125. 2004. ‘The Origins of Modal Error,’ Dialectica 58.1, 11–42. Bennett, K. 2009. ‘Composition, Colocation and Metaontology.’ in D. Chalmers, D. Manley, and R. Wasserman (eds.), Metametaphysics. Oxford University Press, pp. 38–76. 2011. ‘By Our Bootstraps,’ Philosophical Perspectives 25, 27–41. Benovsky, J. 2013. ‘From Experience to Metaphysics: On Experience-based Intuitions and their Role in Metaphysics,’ Noûs, (online) (forthcoming). Ben-Yami, H. 2004. Logic and Natural Language. Surrey: Ashgate.

243

244

Bibliography

2009. ‘Plural Quantification Logic: A Critical Appraisal,’ Review of Symbolic Logic 2.1, 208–232. Berto, F. 2011. ‘Modal Meinongianism and Fiction: The Best of Three Worlds,’ Philosophical Studies 152, 313–334. Berto, F. and M. Plebani. 2015. Ontology and Metaontology: A Contemporary Guide. London: Bloomsbury. Bird, A. 2007. Nature’s Metaphysics. Oxford University Press. Bliss, R. L. 2014. ‘Viciousness and Circles of Ground,’Metaphilosophy 45.2, 245–256. 2013. ‘Viciousness and the Structure of Reality,’ Philosophical Studies 166.2, 399–418. Bliss, R. L. and K. Trogdon. 2014. ‘Metaphysical Grounding,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2014 edn), see http://plato. stanford.edu/archives/win2014/entries/grounding/ Bohm, D. 1984. Causality and Chance in Modern Physics, (Second edn). London: Routledge. BonJour, L. 1998. In Defense of Pure Reason. Cambridge University Press. Booth, A. R. and D. P. Rowbottom, 2014. (eds.). Intuitions. Oxford University Press. Bricker, P. 2014. ‘Ontological Commitment,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2014 edn), see http://plato.stanford.edu/ archives/win2014/entries/ontological-commitment/ Callender, C. 2011. ‘Philosophy of Science and Metaphysics,’ in S. French and J. Saatsi (eds.), The Continuum Companion to the Philosophy of Science. London: Continuum, pp. 33–54. Cameron, R. P. 2008. ‘Turtles All the Way Down: Regress, Priority and Fundamentality,’ Philosophical Quarterly 58, 1–14. Cappelen, H. 2012. Philosophy without Intuitions. Oxford University Press. Carnap, R. 1931. ‘The Elimination of Metaphysics Through Logical Analysis of Language’ (A. Pap, trans.), in A. J. Ayer (ed.), Logical positivism. New York: The Free Press, 1959, pp. 60–81; srcinally published in Erkenntnis 2.1 as ‘Überwindung der Metaphysik durch logische Analyse der Sprache,’ pp. 219–241. 1950. ‘Empiricism, Semantics, and Ontology.’ Revue Internationale de Philosophie 4; reprinted in his Meaning and Necessity: A Study in Semantics and Modal Logic. University of Chicago Press, 1956. Casullo, A. 2003. A Priori Justification. Oxford University Press. 2013. ‘Four Challenges to the A Priori–A Posteriori Distinction,’Synthese, (online) (forthcoming). Casullo, A. and J. C. Thurow 2013. (eds.), The A Priori in Philosophy. Oxford University Press.

Bibliography

Chalmers, D. 1996. The Conscious Mind: In Search of a Fundamental Theory. Oxford University Press. 2002. ‘Does Conceivability Entail Possibility?’, in T. S. Gendler and J. Hawthorne (eds.), Conceivability and Possibility. Oxford University Press, pp. 145–200. 2006. ‘The Foundations of Two-Dimensional Semantics,’ in M. Garcia-Carpintero and J. Macia (eds.), Two-Dimensional Semantics: Foundations and Applications. Oxford University Press, pp. 55–140. 2009. ‘Ontological Anti-Realism,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.), Metametaphysics. Oxford University Press, pp. 77–129. 2011. ‘Verbal Disputes,’ Philosophical Review 120.4, 515–566. 2014. ‘Intuitions in Philosophy: A Minimal Defense,’ Philosophical Studies 171.3, 535–544. Chalmers, D., D. Manley, and R. Wasserman 2009. (eds.). Metametaphysics. Oxford University Press. Chudnoff, E. 2013. Intuition. Oxford University Press. Cohnitz, D. 2006. ‘When are Discussions of Thought Experiments Poor Ones? A Comment on Peijnenburg and Atkinson,’ Journal for General Philosophy of Science 37.2, 373–392. Cornell, E. A. and C. E. Wieman. 1998. ‘The Bose–Einstein Condensate,’ Scientific American 278.3, 40–45. Correia, F. 2008. ‘Ontological Dependence,’ Philosophy Compass 3, 1013–1032. 2010. ‘Grounding and Truth-functions,’ Logique & Analyse 211, 251–279. Correia, F. and B. Schnieder. 2012. ‘Grounding: An Opinionated Introduction,’ in F. Correia and B. Schnieder (eds.), Metaphysical Grounding. Cambridge University Press, pp. 1–36. 2012. (eds.). Metaphysical Grounding: Understanding the Structure of Reality. Cambridge University Press. Crane, T. 2013. The Objects of Thought. Oxford University Press. Creath, R. ‘Logical Empiricism,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2014 edn), see http://plato.stanford.edu/archives/spr2014/ entries/logical-empiricism/ Daly, C. 2010. Introduction to Philosophical Methods. Peterborough, ON: Broadview Press. 2012. ‘Scepticism about Grounding,’ in F. Correia and B. Schnieder (eds.), Metaphysical Grounding. Cambridge University Press, pp. 81–100. Daly, C. and D. Liggins. 2014. ‘In Defence of Existence Questions,’ Monist 97.7, 460–478. Davidson, D. 1987. ‘Knowing One’s Own Mind,’ Proceedings and Addresses of the American Philosophical Association 60.3, 441–458.

245

246

Bibliography

de Cruz, H. 2015. ‘Where Philosophical Intuitions Come From,’ Australasian Journal of Philosophy 93.2, 233–249. Dehmelt, H. 1989. ‘Triton,… Electron,… Cosmon,…: An Infinite Regression?’, Proceedings of the National Academy of Sciences 86, 8618–8619. Dennett, D. 2013. Intuition Pumps and Other Tools for Thinking. New York: W. W. Norton & Co. deRosset, L. 2013. ‘What Is Weak Ground?’, Essays in Philosophy 14.1, Article 2. Dixon, S. ‘What Is the Well-Foundedness of Grounding?’, Mind (forthcoming). Dorato, M. and M. Morganti. 2013. ‘Grades of Individuality. A Pluralistic View of Identity in Quantum Mechanics and in the Sciences,’ Philosophical Studies 163, 591–610. Dorr, C. 2005. ‘What We Disagree about When We Disagree about Ontology,’ in M. E. Kalderon (ed.), Fictionalism in Metaphysics. Oxford University Press, pp. 234–286. 2010. ‘Review of Every Thing Must Go,’ Notre Dame Philosophical Reviews, seehttps:// ndpr.nd.edu/news/24377-every-thing-must-go-metaphysics-naturalized/ Dummett, M. 1981. Frege. Philosophy of Language, (Second edn). Cambridge, MA: Harvard University Press. Eklund, M. 2009. ‘Carnap and Ontological Pluralism,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.), Metametaphysics. Oxford University Press, pp. 130–156. 2011. ‘Review of Quantifier Variance and Realism:Essays in Metaontology,’ Notre Dame Philosophical Reviews, see https://ndpr.nd.edu/news/24764-quantifier-va riance-and-realism-essays-in-metaontology/ 2013. ‘Carnap’s Metaontology,’ Noûs 47.2, 229–249. 2014. ‘Rayo’s Metametaphysics,’ Inquiry: An Interdisciplinary Journal of Philosophy 57.4, 483–497. Enderton, H. B. 2012. ‘Second-order and Higher-order Logic,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Fall 2012 edn), see http://plato.stanford.edu/archives/fall2012/entries/logic-higher-order/ Fine, K. 1994a. ‘Essence and Modality,’ Philosophical Perspectives 8, 1–16. 1994b. ‘OntologicalDependence,’ Proceedings of the Aristotelian Society 95, 269–290. 1995. ‘Logic of Essence,’ Journal of Philosophical Logic 24, 241–273. 2001. ‘The Question of Realism,’ Philosophers Imprint 1, 1–30. 2009. ‘The Question of Ontology,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.), Metametaphysics. Oxford University Press, pp. 157–177. 2010. ‘Towards a Theory of Part,’ Journal of Philosophy 107.11, 559–589. 2012. ‘Guide to Ground,’ in F. Correia and B. Schnieder (eds.), Metaphysical Grounding. Cambridge University Press, pp. 37–80.

Bibliography

2012. ‘What Is Metaphysics?’, in T. E. Tahko (ed.), Contemporary Aristotelian Metaphysics. Cambridge University Press, pp. 8–25. Fiocco, M. O. 2007. ‘Conceivability, Imagination, and Modal Knowledge,’ Philosophy and Phenomenological Research, 74.2, 364–380. Fischer, R. W. and F. Leon (eds.). 2015. Modal Epistemology After Rationalism, Synthese Library. Dordrecht: Springer (forthcoming). French, S. 2011. ‘Metaphysical Underdetermination: Why Worry?’, Synthese 180, 205–221. 2014. The Structure of the World:Metaphysics and Representation. Oxford University Press. Gendler, T. S. and J. Hawthorne (eds.). 2002. Conceivability and Possibility. Oxford University Press. Georgi, H. 1989. ‘Effective Quantum Field Theories,’ in P. Davies (ed.), The New Physics. Cambridge University Press, pp. 446–457. Gibbard, A. 1975. ‘Contingent Identity,’Journal of Philosophical Logic 4, 187–221. Hackett, J. 2007. ‘Locality and Translations in Braided Ribbon Networks,’ Classical and Quantum Gravity 24.23, 5757–5766. Hale, B. 2013. Necessary Beings: An Essay on Ontology, Modality, and the Relations Between Them. Oxford University Press. Haug, M. C. (ed.). 2014. Philosophical Methodology: The Armchair or the Laboratory? Abingdon: Routledge. Hawley, K. 2006. ‘Science as a Guide to Metaphysics?’, Synthese 149.3, 451–470. Hawthorne, J. 2008. ‘Three-dimensionalism vs. Four-Dimensionalism,’ in T. Sider, J. Hawthorne and D. W. Zimmerman (eds.), Contemporary Debates in Metaphysics. Oxford: Blackwell Publishing, pp. 263–282. Heil, J. 2003. From an Ontological Point of View. Oxford: Clarendon Press. 2012. The Universe As We Find It. Oxford University Press. Hirsch, E. 2002. ‘Quantifier Variance and Realism,’ Philosophical Issues 12, 51–73. 2005. ‘Physical-Object Ontology, Verbal Disputes, and Common Sense,’ Philosophy and Phenomenological Research 70, 67–97. 2009. ‘Ontology and Alternative Languages,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.),Metametaphysics. Oxford University Press, pp. 231–259. Hofweber, T. 2009. ‘Ambitious, Yet Modest, Metaphysics,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.), Metametaphysics. Oxford University Press, pp. 260–289. Horvath, J. and T. Grundmann (eds.). 2012. Experimental Philosophy and Its Critics. London: Routledge. Hüttemann, A. and D. Papineau. 2005. ‘Physicalism Decomposed,’ Analysis 65, 33–39.

247

248

Bibliography

Jackson, F. 1986. ‘What Mary Didn’t Know,’ The Journal of Philosophy 83, 291–295. 1998. From Metaphysics to Ethics: A Defence of Conceptual Analysis. Oxford: Clarendon Press. Jenkins, C. S. 2010. ‘Concepts, Experience and Modal Knowledge,’ Philosophical Perspectives 24, 255–279. 2010. ‘What Is Ontological Realism?’, Philosophy Compass 5/10, 880–890. 2011. ‘Is Metaphysical Dependence Irreflexive?’, The Monist 94.2, 267–276. 2014. ‘Intuition, “Intuition”, Concepts and the A Priori,’ in A. R. Booth and D. P. Rowbottom (eds.), Intuitions. Oxford University Press, pp. 91–115. 2014. ‘Merely Verbal Disputes,’ Erkenntnis 79.1, 11–30. Kim, J. 2010. Essays in the Metaphysics of Mind. Oxford University Press. Koslicki, K. 2005. ‘On the Substantive Nature of Disagreements in Ontology,’ Philosophy and Phenomenological Research 71, 85–105. 2012. ‘Varieties of Ontological Dependence,’ in F. Correia and B. Schnieder (eds.), Metaphysical Grounding. Cambridge University Press, pp. 186–213. Kripke, S. 1980. Naming and Necessity . Cambridge, MA: Harvard University

Press. 2013. Reference and Existence. Oxford University Press. Ladyman, J. 1998. ‘What Is Structural Realism?’, Studies in History and Philosophy of Science Part A 29, 409–424. 2012. ‘Science, Metaphysics and Method,’ Philosophical Studies 160.1, 31–51. Ladyman, J. and D. Ross (with D. Spurrett and J. Collier). 2007. Every Thing Must Go. Oxford University Press. Lange, M. 2008. ‘Could the Laws of Nature Change?’, Philosophy of Science 75.1, 69–92. Lewis, D. 1990. ‘Noneism or Allism?’, Mind 99.393, 23–31. Lowe, E. J. 2006. The Four-Category Ontology: A Metaphysical Foundation for Natural Science. Oxford: Clarendon Press. 2008. ‘Two Notions of Being: Entity and Essence,’ Royal Institute of Philosophy Supplements 83.62, 23–48. 2010. ‘Ontological Dependence,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2010 edn), see http://plato.stanford.edu/archives/spr2010/ entries/dependence-ontological/ 2011. ‘Locke on Real Essence and Water as a Natural Kind: A Qualified Defence,’ Aristotelian Society Supplementary Volume 85, 16–17. 2011. ‘The Rationality of Metaphysics,’ Synthese 178.1, 99–109. 2012. ‘What Is the Source of Our Knowledge of Modal Truths?’, Mind 121, 919–950.

Bibliography

2014. ‘Essence vs. Intuition: An Unequal Contest,’ in A. R. Booth and D. P. Rowbottom (eds.), Intuitions. Oxford University Press, pp. 256–268. Lowe, E. J. and S. McCall. 2006. ‘3D/4D Controversy: A Storm in a Teacup,’ Noûs 40, 570–578. Markosian, N. 2005. ‘Against Ontological Fundamentalism,’ Facta Philosophica 7, 69–84. 2014. ‘Time,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2014 edn), see http://plato.stanford.edu/archives/spr2014/entries/time/ Marcus, R.B. 1967. ‘Essentialism in Modal Logic,’ Noûs 1.1, pp. 91–6. Marsh, G. 2010. ‘Is the Hirsch–Sider Dispute Merely Verbal?’, Australasian Journal of Philosophy 88.3, 459–469. Maudlin, T. 2007. The Metaphysics Within Physics. Oxford University Press. McDaniel, K. 2013. ‘Degrees of Being,’ Philosophers’ Imprint 13.19, 1–18. ‘Ways of Being,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.), 2009. Metametaphysics. Oxford University Press, pp. 290–319. McKenzie, K. 2014. ‘Priority and Particle Physics: Ontic Structural Realism as a Fundamentality Thesis,’ British Journal for the Philosophy of Science 65, 353–380. McTaggart, J. M. E. 1993. ‘The Unreality of Time,’ Mind 17, 1908, 457–473; reprinted in R. Le Poidevin and M. McBeath (eds.), The Philosophy of Time. Oxford University Press, pp. 23–34. Merricks, T. 2001. Objects and Persons. Oxford University Press. Miller, K. 2005. ‘The Metaphysical Equivalence of Three and Four Dimensionalism,’ Erkenntnis 62.1, 91–117. Morganti, M. 2013.Combining Science and Metaphysics. New York: Palgrave Macmillan. 2014. ‘Metaphysical Infinitism and the Regress of Being,’ Metaphilosophy 45.2, 232–244. 2015. ‘Dependence, Justification and Explanation: Must Reality Be Well-Founded?’, Erkenntnis 60.3, 555–572. Mulligan, K., P. Simons and B. Smith. ‘Truth-Makers,’ 1984. Philosophy and Phenomenological Research 44.3, 287–321. Nelson, M. ‘Existence,’ in E. N. Zalta (ed.), 2012. The Stanford Encyclopedia of Philosophy (Winter 2012 edn), see http://plato.stanford.edu/archives/ win2012/entries/existence/ Ney, A. 2014. Metaphysics: An Introduction. Abingdon: Routledge. Nolan, D. 1997. ‘Impossible Worlds: A Modest Approach,’ Notre Dame Journal of Formal Logic 38.4, 535–572. 2015. ‘The A Posteriori Armchair,’Australasian Journal of Philosophy 93.2, 211–231. O’Conaill, D. 2014. ‘Ontic Structural Realism and Concrete Objects,’ The Philosophical Quarterly 64.255, 284–300.

249

250

Bibliography

Oppenheim, P. and H. Putnam. 1958. ‘Unity of Science as a Working Hypothesis,’ in H. Feigl et al. (eds.), Concepts, Theories, and the Mind-Body Problem, Minnesota Studies in the Philosophy of Science Vol. II. Minneapolis: University of Minnesota Press, pp. 3–36. Papineau, D. 2012. Philosophical Devices: Proofs, Probabilities, Possibilities, and Sets. Oxford University Press. Parsons, T. 1980. Nonexistent Objects. New Haven, CO: Yale University Press. Paul, L. A. 2004. ‘The Context of Essence,’ Australasian Journal of Philosophy 82, 170–184. 2010. ‘Temporal Experience,’ Journal of Philosophy 107.7, 333–359. 2012. ‘Metaphysics as Modeling: The Handmaiden’s Tale,’Philosophical Studies 160, 1–29. Peijnenburg, J. and D. Atkinson. 2003. ‘When Are Thought Experiments Poor Ones?’, Journal for General Philosophy of Science 34.2, 305–322. Plato. Phaedrus (R. Hackforth, trans.). 1972. Cambridge University Press. Priest, G. 2005. Towards Non-Being:The Logic and Metaphysics of Intentionality. Oxford University Press. 2008. ‘The Closing of the Mind: How the Particular Quantifier Became Existentially Loaded Behind Our Backs,’ The Review of Symbolic Logic 1.1, 42–55. Pust, J. 2014. ‘Intuition,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy(Fall 2014 edn), seehttp://plato.stanford.edu/archives/fall2014/entries/intuition/ Putnam, H. 1979. ‘The Meaning of “Meaning,”’ reprinted in his Mind, Language and Reality: Philosophical Papers , Vol. 2. Cambridge University Press, pp. 215–271. 1990. ‘Is Water Necessarily H 2O?’, in J. Conant (ed.), Realism with a Human Face. Cambridge, MA: Harvard University Press, pp. 54–79. Pylkkänen, P., B. J. Hiley and I. Pättiniemi. ‘Bohm’s Approach and Individuality,’ in A. Guay and T. Pradeu (eds.) 2015. Individuals Across the Sciences. Oxford University Press, Ch. 12. Quine, W. V. 1969. ‘Existence and Quantification,’ in his Ontological Relativity and Other Essays New York: Columbia University Press. 1980. ‘On What There Is,’ The Review of Metaphysics 2, 1948, 21–38; reprinted in his From a Logical Point of View. Cambridge, MA: Harvard University Press, pp. 1–19. 1986. Philosophy of Logic, (Second edn). Cambridge, MA: Harvard University Press. Raven, M. 2013. ‘Is Ground a Strict Partial Order?’, American Philosophical Quarterly 50.2, 191–199. Rayo, A. 2007. ‘Ontological Commitment,’ Philosophy Compass 2/3, 2013. 428–444.

Bibliography

2013. Construction of Logical Space. Oxford University Press. Reicher, M. ‘Nonexistent Objects,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Summer 2014 edn), see http://plato.stanford.edu/archives/ sum2014/entries/nonexistent-objects/ Ritchie, J. 2008. Understanding Naturalism. Stocksfield: Acumen. Roca-Royes, S. 2011. ‘Conceivability and De Re Modal Knowledge,’ Noûs 45.1, 22–49. Rodriguez-Pereyra, G. 2005. ‘Why Truthmakers?’, in H. Beebee and J. Dodd (eds.), Truthmakers: The Contemporary Debate. Oxford University Press, 2011, pp. 17–31. ‘Grounding Is Not a Strict Order,’ Journal of the American Philosophical Association (forthcoming). Rose, D. and D. Danks. 2012. ‘Causation: Empirical Trends and Future Directions,’ Philosophy Compass 7.9, 643–653. Rosen, G. 2010. ‘Metaphysical Dependence: Grounding and Reduction,’ in B. Hale and A. Hoffman (eds.), Modality: Metaphysics, Logic, and Epistemology. Oxford University Press, pp. 109–135. Routley, R. ‘On What There Is Not,’ 1982. Philosophy and Phenomenological Research 43, 151–178. Rowbottom, D. P. 2014. ‘Intuitions in Science: Thought Experiments as Argument Pumps,’ in A. R. Booth and D. P. Rowbottom (eds.), Intuitions. Oxford University Press, pp. 119–134. Rueger, A. and P. McGivern. 2010. ‘Hierarchies and Levels of Reality,’ Synthese 176, 379–397. Russell, B. 1905. ‘On Denoting,’ Mind 14, 479–493. Ryckman, T. A. ‘Early Philosophical Interpretations of General Relativity,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2014 edn), see http://plato.stanford.edu/archives/spr2014/entries/genrel-early/ Saunders, S., J. Barrett, A. Kent, and D. Wallace (eds.). 2010. Many Worlds? Everett, Quantum Theory, and Reality. Oxford University Press. Schaffer, J. 2003. ‘Is There a Fundamental Level?’, Noûs 37, 498–517. 2010. ‘Monism: The Priority of the Whole,’ Philosophical Review 119, 31–76. ‘On What Grounds What,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.), 2009. Metametaphysics. Oxford University Press, pp. 347–383. 2012. ‘Grounding, Transitivity, and Contrastivity,’ in F. Correia and B. Schnieder,Metaphysical Grounding. Cambridge University Press, pp. 122–138. Schnieder, B. 2006. ‘Truth-Making Without Truth-Makers,’ Synthese 152.1, 21–46. Schrödinger, E. 1935. ‘Discussion of Probability Relations Between Separated Systems,’ Proceedings of the Cambridge Philosophical Society 31, 555–563.

251

252

Bibliography

Sidelle, A. 2002. ‘On the Metaphysical Contingency of Laws of Nature,’ in T. S. Gendler and J. Hawthorne (eds.), Conceivability and Possibility. Oxford University Press, pp. 309–336. 2009. ‘Conventionalism and the Contingency of Conventions,’ Noûs 43.2, 224–441. 2010. ‘Modality and Objects,’ The Philosophical Quarterly 60.238, 109–125. Sider, T. Ontological Realism,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.), 2009. Metametaphysics. Oxford University Press, pp. 384–423. 2011. Writing the Book of the World. Oxford University Press. 2014. ‘Hirsch’s Attack on Ontologese,’ Noûs 48.3, 565–572. Simons, P. 1987. Parts: A Study in Ontology. Oxford: Clarendon Press. Smith, B. and A. C. Varzi. 1997. ‘Fiat and Bona Fide Boundaries,’ Philosophy and Phenomenological Research 60.2, 401–420. Smith, B. and K. Mulligan. 1983. ‘Framework for Formal Ontology,’Topoi 3, 73–85. Smolin, L. 2006. The Trouble with Physics. London: Penguin Books. Sorensen, R. 1992. Thought Experiments. Oxford University Press. Sosa, E. 2014. ‘Intuitions: Their Nature and Probative Value,’ in A. R. Booth and D. P. Rowbottom (eds.), Intuitions. Oxford University Press, pp. 36–49. Stoljar, D. ‘Physicalism,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2015 edn, see http://plato.stanford.edu/archives/spr2015/entries/ physicalism/ Su, G. and M. Suzuki. 1999. ‘Towards Bose–Einstein Condensation of Electron Pairs: Role of Schwinger Bosons,’ International Journal of Modern Physics B 13.8, 925–937. Szabó, Z. G. 2003. ‘Believing in Things,’ Philosophy and Phenomenological Research 66, 584–611. Tahko, T. E. 2008. ‘A New Definition ofA Priori Knowledge: In Search of a Modal Basis,’ Metaphysica 9.2, 57–68. 2011. ‘ A Priori and A Posteriori: A Bootstrapping Relationship,’ Metaphysica 12.2, 151–164. 2012. ‘Boundaries in Reality,’ Ratio 25.4, 405–424. (ed.). 2012. Contemporary Aristotelian Metaphysics. Cambridge University Press. 2012. ‘Counterfactuals and Modal Epistemology,’ Grazer Philosophische Studien 86, 93–115. 2012. ‘In Defence of Aristotelian Metaphysics,’ in T. E. Tahko (ed.), Contemporary Aristotelian Metaphysics. Cambridge University Press, pp. 26–43. 2013. ‘Truth-Grounding and Transitivity,’ Thought: A Journal of Philosophy 2.4, 332–340. 2014. ‘Boring Infinite Descent,’ Metaphilosophy 45.2, 257–269. 2015. ‘Natural Kind Essentialism Revisited,’ Mind 124.495, 2015, 795–822.

Bibliography

2015. ‘The Modal Status of Laws: In Defence of a Hybrid View,’ The Philosophical Quarterly, (online) (forthcoming). ‘Empirically-Informed Modal Rationalism,’ in R. W. Fischer and F. Leon (eds.), Modal Epistemology After Rationalism, Synthese Library. Dordrecht: Springer (forthcoming). Tahko, T. E. and E. J. Lowe. ‘Ontological Dependence,’ in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2015 edn), see http://plato.stanford.edu/entries/dependence-ontological/ Tahko, T. E. and D. O’Conaill. 2015. ‘Minimal Truthmakers,’ Pacific Philosophical Quarterly, (online) (forthcoming). Tallant, J. 2013. ‘Intuitions in Physics,’ Synthese 190, 2959–2980. Thomas, C. 2008. ‘Speaking of Something: Plato’s Sophist and Plato’s Beard,’ Canadian Journal of Philosophy 38.4, 631–667. Thomasson, A. 2007. Ordinary Objects. Oxford University Press. 2009. ‘Answerable and Unanswerable Questions,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.), Metametaphysics. Oxford University Press, 444–471. 2009. ‘The Easy Approach to Ontology,’ Axiomathes 19, 1–15. 2015. Ontology Made Easy. Oxford University Press. Trogdon, K. 2009. ‘Monism and Intrinsicality,’ Australasian Journal of Philosophy 87, 127–148. 2013. ‘Grounding: Necessary or Contingent?’, Pacific Philosophical Quarterly 94, 465–485. 2013. ‘An Introductionto Grounding,’ in M. Hoeltje,B. Schnieder, andA. Steinberg (eds.), Varieties of Dependence. Munich:Philosophia Verlag, pp. 97–122. Vaidya, A. 2010. ‘Understanding and Essence,’ Philosophia 38, 811–833. 2012. ‘Intuition and Inquiry,’ Essays in Philosophy 13.1, Article 16. Van Brakel, J. 2010. ‘Chemistry and Physics: No Need for Metaphysical Glue,’ Foundations of Chemistry 12.2, 123–136. Van Fraassen, B. C. 2002. The Empirical Stance . New Haven: Yale University Press. Van Inwagen, P. 1990. Material Beings. Ithaca, NY: Cornell University Press. 1998. ‘Meta-ontology,’ Erkenntnis 48, 233–250. 2009. ‘Being, Existence, and Ontological Commitment,’ in D. Chalmers, D. Manley, and R. Wasserman (eds.), Metametaphysics. Oxford University Press, pp. 472–506. Varzi, A. C. 2011. ‘Boundaries, Conventions, and Realism,’ in J. K. Campbell, M. O’Rourke, and M. H. Slater (eds.), Carving Nature at Its Joints: Natural Kinds in Metaphysics and Science. Cambridge, MA: MIT Press, pp. 129–153. Von Solodkoff, T. and R. Woodward. 2013. ‘Noneism, Ontology, and Fundamentality,’ Philosophy and Phenomenological Research 87.3, 558–583.

253

254

Bibliography

Williamson, T. 2007. The Philosophy of Philosophy. Oxford: Blackwell Publishing. 2013. ‘How Deep is the Distinction Between A Priori and A Posteriori Knowledge?’ in A. Casullo and J. C. Thurow (eds.), The A Priori in Philosophy. Oxford University Press, pp. 291–312. Wilson, J. 2014. ‘No Work for a Theory of Grounding,’ Inquiry 57.5–6, 1–45. Woodward, R. 2013. ‘Towards Being,’ Philosophy and Phenomenological Research 86.1, 183–193. Yablo, S. 1993. ‘Is Conceivability a Guide to Possibility?’, Philosophy and Phenomenological Research, 53, 1–42. 2009. ‘Must Existence-Questions have Answers?’, in D. Chalmers, D. Manley, and R. Wasserman (eds.), Metametaphysics. Oxford University Press, pp. 507–525.

Index

a posteriori necessity, 80, 154, 164

constitutive facts, 161–63, 229, 230, 236

allism, 22, 236. See also noneism

conventional, conventionalism,9, 63, 65,

analyticity, 27, 73, 80, 113, 155, 183, 191 application conditions, 74–76, 236 apriority, 10, 11, 80, 146, 151, 152–55, 159, 164, 168–76, 178, 180, 181, 190, 193, 203, 206, 217, 231, 232 Aristotle, Aristot elian, 3, 4, 9, 46, 57,

76–83, 86, 89, 90, 169, 233, 236, 238, 240 Correia, Fabrice, 108 counterfactuals, 81, 82, 160, 161, 162, 167, 229, 230, 231, 236 Crane, Tim, 23, 25–27, 44, 46, 47, 62

64, 67, 93, 100, 101, 103, 127, 128, 163, 195, 196, 206, 207, 208, 209,

definite descriptions, 19, 49

240, 241

Dennett, Daniel, 177

armchair metaphysics, 152, 169, 170, 171, 193, 194, 206

dependence. See ontological dependence Dorr, Cian,32, 33, 34, 35, 220

Armstrong, D. M., 137 atomism, 120, 125–30, 133, 134, 136, 139, 142, 236, 238

'easy' ontology,65, 66, 73, 74, 75, 91 Einstein, Albert, 88, 178, 196 Eklund, Matti, 29, 37

Bealer, George, 156, 178, 190–93

emergence, emergent, 143, 146, 208, 237

Bennett, Karen, 68–71, 72, 77, 89, 135

endurantism and perdurantism, 53–55,

Bohm, David,129, 141

187, 237. See also three- and fourdimensionalism

Cameron, Ross, 130–33

epistemicism, 69, 70, 71, 72, 77, 83

Carnap, Carnapian, 8, 13–15, 20, 27, 28, 29, 31, 32, 33, 35, 37, 38, 41, 49, 50,

essence, 82, 95, 98, 100–4, 115, 152, 156, 157, 162–66, 169, 173, 174, 225, 232,

51, 53, 64, 75, 76, 203, 211, 237, 238 causation, 9, 56, 93, 94, 108, 111, 113, 114, 144, 185, 186, 187, 218, 229, 230, 235, 238 Chalmers, David, 5, 14, 32, 68, 73, 156, 158, 184 chemical substance, 125, 174, 180, 202, 233 conceptual analysis, 169, 182, 183, 200, 206, 217, 236, 237

233, 234, 236, 237, 239, 240 essentialism, 100, 156, 235, 237 existence, 6, 8, 10, 13, 15, 16–29, 32, 35, 37, 39, 40, 41, 42, 44, 46, 47, 51, 57–62, 69, 73, 74, 75, 85, 91, 93, 95–100, 102, 104, 110, 112, 120, 131, 136, 144, 166, 210, 221, 237, 240 external questions. See internal and external questions

255

256

Index

factivity, 29, 35 fallibilism, 85, 167, 173, 182, 191, 192, 215, 218, 222, 237 Fine, Kit, 8, 40, 57–63, 67, 71, 73, 98, 101, 102, 103, 104, 105, 106, 107, 108, 113, 115, 116, 118, 139, 156, 162, 163, 165, 225 fundamental physics, 7, 86, 124, 129, 143, 186, 205, 208, 219, 222 fundamentality, 9, 11, 34, 41, 43, 62, 79, 83, 85, 86, 91, 105, 110, 115, 119, 120–51, 204, 208, 218, 221, 223, 226, 229, 238, 239, 241, 242

Ladyman, James, 11, 123, 125–27, 128, 129, 132, 133, 137, 140, 141, 142, 144, 193, 194, 213, 217–23, 228 language pluralism, 8, 13, 15, 29, 35–39, 53, 76, 79, 238 levels (of reality), 9, 83, 107, 120–50, 152, 221, 238 Lewis, David, 21, 22, 23, 240 logical empiricism. See logical positivism logical positivism, 14, 155, 203, 204, 205, 219 Lowe, E. J., 5, 98, 101, 102, 104, 105, 115, 156, 163, 164, 165, 166, 169, 170, 172, 173, 216, 217, 218, 225, 231, 232, 233,

general relativity, 132, 215

234, 235

ground, grounding, 9, 62, 91, 93, 94, 104–19, 127, 130, 134, 135, 138, 140,

Markosian, Ned, 130

156, 165, 227, 231, 238, 242

Maudlin, Tim, 204

gunk, 128, 130, 132, 133, 134, 138, 238

McDaniel, Kris, 127 McTaggart, John, 20

Hawley, Katherine, 220, 221 Hawthorne, John, 131, 132

Meinong, Alexius, 8, 15, 20–27, 39, 46, 240

Heil, John, 147

Meinongianism, 20–27, 39, 41

Hiley, Basil J., 130

mentality, mental states, 93, 113, 114,

Hirsch, Eli, 8, 9, 14, 36, 40, 50–57, 62, 65, 66, 68, 69, 71, 72, 73, 78

116, 153, 154, 198 mereological atomism. See atomism

Hofweber, Thomas,30–32, 36

mereology, 9, 33, 55, 120, 238, See also

identity (criteria of), 26, 77, 78, 100, 101,

metaontology, 1, 3–6, 7, 8, 11, 15, 16, 17,

atomism 165, 166, 173, 175, 222 infinite regress. See regress

18, 24, 25, 35, 38, 39, 40, 43, 49, 50, 53, 54, 57, 59, 60, 62, 63, 70

infinitism, 10, 122, 133, 140, 147, 148, 239

metaphysical infinitism. See infinitism

internal and external questions, 8, 15,

microstructure, 81, 115, 173–76, 180, 183,

27–35, 37, 76, 237, 238 intuitions, 10, 79, 112, 131, 132, 148, 156,

202 Miller, Kristie, 54

172, 177–94, 198, 199, 202, 203, 210, 217, 221, 232, 237, 239, 240. See also rational intuition

modal empiricism,10, 152, 166, 167–72, 239 modal rationalism, 10, 152, 155–63, 166, 167, 168, 172, 232, 239

Jackson, Frank, 200, 201

modality

Jenkins, Carrie, 68, 167–70, 181

conceptual, 157, 191, 236

joint-carving, 50, 57, 66, 68, 77, 79, 83–90,

metaphysical, 95, 97, 115, 146, 158, 165,

138, 140, 205, 238

174, 175, 176, 180, 210, 224, 230, 236, 239

Kim, Jaegwon, 132 Koslicki, Kathrin, 102, 103, 104 Kripke, Saul, 163, 173

physical, 144, 239 monism, 128, 133, 137, 208, 239. See also pluralism

Index

natural kinds, 89, 166, 169, 172, 173, 174, 205, 232, 237, 239 natural necessity. See modality:physical naturalism, 85, 191, 207, 212, 225, 239

principle of charity, 52, 240 principle of independence, 21, 23, 25, 240 Putnam, Hilary, 81, 124, 125, 173, 177, 179, 180, 183, 200

necessity. See modality nihilism (about composition), 34, 35, 55, 187, 239, See also universalism (about composition) Nolan, Daniel, 169–72 nominalism, 17, 53 nomological necessity.See modality: physical noneism, 22, 23, 24, 25, 39, 236, See also

allism

quantification, quantifiers, 4, 6, 8, 14, 16–18, 19, 21, 23, 26, 27, 30, 34, 35, 36, 39–52, 56, 57, 58, 59, 60, 61, 62, 66, 68, 69, 85, 228 quantifier variance, 8, 9, 36, 40, 42, 49–57, 66, 68, 69, 73, 75, 77, 240 quantum field theory, 132, 148 quantum mechanics, 28, 125, 129, 132, 148, 197, 204, 214, 221, 222 Quine, Quinean, 4, 5, 8, 13–27, 30, 35, 38,

ontic structural realism. See structuralism

39, 40, 41, 42, 44, 45, 46, 47, 48, 49,

Ontologese, 32, 49, 56, 239

50, 57, 59, 64, 138, 155, 191, 192, 203,

ontological anti-realism, 9, 64, 72, 76

216, 217, 228

ontological commitment, 4, 6, 8, 14, 16, 18, 19, 23, 39–49, 58, 63, 64, 227 ontological deflationism, 9, 14, 36, 62, 64,

rational intuition, 11, 190–94, 240, See also intuitions

65, 66, 67, 71–76, 78, 79, 84, 85, 90,

Rayo, Augustín, 43, 44, 45, 47, 51

91, 92, 236

real (metaphysically robust), 58, 60, 61,

ontological dependence essential, 95, 98, 100, 101–4, 165, 237 generic existential, 95, 96, 97, 135, 146 identity-dependence, 95, 98–101, 104, 223, 238 rigid existential, 96, 97, 102, 110, 136 ontological realism, 2, 9, 50, 64–73, 77,

233 real definition, 101, 104, 163, 164, 165, 237, 240 reduction, reductionism, 9, 13, 52, 94, 102, 113, 114–16, 124–26, 152, 167, 168, 176, 188, 189, 223, 225, 237, 240 reference magnetism, 86, 87 regress, 131, 139, 148, 149, 163, 214, 238

83, 85, 89, 240, See also ontological

relativism, 29, 72, 241

anti-realism; ontological deflationism

Ross, Don, 11, 123, 125–27, 128, 129, 132,

paraphrase, 19, 49

133, 137, 140, 141, 142, 144, 193, 194, 213, 217–23

Parsons, Terence,24

Routley, Richard, 24, 25

Paul, L.A., 225–31, 234, 235

Rowbottom, Darrell, 178

physical constants, 136, 144–47

Russell, Bertrand, 19, 20, 24, 45, 46

Plato, Platonism, 8, 15, 18, 19, 20, 36, 53, 59, 84 pluralism (as opposend to monism), 105, 128, 133, 137, See also monism

Schaffer, Jonathan,62, 114, 128, 133, 135, 138, 140, 141, 208 Schrödinger, Erwin, 129

possibility. See modality

Sidelle, Alan, 80–83

possible worlds, 23, 25, 116, 157, 180, 224,

Sider, Ted,8, 9, 33, 34, 40, 49, 50, 54,

230, 236, 240 powers, 56 Priest, Graham, 21, 24, 25, 45–47

56, 57, 62, 65, 66, 67, 68, 69, 71, 73, 83–90, 138 Simons, Peter, 99

257

258

Index

Smith, Barry, 78 Smolin, Lee, 214

Twin Earth, 81, 177, 179, 180, 183, 184, 192, 200, 201

Sorensen, Roy,199, 200, 201 Sosa, Ernest, 183, 184, 193 special composition question, 33, 34, 35, 55, 241 structuralism, 79, 223, 241 structure, 77, 79, 83, 84, 119, 120, 121,

universalism (about composition), 34, 55, 56, 75, 187, 241. See also nihilism (about composition) universals, 17, 18, 19, 20, 22, 32, 208, 215, 233, 236

124, 127, 131, 136, 137, 138, 139, 140, 146, 147, 185, 204, 205, 206, 217, 223, 227, 228

vague, vagueness, 4, 5, 69, 153, 172, 181, 202, 207, 241

substance, 95, 99, 128, 215, 227, 239, 241

Vaidya, Anand, 156, 186, 187

supervenience, 113, 116, 132

Van Fraassen, Bas, 211–17, 218

Szabó, Zoltán Gendler, 47, 48

Van Inwagen, Peter, 4–6, 15, 16, 33, 42, 50,

theoretical virtues, 170, 171, 228, 230,

Varzi, Achille C., 78–80

55, 70, 170, 241 231, 235, 241 Thomas, Christie, 18, 19

verbal disputes, 50, 51–57, 62, 69, 70, 71, 72, 75, 240

Thomasson, Amie, 71, 72, 73–76, 78, 91 thought experiments, 10, 177–81, 185, 186, 192, 194–202, 229, 232 three- and four-dimensionalism, 53, 55, 66, 241. See also endurantism and perdurantism

well-foundedness, 123, 133–35, 136, 150, 242 Williamson, Timothy, 155, 160–63, 167, 168, 229, 230, 231 Woodward, Richard, 22

Trogdon, Kelly, 118 truthmaking, 9, 94, 116–19, 137, 139, 241

Yablo, Stephen, 70, 158