Form, Matter, Substance 9780198823803

Table of contents :
Dedication
Contents
Acknowledgments
Introduction
Part I. A Hylomorphic Analysis of Concrete Particular Objects
1. Concrete Particular Objects
2. Matter
3. Form
4. Hylomorphic Relations
Part II. Substance
5. Ontological Dependence
6. Independence Criteria of Substancehood
7. Unity
8. Artifacts
Conclusion
Bibliography
Index of Names
General Index

Citation preview

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

Form, Matter, Substance

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

Form, Matter, Substance Kathrin Koslicki

1

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

3 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Kathrin Koslicki 2018 The moral rights of the author have been asserted First Edition published in 2018 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2018935623 ISBN 978–0–19–882380–3 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

To Bob

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

Contents Acknowledgments

xi

Introduction

1

Part I. A Hylomorphic Analysis of Concrete Particular Objects 1. Concrete Particular Objects

11 11 11 12 13 15 19 30

2. Matter

32 32 33 33 35 37 45 45 48 51 57 61

3. Form

62 62 63 63 64 66 71 73 76 77 82 89 103

1.1 Introductory Remarks 1.2 Constituent vs. Non-Constituent Ontologies 1.3 Non-Constituent Ontologies: Platonism and Austere Nominalism 1.4 Constituent Ontologies: Bundle Theories and Substratum Theories 1.5 Difficulties for Non-Hylomorphic Competitors 1.6 Desiderata and Decision Points for Hylomorphists 1.7 Conclusion 2.1 Introductory Remarks 2.2 Matter as Prime Matter 2.2.1 Prime Matter in Aristotle 2.2.2 Thomistic Prime Matter 2.2.3 Difficulties for Oderberg’s Thomistic Prime Matter Hypothesis 2.3 Matter as Stuff 2.3.1 Thomistic Prime Matter as Stuff 2.3.2 Difficulties for Brower’s Thomistic Prime Matter as Stuff Hypothesis 2.3.3 Difficulties for Markosian’s Matter as Stuff Hypothesis 2.4 The Hylomorphic Conception of Matter 2.5 Conclusion 3.1 Introductory Remarks 3.2 The Ontological Category of Form 3.2.1 Areas of Agreement and Disagreement 3.2.2 The Universal Forms Hypothesis 3.2.3 The Individual Forms Hypothesis 3.2.4 The Hybrid Position 3.3 The Individual vs. Universal Forms Debate 3.4 In Defense of Individual Forms 3.4.1 Causation, Explanation, and Change 3.4.2 Essence and Accident 3.4.3 Identity and Indiscernibility 3.5 Conclusion

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

viii  Contents 4. Hylomorphic Relations 4.1 4.2 4.3 4.4

Introductory Remarks The Grounding Problem Sidelle’s Challenge to Hylomorphists Towards a Hylomorphic Solution to the Grounding Problem 4.4.1 The Scenario 4.4.2 Hylomorphic Monism 4.4.3 Non-Robust Hylomorphic Pluralism 4.4.4 Robust Hylomorphic Pluralism: A Compositional Approach 4.4.4.1 Robust Mereological Hylomorphic Pluralism: The Basics 4.4.4.2 Essential Differences between Coincident Objects 4.5 Explanatory Work for Non-Modal Essences 4.6 Conclusion

104 104 105 107 109 109 111 112 113 114 116 120 123

Part II.  Substance Introduction to Part II  5. Ontological Dependence

125 135 135 139 139 142 142 143 144 144 146 149 150 150 154 159 160 162

6. Independence Criteria of Substancehood

163 163 164 167 171 172 173 175 176 177 177

5.1 Preliminaries 5.2 Varieties of Existential Dependence 5.2.1 Ontological Dependence in Aristotle’s Categories 5.2.2 Modal Existential Dependence 5.2.2.1 Rigid Existential Necessary Dependence 5.2.2.2 Generic Existential Necessary Dependence 5.2.3 Other Forms of Existential Dependence 5.2.3.1 Necessary vs. Essential Existential Dependence 5.2.3.2 Rigid and Permanent Existential Dependence 5.2.4 Being vs. Existence 5.3 Varieties of Essential Dependence 5.3.1 Essential Identity Dependence 5.3.2 Constitutive Essential Dependence 5.3.3 Constitutive Definitional Dependence 5.3.4 A Potential Difficulty 5.4 Conclusion 6.1 Introductory Remarks 6.2 Substancehood 6.3 Independence Criteria of Substancehood 6.4 Alternative Strategies 6.4.1 Lowe’s Account 6.4.2 Gorman’s Modifications of Lowe’s Account 6.4.3 The Stipulative Exclusion of Non-Particulars 6.4.4 The Stipulative Exclusion of Proper Parts 6.4.4.1 The Possibility of Simple Substances Simpliciter 6.4.4.2 The Threat of Heterogeneity

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

Contents  ix 6.4.4.3 Proper Parts vs. Constituents 6.4.4.4 Intrinsicness 6.4.5 Matter–Form Compounds 6.4.5.1 Lowe’s Strategy 6.4.5.2 Gorman’s Exemption for Part-Dependence 6.4.5.3 Form as Principle of Unity 6.5 The Status of Forms 6.6 Conclusion

179 182 185 185 186 187 188 190

7. Unity

191 191 192 197 206 208 214

8. Artifacts

216 216 217 220 220 222 224 226 227 230 234 235 237 239

Conclusion

241

Bibliography Index of Names General Index

247 265 269

7.1 Introductory Remarks 7.2 Wholeness, Structure, and Unity 7.3 Wholeness, Integrity, and Unity 7.4 Ontological Dependence 7.5 Interactional Dependence 7.6 Conclusion 8.1 Introductory Remarks 8.2 Artifact Kinds and Natural Kinds 8.3 Artifact-Essences 8.3.1 Maker’s Intentions 8.3.2 Creative Acts 8.3.3 Functions 8.4 Challenges for Author-Intention-Based Accounts of Artifact-Essences 8.4.1 User-Intentions 8.4.2 Easy Ontology 8.4.3 Further Objections: Mass-Production and Scope 8.4.4 The Limits of Human Creative Intentions 8.5 Challenges for Existing Anti-Essentialist Frameworks 8.6 Conclusion

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

Acknowledgments The material in this book has been presented over the years in numerous talks, workshops, conferences, and seminars and has benefited greatly from the help of many friends, colleagues, and students. I would like to express my thanks to Christof Rapp for organizing a very stimulating workshop on my book manuscript at MUSAΦ (Munich School of Ancient Philosophy), Ludwig Maximilians Universität, München, Germany, in the fall of 2017. I learned a great deal from the very thoughtful comments delivered by Michail Peramatzis at this workshop as well as from the contributions of the other workshop participants. In the spring of 2016, I had the pleasure of traveling to Istanbul, Turkey, to participate in a very thought-provoking workshop on Aristotelian Metaphysics and my book manuscript at Bogaziçi University. I am grateful to Mark Steen for organizing this workshop and to the other participants, who traveled to Istanbul for this event, for their helpful and interesting comments. I would like to acknowledge the support of the following foundations and institutions which had a hand in making this research possible. Portions of this research were undertaken during my tenure as the Tier I Canada Research Chair in Epistemology and Metaphysics at the University of Alberta, which began in January 2014. I am grateful to the Canada Research Chairs program for its support. The University of Alberta and the Kule Institute for Advanced Study have also been very generous in providing funding for various research initiatives related to this book project: in 2017, the Kule Institute for Advanced Study awarded me a Dialogue Grant for the project, “2017 Hylomorphism Conference: an examination of the Aristotelian doctrine of matter (hylē) and form (morphē),” which was also funded in part through a SSHRC (Social Science and Humanities Research Council of Canada) Connection Grant; in 2014/15, the University of Alberta awarded me a Killam Research Fund Connection Grant and the Kule Institute for Advanced Study awarded me a Dialogue Grant for the project, “Artifacts and Metaphysical Explanation.” I am grateful to the Social Science and Humanities Research Council of Canada, the University of Alberta, and the Kule Institute for Advanced Study, as well as its director, Geoffrey Rockwell, for their support. During the academic year of 2012/13, I held an Alvin Plantinga Fellowship at the University of Notre Dame’s Center for Philosophy of Religion. I would like to thank the co-directors, Mike Rea and Sam Newlands, for inviting me to spend a very enjoyable year at the Center, and all those at Notre Dame during that year who helped create such an excellent research environment. Finally, I would like to express my sincere thanks to the National Endowment of the Humanities for awarding me a fellowship during the academic year of 2010/11. Portions of this book are based on previously published work as follows. Chapter 3 incorporates material from Koslicki (2014), “The Causal Priority of Form in Aristotle,”

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

xii  Acknowledgments in Studia Philosophica Estonica, Special Issue: “Aristotelian Metaphysics: Essence and Ground,” edited by Riin Sirkel and Tuomas E. Tahko, Vol. 7.2 (2014), pp. 113–41; and Koslicki (2018a), “Essence and Identity,” (forthcoming) in Metaphysics, Meaning and Modality: Themes from Kit Fine, edited by Mircea Dumitru, Oxford University Press, Oxford. Chapter 4 incorporates material from Koslicki (2012b), “Essence, Necessity and Explanation,” in Contemporary Aristotelian Metaphysics, edited by Tuomas Tahko, Cambridge University Press, Cambridge, pp. 187–206; and Koslicki (2018b), “Towards a Hylomorphic Solution to the Grounding Problem,” to appear in Royal Institute of Philosophy Supplements to Philosophy, Cambridge University Press, Cambridge. Chapter  5 incorporates material from Koslicki (2012a), “Varieties of Ontological Dependence,” in Metaphysical Grounding: Understanding the Structure of Reality, edited by Fabrice Correia and Benjamin Schnieder, Cambridge University Press, Cambridge, pp. 186–213; and Koslicki (2013a), “Ontological Dependence: An Opinionated Survey,” in Varieties of Dependence: Ontological Dependence, Grounding, Supervenience, Response-Dependence (Basic Philosophical Concepts), edited by M. Hoeltje, B. Schnieder, and A. Steinberg, Philosophia Verlag, München, pp. 31–64. Chapter 6 incorporates material from Koslicki (2013b), “Substance, Independence and Unity,” in Aristotle on Method and Metaphysics, edited by Edward Feser, Palgrave Macmillan, Basingstoke, pp. 169–95; and Koslicki (2015b), “In Defense of Substance,” in Themes from Ontology, Mind, and Logic: Present and Past, Essays in Honour of Peter Simons, Grazer Philosophische Studien, Vol. 91 (2015), edited by Sandra Lapointe, Brill Rodopi, Leiden, pp. 59–80. I am grateful to Brill, Cambridge University Press, Palgrave Macmillan, Philosophia Verlag, Oxford University Press, and Studia Philosophica Estonica for giving me permission to reproduce this previously published material here. Over the years, I have greatly benefited from feedback I have received on my work  from many colleagues and students, including (but not limited to): Andreas Anagnostopoulos, Lucas Angioni, Paul Audi, Sina Azizi, Ralf Bader, Mark Balaguer, Karen Bennett, Sara Bernstein, Philipp Blum, Ingo Brigandt, Teresa Britton, Jeff Brower, Danielle Brown, Margaret Cameron, Anjan Chakravartty, Rebecca Chan, David Charles, Eli Chudnoff, Phil Corkum, Fabrice Correia, Andrew Cortens, Richard Cross, Shamik Dasgupta, Heather Demarest, Louis deRosset, Joseph Diekemper, Maureen Donnelly, Travis Dumsday, Antoine Dussault, Brian Epstein, Simon Evnine, Kit Fine, Mark Fiocco, Graeme Forbes, Christopher Frey, Martin Glazier, Bob Hale, Ata Hashemi, Pieter Sjoerd Hasper, Katherine Hawley, Allen Hazen, Chris Heathwood, Thomas Hofweber, Paul Hovda, Ferenc Huoranszki, Ross Inman, Nurbay Irmak, Christoph Jäger, Ludger Jansen, Thomas Johansen, Christian Kanzian, Boris Kment, Dan Korman, David Kovacs, Irem Kurtsal, Ka Ho Lam, Sandra Lapointe, Hannah Laurens, Mitzi Lee, David Liebesman, Bernie Linsky, Dan Lopez de Sa, Alan Love, Jonathan Lowe, Josep Macià, Marko Malink, Wolfgang Mann, Ned Markosian, Sanna Mattila, Kris McDaniel, Andrew McFarland, Michaela McSweeney, Chris Menzel, Trenton Merricks, Friederike Moltmann, Kevin Mulligan, Sam Newlands, Bruno Niederbacher, Howard Nye, David Oderberg, Marilú Papandreou, Bob Pasnau, Ville Paukkonen, Laurie Paul,

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

Acknowledgments  xiii Tim Pawl, Jeff Pelletier, Michail Peramatzis, Christian Pfeiffer, Christof Rapp, Mike Raven, Mike Rea, Esther Rosario, Gideon Rosen, Liva Rotkale, Noël Saenz, Thomas Sattig, Theodore Scaltsas, Jonathan Schaffer, Amy Schmitter, Benjamin Schnieder, Alan Sidelle, Ted Sider, Ori Simchen, Peter Simons, Riin Sirkel, Alex Skiles, Jeff Snapper, Mark Steen, Leopold Stubenberg, Mack Sullivan, Meghan Sullivan, Tuomas Tahko, Amie Thomasson, Patrick Todd, Niels Tolkhien, Christopher M. P. Tomaszewski, Patrick Toner, Sirio Trentini, Kelly Trogdon, Christina van Dyke, Peter van Inwagen, Achille Varzi, Barbara Vetter, Meg Wallace, Gloria Wasserman-Frost, Ken Waters, Marcel Weber, Kelly Weirich, Jessica Wilson, Bill Wimsatt, Andrea Woody, Lianghua Zhou, Jack Zupko, and Justin Zylstra. I would like to express my appreciation to the founding members of the Canadian Metaphysics Collaborative () for their support in helping me create this initiative in 2015 to facilitate collaboration among Canadian and Canadianaffiliated metaphysicians: Margaret Cameron, Phil Corkum, Carrie Jenkins, Mike Raven, Ori Simchen, Jessica Wilson, and Jack Zupko. I thank Margaret Cameron and Mike Raven, in particular, for serving with me on the CMC Steering Committee and as co-editors-in-chief of our new peer-reviewed, open-access, online journal, Metaphysics (). Jeff Brower, Tuomas Tahko, and an anonymous reader for Oxford University Press deserve special thanks for their extensive, helpful, and interesting comments on the penultimate draft of this manuscript. I have also benefited greatly from exchanges about this project with David Oderberg and Simon Evnine. In addition, I am very grateful to Ka Ho Lam for his outstanding work as my graduate research assistant at the University of Alberta from spring 2015 to summer 2017. Peter Momtchiloff has been an absolute pleasure to work with on this and my previous book; we all owe him much appreciation for the amazing work he has done over the years for the philosophical community as a whole as being the philosophy editor for Oxford University Press. My deepest gratitude, however, is due to Bob Semborski for his love and patience during these past ten years while I was working on this book; I dedicate this book to him.

OUP CORRECTED PROOF – FINAL, 07/24/2018, SPi

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Introduction The Aristotelian doctrine of hylomorphism holds that those entities which are subsumed under it are compounds of matter (hulē) and form (morphē or eidos). Aristotle first introduced the distinction between matter and form in response to a problem concerning the possibility of change he had inherited from his pre-Socratic predecessors, according to which change is logically impossible since it requires a transition from nothing to something or from something to nothing. In response, Aristotle argued in Physics I that change is possible after all, since something (viz., the matter) always persists through any change that occurs, while something else (viz., the form or privation) does not persist through the change. For example, when some wood (the matter) goes from not being a bed (the privation) to being a bed (the form), the wood remains throughout the change, but the wood’s not being a bed does not persist through the change. After Aristotle’s initial application of the doctrine of hylomorphism to the problem of change, the distinction between matter and form became a cornerstone of his approach to nearly every central phenomenon discussed elsewhere in the corpus, including in his treatises on logic, natural science, metaphysics, theology, ethics, and political philosophy. Since Aristotle’s introduction of the doctrine of hylomorphism, philosophers have continued to engage with it and have found it to be a highly flexible, effective, and powerful tool with which to engage sympathetically or critically across a wide range of domains, including applied ethics, philosophy of biology, continental philosophy (in particular, the philosophy of technology), ethics, feminist philosophy, logic, ­philosophy of mind, metaphysics, philosophy of religion, and philosophy of science. To illustrate, in applied ethics, the hylomorphic conception of human beings as composed of a body (the matter) configured by a soul (the form) impacts questions of life and death, e.g., as they arise in connection to hospice care and abortion (Hershenov and Koch (2005); Spencer (2010)). In the philosophy of biology, Aristotelian views concerning taxonomic classification as well as the mechanism of genetic inheritance between offspring and parent are still debated today (Henry (2006a, 2006b); Jaworski (2016a); Lennox (2009); O’Rourke (2004); Vinci and Robert (2005)). In continental philosophy, hylomorphism is taken to contribute a certain model of individuation, as resulting from the combination of form with matter (Deleuze and Guattari (1987); Simondon (1964)). In ethics, hylomorphists tend to be attracted to a virtue-centered

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

2  introduction conception of the good life for humans as focused on the exercise of a range of ­characteristic or unique capacities of the human soul, e.g., practical deliberation and action or theoretical contemplation (Barney 2008); Johnson (2005); Korsgaard (2008); MacIntyre (2007); Nussbaum (2014); Reeve (2009, 2012); Shields (2015)). Feminist philosophers have argued that Aristotle’s doctrine of hylomorphism is not a neutral thesis concerning the natural world, but an inherently normative claim that leads to troublesome conceptions of gender and the status of women in society (Bar On (1994); Connell (2016); Freeland (1998); Witt (2004)). In logic, hylomorphism is relevant to the thesis that the subject matter of this discipline is purely formal (Novaes (2012a, 2012b)). In the philosophy of mind, hylomorphism has been heralded as yielding an attractive middle ground between dualism and physicalism (Jaworski (2016b); Mouracade (2008); Pasnau (2012); Shields (2009)). Metaphysicians have appealed to the distinction between matter and form in the service of defending a structure-based theory of parts and wholes (Fine (1999); Johnston (2002, 2006); Koslicki (2008a)). In the philosophy of religion, hylomorphic conceptions of the human person are taken to have implications for the doctrines of resurrection and personal immortality (Brower (2014); Stump (1995)). In the philosophy of science, the Thomistic reading of hylomorphism, as involving a commitment to prime matter, has been taken to be surprisingly compatible with conceptions of matter found in modern physics (Nicolescu (2000); Oderberg (2007); Spencer (2016)). This study focuses on a particular application of the doctrine of hylomorphism within the domain of metaphysics, viz., to the specific case of concrete particular objects. Concrete particular objects (e.g., individual living organisms) figure saliently in our everyday experience as well as in our scientific theorizing about the world. Since they are concrete rather than abstract, they are located both in space and time (unlike, say, the number 2). Since they are particular, rather than universal, they cannot be wholly present in multiple different regions of space at a single time (unlike, say, the property of redness which is shared by tomatoes, fire trucks, and sunsets). And since they are objects rather than events or properties, they are wholly present at every time at which they exist (unlike, say, a football game); they are capable of persisting through intrinsic change; and they are the bearers of characteristics while not themselves characterizing other entities (unlike, say, the property of redness). In what follows, I develop and defend a hylomorphic approach to the metaphysics of concrete particular objects. Such an account, I will argue, is well equipped to compete with alternative approaches when it is measured against familiar criteria of success. Furthermore, hylomorphism is designed to meet further challenges which are less widely emphasized in the recent literature on these topics. A successful development of this doctrine, however, hinges on how hylomorphists conceive of the matter composing a concrete particular object; its form; and the hylomorphic relations which hold between a matter–form compound and its form (viz., the “compound–form relation”), between the matter and the form (viz., the “matter–form relation”), and between a matter–form compound and its matter (viz., the “matter–compound” relation or

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

introduction  3 “material constitution”). In addition, we would ideally want hylomorphists to clarify the further commitments and the explanatory value which are associated with the application of this doctrine to the specific case of concrete particular objects. We will have occasion to examine all of these questions in detail in what follows. I begin Part I of this study (“A Hylomorphic Analysis of Concrete Particular Objects,” Chapters 1–4) by situating hylomorphism with respect to existing approaches to the metaphysics of concrete particular objects. After setting out a series of decision points, desiderata, and options that are available to hylomorphists, I argue for a particular conception of the matter composing a concrete particular object, its form, and the hylomorphic relations between a matter–form compound, its matter and its form. In Part II (“Substance,” Chapters 5–8), I turn to the question of whether concrete particular objects, when construed as matter–form compounds, deserve a privileged position of some kind within a hylomorphic ontology, despite their metaphysical complexity, and, if so, according to what criterion or criteria of “ontological privilege” or substancehood. I go on to present my own proposal as to how Aristotelians should proceed, in a way that is maximally consistent with their other central commitments. This proposal is intended to help resolve the apparent tension which apparently exists between, on the one hand, attributing to concrete particular objects a hylomorphic structure and, on the other hand, assigning these entities to the ontologically privileged category of substances. In Chapter 1 (“Concrete Particular Objects”), I review existing approaches to the metaphysics of concrete particular objects and position the doctrine of hylomorphism, as applied to the specific case of concrete particular objects, with respect to these competing accounts. The literature is divided over whether concrete particular objects are or are not further analyzable into constituents which do not themselves belong to the ontological category of concrete particular objects and in terms of which the character of these latter entities is to be explained. Constituent ontologies (e.g., bundle theories or substratum theories) attribute such a constituent structure to concrete ­particular objects and appeal to this structure in explaining their character, while nonconstituent ontologies (e.g., Platonism or austere nominalism) deny that concrete particular objects have constituents which are not themselves concrete particular objects and therefore do not attempt to explain their character by appeal to any such constituent structure. I briefly survey these non-hylomorphic approaches to the metaphysics of concrete particular objects and discuss prominent objections which have been raised against these accounts. These considerations in turn give rise to a set of desiderata and decision points which guide us in our subsequent development of a hylomorphic analysis of concrete particular objects in the chapters to follow. In Chapter  2 (“Matter”), I examine the question of how hylomorphists should ­conceive of the matter composing concrete particular objects. I consider three conceptions of matter: the traditional Thomistic doctrine of prime matter; the matter-as-stuff hypothesis; and the hylomorphic conception of matter, according to which the matter of a concrete particular object is nothing other than its material parts and these

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

4  introduction are  themselves conceived of as matter–form compounds. I argue that the third ­conception is preferable to the previous two. In Chapter 3 (“Form”), I turn to the question of how hylomorphists should conceive of the form of concrete particular objects. I argue that hylomorphists should endorse the individual forms hypothesis, and reject the universal forms hypothesis, on grounds primarily having to do with the cross-world identification of concrete particular objects. Other issues, e.g., the causal roles ascribed to form or the relation between form and essence, perhaps surprisingly, turn out to be neutral between the individual forms hypothesis and the universal forms hypothesis. When the conclusions of this chapter are combined with those of Chapter 4, we arrive at my preferred conception of forms as “robust” particulars which do not simultaneously bear the same relation both to the matter–form compound (essentially) and to the matter composing it (accidentally). This chapter incorporates material from Koslicki (2012b, 2014, 2018a). In Chapter 4 (“Hylomorphic Relations”), I take up the question of how hylomorphists should conceive of the relations which hold between a matter–form compound, its matter, and its form. In this chapter, I outline my response to the challenge issued to hylomorphists by the Grounding Problem, which poses the question of what (if anything) explains the apparent differences between numerically distinct spatiotemporally coincident objects. I argue that, in response to the Grounding Problem, hylomorphists should opt for a “robust” construal of form according to which something in the nature of forms or the compound–form relation blocks forms from simultaneously belonging, in the same way, to a matter–form compound (essentially) and to the matter composing it (accidentally). Armed with a robust conception of forms, hylomorphists can solve the Grounding Problem by invoking independently motivated differences between numerically distinct spatiotemporally coincident objects. According to my own preferred version of robust hylomorphic pluralism, the differences between numerically distinct spatiotemporally coincident objects are to be explained by appeal to a non-modal conception of essence and a mereological approach, developed in Koslicki (2008a), to the relation between a matter–form compound, its matter, and its form. This chapter incorporates material from Koslicki (2012b, 2018a). With Chapter 5 (“Ontological Dependence”), I begin Part II (“Substance”) whose main focus is on the question of whether concrete particular objects deserve to be assigned the ontologically privileged status of substancehood within a hylomorphic ontology and, if so, according to what notion of “ontological privilege.” As noted earlier, this assignment becomes potentially problematic once concrete particular objects are analyzed as metaphysically complex due to their hylomorphic structure. It is common to conceive of the substances as ontologically independent, according to some preferred sense of “independence.” But what is this preferred sense of “ontological independence” and do matter–form compounds qualify as substances when we apply this notion of ontological independence to them? This chapter discusses various relations which have been defined in the literature under the heading of “ontological dependence.”

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

introduction  5 I examine first existential construals of ontological dependence and turn next to ­construals of ontological dependence which are formulated in terms of a non-modal conception of essence. I argue in this chapter and Chapter 6 that even the most promising ones among these candidate relations are nevertheless open to objections when evaluated against various plausible measures of success. Chapter 5 incorporates material from Koslicki (2012a, 2013a). Chapter  6 (“Independence Criteria of Substancehood”) examines some initially attractive attempts at formulating an independence criterion of substancehood in terms of a particular essentialist construal of ontological dependence. I argue that, in order for a criterion of substancehood to yield the desired results when applied to hylomorphic compounds, a unity criterion of substancehood for composite entities is more suitable to the task at hand than an independence criterion, despite a general preference among Aristotelians for the latter. This chapter incorporates material from Koslicki (2013b, 2015b). Given the results of Chapters 5 and 6, we stand in need of a serviceable account of unity which can capture the idea that matter–form compounds are more unified than other types of composite entities (e.g., heaps, collections, or mereological sums). In Chapter 7 (“Unity”), I develop a conception of unity according to which a structured whole derives its unity from the way in which its parts interact with other parts to allow both the whole and its parts to manifest those of their capacities which require “team work” among the parts. With this conception of unity in place, interesting differences emerge between paradigmatic matter–form compounds belonging to natural (e.g., physical, chemical, or biological) kinds and composite entities belonging to social kinds, in particular artifacts. In the latter case, we find that the interactional dependencies that connect the components of a system can be traced to mind-dependent factors which are extrinsic or external to the system in question, viz., the mental states of intentional agents who invent, design, produce, or use an artifact. We continue our examination of the special features of artifacts in Chapter  8 (“Artifacts”). Based on my evaluation of existing essentialist theories of artifacts in Chapter 8, I reach a skeptical conclusion as to whether such approaches can be made to work. As it turns out, however, existing anti-essentialist frameworks are also not as accommodating toward the special case of artifacts as we might have expected them to be. For the time being, the special case of artifacts thus eludes an entirely satisfactory treatment and we must await the further development and refinement of existing essentialist and anti-essentialist frameworks before the status of artifacts within a hylomorphic ontology can be fully resolved. In a sense, the present study can be regarded as a kind of sequel to my earlier project in Koslicki (2008a). The basic approach that is advanced in both contexts has stayed the same: that concrete particular objects ought to be viewed as having a hylomorphic structure, i.e., as being in some sense compounds of matter and form. But in the earlier treatment my focus was primarily mereological: I was concerned most of all with the

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

6  introduction question of what the most plausible theory of parts and wholes is for those material (i.e., concrete particular) objects to which we are committed as part of our scientifically informed common-sense ontology. In particular, my aim there was to argue against the exceedingly deflationary conception of what it means to be an object, advocated, for example, in Lewis (1986), Sider (2001), and Thomson (1983), which follows from a commitment to the still popular theory of standard or classical extensional mereology (see, for example, Leonard and Goodman (1940); Lewis (1991); Simons (1987)). In its place, I recommended that we adopt a more full-blooded neo-Aristotelian account of parthood and composition, according to which objects are structured wholes: on this conception, it is integral to the existence and identity of a whole of a certain kind that its material parts satisfy certain type and configuration constraints. To illustrate, an H2O molecule, for example, must be composed of two hydrogen atoms and one oxygen atom and these material parts must be arranged in the manner of chemical bonding which requires the atoms in question to share electrons. In this present study, I take the earlier structure-based theory of parthood and ­composition for granted and build on it in order to develop other aspects of my defense of a hylomorphic conception of concrete particular objects which were not already addressed or explicitly settled in Koslicki (2008a). This opportunity also enables me to comment on more recent developments in the flourishing literature on hylomorphism, weigh in on concerns raised by critics of my earlier work, and take a stand on questions on which I remained neutral or on which I have changed my mind in the intervening years.1 I highlight a few of the main respects in which the current study goes beyond what was already accomplished in Koslicki (2008a). In Part I, in contrast to the primarily mereological focus of the previous work, I now situate my particular hylomorphic theory more broadly with respect to other hylomorphic and non-hylomorphic approaches to the metaphysics of concrete particular objects and develop a series of decision points and desiderata with respect to which these competing approaches can be developed and evaluated (Chapter 1). In Koslicki (2008a), I already advocated a hylomorphic conception of matter, according to which the material parts of matter–form compounds are themselves to be viewed as matter– form compounds, as long as these material parts are also structured wholes. In Chapter 2, I offer new support for this approach by evaluating possibilities that were not explicitly considered in the earlier study and by engaging with a wider range of competing views, in particular, the Thomistic doctrine of prime matter as interpreted by Oderberg (2007) 1  For reviews or critical discussions of Koslicki (2008a), see, for example, Bader (2010); Bennett (2011c); Britton (2012); Donnelly (2011); Evnine (2011); Hawley (2010); Hovda (2009); Inman (2011); McFarland (2011); Merricks (2009); Oderberg (2014); Sidelle (2010); Tahko (2011); and Tomaszewski (2016). For recent contributions to the ever-growing literature on hylomorphism, see, for example, Brower (2010, 2012, 2014); Evnine (2016a); Feser (2013); Groff and Greco (2013); Jaworski (2014, 2016a, 2016b); Kanzian (2009, 2013, 2015, 2016); Koons (2014); Marmodoro (2013); Novotný and Novák (2014); Oderberg (2007); Rea (2011); Sattig (2015); Shields (2010a, 2010b, 2010c); Sirkel and Tahko (2014); Tahko (2012); and Toner (2013).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

introduction  7 and Brower (2014) as well as the conception of stuff as developed by Markosian (1998, 2004a, 2004b, 2015). In Chapter 3, I turn to a particularly thorny set of issues—the ontological category to which forms should be assigned—on which I remained largely neutral in the earlier study. Based on work developed during the intervening period (especially Koslicki (2012b, 2018a, 2018b)), I have now come to believe that forms should be taken to be robust particulars of a type that bars them from standing in the same relation both to the matter (accidentally) and to the matter–form compound (essentially). My endorsement of the individual forms hypothesis is motivated primarily by my desire to settle certain tricky questions concerning the cross-world identity of concrete particular objects. The robust construal of forms is motivated, in Chapter 4, by my desire to formulate a hylomorphic solution to the Grounding Problem. (The Grounding Problem asks how an object and its matter can apparently differ, among other things, with respect to their modal profile.) In contrast to what I suggested in Koslicki (2008a), I now no longer hold that positing a difference in formal parts between numerically distinct spatiotemporally coincident objects by itself suffices to generate a hylomorphic solution to the Grounding Problem. Instead, I have since come to believe that a certain non-modal or non-reductive conception of essence, of the type encountered, for example, in the works of Aristotle and Kit Fine, is also needed to formulate a credible hylomorphic or non-hylomorphic response to the Grounding Problem. In Part II, I discuss the alleged substance status of concrete particular objects, an issue which was not taken up at all in Koslicki (2008a). My views concerning ontological dependence as well as the interconnections between ontological independence, ­substancehood, and fundamentality, discussed in Chapters 5 and 6, have all been developed since the publication of the earlier study (see especially Koslicki (2012a, 2013a, 2013b, 2015a, 2015b, 2016a, 2016b)). My take on unity, as described in Chapter 7, is also new and no version of it has appeared in print. In Koslicki (2008a), I suggest that the problem of unity (i.e., the problem of how a whole which has many parts can nevertheless be unified) can be solved fairly straightforwardly simply by formulating a restricted composition principle of the sort I put forward there. Since then, I have come to believe that a proper account of unity requires not only the static notion of structure, but also a dynamic apparatus suitable to capturing the complex network of interactional dependencies which exist between the proper parts of an integrated whole and the whole itself. Finally, although I have by now been guilty of using illustrations from the realm of artifacts for decades, the discussion of artifacts in Chapter 8 constitutes my first (though still very tentative) attempt at sorting out the ontological status of artifacts from the point of a view of a realist hylomorphic ontology. I take this occasion primarily to lay out some directions for essentialist and anti-essentialist treatments of artifacts which I intend to pursue further in future research. As I hope to have indicated in this brief sketch, despite all that has already been accomplished, much work still remains to be done in filling out the details of a hylomorphic approach to the metaphysics of concrete particular objects. In what follows,

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

8  introduction I intend to contribute to this larger undertaking by offering my own thoughts on some of the challenges faced by hylomorphists and thereby join others in carrying forward the momentum currently experienced by Aristotelian approaches to metaphysics and other parts of philosophy.2

2   “Aristotelian” or “neo-Aristotelian” metaphysics, fittingly, is spoken of in many ways by those who consider themselves to be either part of, or opposed to, this “movement” in philosophy. Even though hylomorphism constitutes only one popular strand that has been taken up by researchers working in the Aristotelian tradition, we have already seen in this Introduction that even in this single instance the doctrine in question can be, and has been, interpreted in a multiplicity of different ways as it is applied across a vast range of phenomena. Some modern-day Aristotelians are deeply concerned with historical faithfulness; others are happy to borrow from Aristotle wherever it suits their purposes, without attempting to put forward an interpretation of Aristotle’s texts. My own views, which are very much inspired by Aristotle, are in some cases intended to be faithful to the text; in other cases, however, I deliberately diverge from what I consider to be Aristotle’s own views in the interest of developing a hylomorphic theory of concrete particular objects that is acceptable to contemporary metaphysicians. In all of this, I do not mean to put forward anything like a “manifesto” to which other researchers who take themselves to be working in the Aristotelian tradition must adhere in order to be considered part of the “movement.”

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

PA RT I

A Hylomorphic Analysis of Concrete Particular Objects

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

1 Concrete Particular Objects 1.1  Introductory Remarks Extant approaches to concrete particular objects in the literature are divided over whether these entities are or are not analyzable into further constituents which belong to an ontological category distinct from that of concrete particular objects and in terms of which the character of concrete particular objects is to be explained. (I use the phrase “character” here in a deliberately open-ended manner, to leave room for a variety of explanatory tasks which arise in connection with the metaphysics of concrete particular objects, such as those identified in the remainder of the chapter.) A constituent ontology (e.g., bundle theory or substratum theory), on the one hand, ascribes such a constituent structure to concrete particular objects; a non-constituent ontology (e.g., Platonism or austere nominalism), on the other hand, denies that concrete particular objects have constituents which are not themselves concrete particular objects and hence does not attempt to explain the character of concrete particular objects by appeal to such a constituent structure.1 The doctrine of hylomorphism, in its application to the specific case of concrete particular objects, can be interpreted as yielding a certain type of constituent ontology. On this reading, a concrete particular object is taken to be analyzable into further constituents, viz., matter (hulē) and form (morphē or eidos), at least one of which belongs to a distinct ontological category and in terms of which the object’s character is to be explained. Although I myself find attractive this construal of hylomorphism as giving rise to a certain type of constituent ontology, it will emerge that other hylomorphists prefer a different reading.

1.2  Constituent vs. Non-Constituent Ontologies It is compatible with a non-constituent ontology that it nevertheless ascribes to concrete particular objects a mereological constituent structure. To illustrate, consider the 1   I opt for the label “non-constituent” instead of the commonly used label “relational” to characterize non-constituent ontologies, since not all non-constituent ontologies (e.g., austere nominalism) are naturally characterized as relational. The contrast between “constituent” and “relational” ontologies comes from Wolterstorff (1970); for further discussion, see also Loux (2005,  2006b, 2014); Lowe (2012b); and van Inwagen (2011). David Armstrong uses a similar distinction between “blob” and “layer cake ontologies” (Armstrong (1989), pp. 76–7). To attempt to classify ontological theories by means of taxonomies such as those just cited is by no means straightforward. I will not pursue these difficult issues further here, since I do not intend to put much philosophical weight on these divisions. I merely use them as a heuristic device in Chapter 1 to introduce alternative approaches to the metaphysics of concrete particular objects with which hylomorphists must compete.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

12  Concrete Particular Objects ontology Aristotle defends in the Categories. Relative to this ontology, concrete particular objects (e.g., individual living organisms), which are classified there as primary substances, may have proper parts which are themselves classified as primary substances (e.g., Socrates and Socrates’ arm).2 The mereological complexity Aristotle attributes to these entities does not yield an analysis of Socrates into constituents which belong to a distinct ontological category and in terms of which Socrates’ character is to be explained, since both Socrates and his proper parts are assigned the same status within the ontology of the Categories, namely, that of primary substances. The Categories is generally taken to be one of the earliest, if not the earliest, of Aristotle’s written works. In contrast, when we turn to such texts as the Physics, De Anima, and the Metaphysics, we notice that Aristotle’s views concerning the metaphysical complexity of such entities as individual living organisms have undergone a definitive shift. For a variety of reasons, Aristotle comes to believe that those concrete particular objects which were previously his ontological front runners in the Categories require further analysis into additional constituents, viz., their matter and their form, in terms of which the character of concrete particular objects is to be explained. Aristotle’s hylomorphic analysis of concrete particular objects as compounds of matter and form will be considered in more detail in subsequent chapters. The important point for present purposes is only that, while the Categories ontology does not ascribe to such entities as individual living organisms the type of structure characteristic of a constituent ontology, it does allow that these entities have proper parts which are assigned to the same ontological category as the wholes to which they belong. Thus, relative to the distinction drawn above, Aristotle’s approach in the Categories can be classified as a non-constituent ontology, while the hylomorphic analysis of concrete particular objects he defends in later texts can be classified as a constituent ontology, assuming that either matter or form (or both) are treated as constituents belonging to a distinct ontological category and as explanatory of the matter–form compound’s character.

1.3  Non-Constituent Ontologies: Platonism and Austere Nominalism Given our terminology, then, a non-constituent ontology is one which treats concrete particular objects, so to speak, as “blobs,” i.e., as metaphysically unstructured, while allowing that these blob-like entities may nevertheless have proper parts that are 2   Aristotle is careful to distinguish his “being in a subject” relation from parthood (see Cat. 2, 1a24–5) and is thereby able to maintain in the Categories that the parts of substances are themselves substances:

We need not be disturbed by any fear that we may be forced to say that the parts of a substance, being in a subject (the whole substance), are not substances. For when we spoke of things in a subject we did not mean things belonging in something as parts. (Cat. 5, 3a29–32) This and all further passages from Aristotle’s Categories come from the translation by J. L. Ackrill (see Barnes (1984)).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

constituent ontologies  13 themselves concrete particular objects. Non-constituent ontologies come in different varieties, e.g., Platonism (also sometimes classified as a type of relational ontology) and austere nominalism.3 Both of these theories are primarily put forward as responses to the so-called Problem of Universals.4 The Problem of Universals, which has occupied philosophers at least since the time of Plato, arises when we ask what (if anything) many concrete particular objects which resemble each other in certain objective respects have in common. For example, since fire trucks, tomatoes, and sunsets are all objectively similar to one another by being red, one might think that there is some entity, viz., redness, to which these concrete particular objects are all related in some way and by appeal to which their red character is to be explained. According to the Platonist, the property or characteristic of redness that is apparently shared by all and only those concrete particular objects that are red should be understood as a mind-independent universal that is “transcendent,” i.e., immaterial and abstract and therefore not located in space and time. (The Platonist conception of universals contrasts with another form of realism known as Aristotelianism, according to which universals are “immanent,” i.e., spatiotemporally located and in some sense present in the objects that exemplify them.) For the Platonist, objectively similar concrete particular objects all stand in some relation (e.g., participation, exemplification, or instantiation) to the corresponding universal.5 In contrast, the austere nominalist (e.g., Quine (1960)) maintains that there is no need to posit properties as separate entities in order to capture the ways in which concrete particular objects objectively resemble one another. Rather, for the austere nominalist, everything that exists is itself a concrete particular object and the objective similarities which obtain between concrete particular objects are not taken to be subject to further explanation at all.

1.4  Constituent Ontologies: Bundle Theories and Substratum Theories A constituent ontology allows that concrete particular objects may be complex in two ways: firstly, by possibly having proper parts that are assigned to the same ontological category as the concrete particular object whose parts they are; and, secondly, by consisting of constituents which belong to a distinct ontological category and in terms of which the character of a concrete particular object is to be explained. Constituent ontologies traditionally come in two main flavors: bundle theories 3   Loux (2006a) uses the term “austere nominalism” to describe a type of nominalist ontology which posits only concrete particular objects and which takes facts concerning objective resemblance between concrete particular objects to be fundamental and not open to further analysis or explanation (see Loux (2006a), pp. 52–3). 4   See, for example, Armstrong (1989) for a useful introductory survey of the Problem of Universals and prominent responses to it. 5   The original Platonist was of course Plato himself: see, for example, the theory of forms as it is defended in Plato’s so-called middle dialogues (e.g., Phaedo, Symposium, Republic). For a more recent defense of Platonism, see, for example, van Inwagen (2004).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

14  Concrete Particular Objects and  substratum theories.6 Both of these varieties of constituent ontologies can themselves be further subdivided into more specific types. Bundle theorists hold that concrete particular objects are (numerically identical to) bundles of properties. But bundle theorists differ over how they approach properties, i.e., over which response to the Problem of Universals they favor. Universalist bundle theorists take the properties or characteristics that are shared by objectively similar concrete particular objects to be universals.7 In contrast, trope-theoretic bundle theorists hold that the properties or characteristics that are shared by objectively similar concrete particular objects are themselves particulars, viz., so-called tropes, moments, or modes.8 Tropes are construed by their proponents as particularized properties or individual qualities, e.g., the particular redness that inheres in a rose.9 Both types of bundle theorists view concrete particular objects as in some sense composed of, or constituted by, the properties that enter into a particular bundle; only they differ over whether the very same properties can be multiply located in distinct bundles at a single time. For substratum theories, concrete particular objects are not exhausted by the properties they exemplify; rather, in addition, each concrete particular object is also associated with a substratum, i.e., an entity of a certain type in which the properties exemplified by the concrete particular object in question inhere. The substratum underlying the properties possessed by a concrete particular object is often taken to be a type of bare particular, an entity whose identity is not in any way tied up with the properties exemplified by the concrete particular object with which it is associated.10 But not all bare particularists conceive of the substratum and its relation to the concrete particular object with which it is associated in the same way. According to one version of bare particularism (thin particularism), concrete particular objects are identified with the bare particulars with which they are associated.11 A different version of 6   In what follows, I use the term “substratum theory” to refer to a type of constituent ontology which holds that each concrete particular object, in addition to the properties it exemplifies, is also associated with a further type of constituent, viz., a substratum of some type. But the phrase, “substratum,” can also be used in a more neutral way (see, e.g., Brower (2014)), according to which a substratum is any type of being (constituent or not) which bears the properties exemplified by a concrete particular object. On this construal, even non-constituent ontologies could in principle be classified as substratum theories: a Platonist, for example, who views concrete particular objects as metaphysically unstructured “blobs,” might nevertheless take them to be substrata for their properties; for such a Platonist, then, a substratum would be an entity (viz., the concrete particular object itself) which stands in certain relations (e.g., exemplification, participation, or instantiation) to transcendent universals. 7   Bertrand Russell, at one point in his career, subscribed to universalist bundle theory (Russell (1940)); for a more recent defense, see Paul (2002). 8   It is best not to refer to tropes as “accidents,” if we want to leave open the possibility that some tropes are essential to their bearers. 9   David Hume is generally read as a proponent of trope-theoretic bundle theory (see Hume (2007)). More recent defenders include Keith Campbell and D. C. Williams (see, for example, Campbell (1990) and Williams (1953a, 1953b)). 10   John Locke famously posited a something “I know not what” which underlies the properties associated with a concrete particular object (see Book II of Locke (1975)). 11   A thin particularist may nevertheless be regarded as a constituent ontologist, even if such a theorist takes the thin particular associated with a concrete particular object itself to be simple and metaphysically unstructured. For presumably the thin particularist will nevertheless accept the existence of thick

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

difficulties for non-hylomorphic competitors  15 bare particularism (thick particularism or factualism) identifies concrete particular objects with complexes of some sort (e.g., facts or states of affairs) which include among their constituents entities that are not themselves concrete particular objects. For example, David Armstrong’s specific version of this type of approach takes those concrete particular objects which we encounter in ordinary experience to be states of affairs and makes room for thin particulars only insofar as these can be “abstracted” from the thick particulars with which they are associated. Nevertheless, the concrete particular objects we encounter in ordinary experience (i.e., thick particulars or states of affairs), in Armstrong’s view, are metaphysically structured in that they consist of non-mereological constituents, viz., thin particulars and properties, the latter of which Armstrong takes to be Aristotelian universals.12, 13

1.5  Difficulties for Non-Hylomorphic Competitors In this section, I briefly survey some well-known difficulties which arise for the nonhylomorphic approaches to the metaphysics of concrete particular objects we have reviewed so far.14 By considering prominent objections which have been raised against these competing accounts, we also begin to develop a sense of what explanatory goals a hylomorphic analysis of concrete particular objects ideally should meet and what pitfalls it should, if at all possible, avoid. Difficulties for Non-Constituent Ontologies: Property Possession, Objective Similarity, and Change. Non-constituent ontologists deny that concrete particular objects are further analyzable into constituents belonging to distinct ontological categories in terms of which the character of concrete particular objects is to be explained. Thus, non-constituent ontologists must meet whatever explanatory challenges arise in connection with these entities without taking recourse to any internal structural features particulars and these entities do contain thin particulars as constituents. See, for example, Sider (2006) for further discussion. 12  For a version of factualism, see, for example, the early Ludwig Wittgenstein’s Tractatus LogicoPhilosophicus; a more recent development of a factualist approach can be found, for example, in Turner (2016). 13   Brower (2014) classifies the version of hylomorphism he ascribes to Aquinas as a type of substratum theory, viz., a version of thick particularism. My decision to treat hylomorphism separately is motivated by a mostly practical concern. I want to preempt the following possible misconception which holds that, if hylomorphism is classified as a type of substratum theory, then it must be made to fit into one of the other two varieties of substratum theory just cited, both of which are versions of bare particularism (viz., thin and thick particularism). The interpretation of hylomorphism I go on to defend differs in important respects from both of these varieties of substratum theory. However, assuming that we can ward off the mistaken assumption that hylomorphists must take on board a commitment to bare particulars, I have no objection to going along with Brower’s classification of hylomorphism as a type of substratum theory. Whether hylomorphism is best interpreted as a type of thick particularism is a further question on which Brower and I reach different conclusions. 14   What follows is familiar ground and my exposition will therefore be relatively condensed. For more detailed discussion, see, for example, Brower (2014), especially Chapters  2,  6, and  7; Laurence and Macdonald (1998); Loux (2002); Macdonald (2005); and Rea (2014).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

16  Concrete Particular Objects that can be ascribed to concrete particular objects themselves. Consider, for example, the question of how (if at all) a concrete particular object (e.g., a tomato) is related to the properties it possesses (e.g., the tomato’s redness). A non-constituent ontologist can either reject or accept the invitation to explain what it takes for a concrete particular object to have the features that it does (e.g., a tomato’s being red) in terms which posit a relation between the concrete particular object (viz., the tomato) and its properties (viz., redness). If she takes the first strategy, she has further work to do to make her claim palatable that in fact no such explanation is needed or called for. If she takes the second strategy, then her account of property possession must be formulated in such a way that it does not locate the properties a concrete particular object possesses “inside” these objects themselves in order to explain why they have the features they do. The non-constituent ontologist faces a similar dialectical situation when it comes to the question of how to explain the objective similarities which hold between pluralities of concrete particular objects (e.g., a sunset, a tomato, and a fire truck).15 Perhaps the most important consideration which drove Aristotle to modify the non-constituent ontology he offered in the Categories was his desire to formulate an adequate response to the Parmenidean problem concerning change. Parmenides of Elea argued that, if change ever were to occur, it would have to involve either a transition from being to non-being or from non-being to being; and since either of these scenarios, in his view, would give rise to logical contradictions, Parmenides concluded that change is therefore impossible. Aristotle agreed with the Eleatics that something cannot come to be from nothing, or nothing from something, without qualification. However, according to Aristotle’s own analysis of change, as put forward in Physics I, something always persists through any change that occurs (viz., the matter or subject underlying the change) and something else does not persist through the change (viz., the form or privation). For example, when some wood goes from not being bed-shaped to being bed-shaped, the wood remains throughout the change. But in order for this process to constitute a change, the starting-point of the transition in question must also in some way differ from its outcome: in the present case, the wood’s not being bedshaped does not persist through the change. In this way, Aristotle concludes that change is possible after all, since it only ever involves something coming to be from nothing (or vice versa) in a qualified way. At the same time, Aristotle’s analysis of change led him to attribute a hylomorphic structure to concrete particular objects which he did not previously recognize in his Categories ontology. Difficulties for Universalist Bundle Theorists: Identity and Indiscernibility. A common objection that is raised against the universalist version of the bundle theory is that it seems to be committed to the truth of the very controversial Principle of the Identity of 15   E. J. Lowe, for example, a contemporary critic of Platonism, is puzzled by how positing an external relationship of some kind between a “blob-like” concrete particular object and a transcendent universal could contribute to an explanation of why concrete particular objects have the features they do (see Lowe (2012b), pp. 234–5). To this kind of criticism, Peter van Inwagen, a modern-day defender of Platonism, replies that, while indeed his account does not yield an explanation of exemplification or objective resemblance, none should be expected either (see van Inwagen (2011), p. 398).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

difficulties for non-hylomorphic competitors  17 Indiscernibles (PII). This principle states that necessarily entities which are qualitatively indiscernible are numerically identical.16 Max Black famously argued that (PII) is false by asking us to consider a universe which allegedly contains nothing but two numerically distinct but qualitatively indiscernible spheres: Isn’t it logically possible that the universe should have contained nothing but two exactly similar spheres? We might suppose that each was made of chemically pure iron, had a diameter of one mile, that they had the same temperature, colour, and so on, and that nothing else existed. Then every quality and relational characteristic of the one would also be a property of the other. Now if what I am describing is logically possible, it is not impossible for two things to have all their properties in common. This seems to me to refute the Principle.  (Black 1952, p. 156)

It seems that universalist bundle theorists are forced to deny that Black’s scenario describes a genuine metaphysical possibility. For, by hypothesis, the spheres in Black’s scenario are qualitatively indiscernible and, hence, by the universalist bundle theorist’s lights, they instantiate the same universals. Since, for the universalist bundle theorist, concrete particular objects are just bundles of those universals that are instantiated by the objects in question, the spheres in Black’s scenario will be associated with the very same bundle of universals and the scenario in question therefore really only contains one concrete particular object, rather than two. The universalist bundle theorist thus appears to be committed to the claim that qualitatively indiscernible objects are by necessity numerically identical, and hence to the truth of PII. Those who find it plausible to think that numerically distinct but qualitatively indiscernible concrete particular objects are metaphysically possible must therefore either abandon the universalist bundle theory or find a way to respond to the objection just raised.17 Difficulties for Trope-Theoric Bundle Theorists: Compresence and Unity. Tropetheoretic versions of the bundle theory are not susceptible to the above worry concerning the Identity of Indiscernibles: since these theorists view properties as particulars, they are able to associate the spheres described in Black’s scenario with numerically distinct bundles of intrinsically similar tropes. Like their universalist counterparts, however, trope-theoretic variations of the bundle theory do face the following over-generation worry. Suppose we understand the operation of bundling as one that is akin to (or simply is) mereological composition. Assuming further that the operation of bundling is unrestricted, every plurality of tropes will compose a trope bundle. If we apply this unrestricted bundling operation to Black’s scenario, for example, the result will 16   It is tricky to say precisely how the distinction between qualitative and non-qualitative properties is to be drawn. For present purposes, we may rely on the rough characterization of this distinction given in Adams (1979): “ . . . a property is purely qualitative—a suchness—if and only if it could be expressed, in a language sufficiently rich, without the aid of such referential devices as proper names, proper adjectives, and verbs (such as ‘Leibnizian’ and ‘pegasizes’), indexical expressions, and referential uses of definite descriptions” (1979, p. 7). 17   See, for example, O’Leary Hawthorne, and Cover (1998) for a response to Black’s scenario on behalf of the universalist bundle theorist. See also Hawley (2006); Paul (2012,  2013, 2016,  2017); Saunders (2003, 2006a, 2006b).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

18  Concrete Particular Objects be that any combination of tropes that are associated with the two spheres itself gives rise to a separate trope bundle. But how, we may ask, does the trope-theoretic bundle theorist then account for the judgment that apparently the scenario described by Black contains exactly two concrete particular objects? Are the trope bundles with which the spheres are identified in some way “special,” e.g., by being more unified than the other combinations of tropes that are present in the scenario in question? In that case, the trope-theoretic bundle theorist might try to avoid the present over-generation worry by claiming that only the more unified trope bundles are to be identified with concrete particular objects, while the less unified ones are “mere” bundles. At this point, trope-theoretic bundle theorists typically appeal to a relation they call compresence or collocation, in order to differentiate between those combinations of tropes that give rise to concrete particular objects and those that do not. According to this line of thought, what distinguishes the trope bundles associated with each of the spheres, for example, from all the other possible combinations of tropes that are present in Black’s scenario is that the former are all thought to be compresent or collocated within a single region of space-time, while the latter are scattered across different and possibly disconnected regions of space-time. Unfortunately, however, the tropetheoretic bundle theorist’s reliance on the notion of compresence or collocation gives rise to well-known difficulties; and, as a result, the overall plausibility of this approach to the metaphysics of concrete particular objects is thereby hampered as well.18 Difficulties for Substratum Theorists: Bare Particulars, Essence, and Accidents. Perhaps the most puzzling feature of the bare particularist version of the substratum theory is just the central thesis itself to which it is committed, viz., the idea that each concrete particular object is associated with a bare particular which in some way “underlies” the properties exemplified by the object in question. Proponents of this approach to the metaphysics of concrete particular objects are quick to point out that we need not think of bare particulars as completely devoid of properties, since it is part of the very job description of these entities that they serve as the bearers of properties. In this way, some of the most obvious objections against the theory, originating from the thought that nothing could be both a particular entity and yet not be qualified in any way, can be avoided.19 Let’s assume, then, that we need not think of bare particulars as completely unqualified. Presumably, bare particulars are nevertheless characterized as “bare” because whatever properties inhere in them do so only accidentally: insofar as bare particulars can be described as having an essence or intrinsic nature of their own at all, it is only in that they are said to be (i) particular, and (ii) capable of acting as the substrata for whatever properties come their way. In that case, however, it becomes difficult to see how, according to thin particularism, some of the properties that apparently belong to a concrete particular object could nevertheless be essential to the object in question, even though they inhere in the bare particular with which the object is identified only 18   See, for example, Simons (1994), pp. 558–65, for a discussion of some of the problems surrounding the trope-theoretic bundle theorist’s use of the notion of compresence or collocation. 19   For further critical discussion of bare particularism, see, for example, Bailey (2012).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

desiderata and decision points for hylomorphists  19 accidentally. Thus, thin particularism seems to lead to the unwelcome consequence that there is only one way in which properties can be associated with concrete particular objects, viz., by accidentally inhering in the bare particular associated with the object in question. Thick particularism, by contrast, has the opposite tendency: it seems to categorize all of the properties that are associated with concrete particular objects, when these objects are conceived of as thick particulars (i.e., facts or states of affairs), as essentially or necessarily belonging to the objects in question. To see this, consider the role  Armstrong assigns to states of affairs as the truth-makers for propositions. In Armstrong’s view, in order for a proposition to be true, a certain state of affairs, which acts as the truth-maker for the proposition in question, must obtain. In addition, Armstrong accepts a position known as “truth-maker necessitarianism,” according to which the relation between true propositions and their truth-makers is governed by the following principle: necessarily, if the state of affairs (e.g., the tomato’s being red) which acts as the truth-maker for a certain proposition (viz., the proposition expressed by the sentence, “The tomato is red”) obtains, then the proposition in question is true. Given truth-maker necessitarianism, it follows that states of affairs cannot undergo any changes with respect to their non-mereological constituents (viz., thin particulars and the universals exemplified by them). To illustrate, suppose for example, that the state of affairs consisting in a tomato’s being red could persist through a change which might be described by saying that the tomato goes from being red to being yellow. In that case, the existence of the state of affairs in question would no longer guarantee the truth of the proposition for which it is supposed to act as a truth-maker (viz., the proposition expressed by the sentence, “The tomato is red”), since the state of affairs in question now makes true a different proposition (viz., the proposition expressed by the sentence, “The tomato is yellow”), whose truth is incompatible with the truth of the proposition for which it previously acted as a truth-maker. Given Armstrong’s theoretical commitments, then, his thick particularism requires that it is necessary or essential to the identity of a state of affairs that it consists of the very non-mereological constituents of which it in fact consists.20

1.6  Desiderata and Decision Points for Hylomorphists In the foregoing section, we have considered prominent objections which have been raised in connection with some of the major non-hylomorphic approaches to the metaphysics of concrete particular objects. These challenges also give rise to a set of desiderata and decision points which can guide hylomorphists in their efforts to formulate the most attractive version of their own competing analysis of concrete particular objects. 20   See Armstrong (1997) and (2004). Other versions of factualism are susceptible to similar worries: see, for example, the arguments provided in Wolterstorff (1970) in connection with Gustav Bergmann’s constituent ontology.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

20  Concrete Particular Objects (1)  Property Possession and the Problem of Universals. As noted earlier, some of the approaches we have encountered (e.g., Platonism, Aristotelianism, austere nominalism, and trope theory) are primarily driven by their commitment to a particular position concerning the metaphysics of properties. These accounts are thus most directly focused on questions concerning property possession and objective resemblance. By contrast, other approaches to the metaphysics of concrete particular objects are in and of themselves compatible with different solutions to the Problem of Universals (e.g., substratum theories or bundle theories). Hylomorphism falls into this latter category: the doctrine’s central thesis, viz., that concrete particular objects are compounds of matter and form, does not automatically legislate a particular solution to the Problem of Universals, although some conceptions of properties are more naturally combined with hylomorphism than others. Proponents of this doctrine can be expected to be more sympathetic to Aristotelianism about universals or trope theory than to Platonism or austere nominalism, given the hylomorphist’s inclination to locate their theory’s central explanatory principles, matter, and form, within the concrete particular objects themselves whose character they are supposed to explain. At the same time, while Aristotelianism about universals or trope theory are both live options for hylomorphists, they would not wish to combine either of these approaches with a bundle-theoretic conception of concrete particular objects, since they do not perceive bundles as sufficiently unified to be able to stand in for matter–form compounds. (We will take up the topic of unity again under point (7)). Decision Point for Hylomorphists. Is the relation between a concrete particular object and its form as well as the relation between matter and form the same as, or different from, the relation between a property-bearer and its properties (property possession)? In addition to having to weigh their preferences with respect to the metaphysics of properties, hylomorphists face a further issue in this area. Since they regard concrete particular objects not only as compounds of matter and form but also as the bearers of properties, one wonders whether the relation of property possession (regardless of whether properties are viewed as immanent universals or tropes) is the same as, or different from, the various hylomorphic relations which hold between a concrete particular object, its matter, and its form. Some versions of hylomorphism, viz., those that conceive of forms as property-like entities, lend themselves to the idea that the relations in question can be thought of along the lines of property possession. Other versions of the doctrine, however, reject this idea and view the relation between a hylomorphic compound and its form, as well as the relation between the matter and the form when together they make up a concrete particular object, as different from property possession. We will have further occasion to examine these options as well as the different versions of hylomorphism to which they give rise.21 21   One issue which has traditionally divided those who endorse different solutions to the Problem of Universals is that of ontological parsimony, i.e., the desire to keep the number of distinct ontological categories to which one’s theory is committed as low as reasonably possible. Although the issue of ontological parsimony is an important dividing line by means of which different ontological theories can be

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

desiderata and decision points for hylomorphists  21 (2)  Causation, Explanation, and Change. The doctrine of hylomorphism is very much designed to respond to questions concerning causation, explanation, and change, among them the following: how do concrete particular objects come into and go out of existence? How do they persist through change? Why are concrete particular objects capable of persisting through some changes but not others? What causes concrete particular objects to be as they are, once they have come into existence? As noted earlier, Aristotle’s own hylomorphic analysis of concrete particular objects was first developed in the context of the analysis of change he proposes in Physics I, in response to the Parmenidean thesis that change is impossible. In Physics II, Aristotle embeds his matter–form distinction into a more general theory of causation and explanation which famously centers around his doctrine of the four causes.22 Aristotle argues there that two explanatory principles, matter and form, are sufficient to account for any kind of change involving natural things (i.e., things which have their own internal principles of change and stability), with matter playing the role of the material cause and form acting as the formal, final, and efficient cause of any such natural change.23 Decision Point for Hylomorphists. Can we deduce anything about the ontological category to which the matter and the form of a matter–form compound belong from the causal roles which are assigned to these two explanatory principles? One of the most hotly debated issues among ancient scholars is the question of whether Aristotelian forms should be interpreted as non-repeatable entities (i.e., as individuals or particulars) or as repeatable entities (i.e., as universals or general entities). According to the first interpretation, the form that is present in each matter–form compound is numerically distinct from the form that is present in every other matter–form compound, even when these matter–form compounds belong to a single species and their individual forms therefore closely resemble one another. In contrast, the second interpretation holds that the form present in one member of a species is not only similar to, but the very same form as, the form present in every other member of the same species. In connection with this debate, some commentators (e.g., Frede and Patzig (1988), Shields (2010c)) argue that only the forms-as-individuals reading can really make sense of the causal work that is supposed to be done by forms within Aristotle’s hylomorphic analysis of concrete particular objects. Others, however, distinguished, I will not include it here as a separate decision point or desideratum for hylomorphists. Rather, ontological parsimony, along with other “intellectual virtues” (e.g., simplicity, explanatory power, etc.), strikes me as more suitably classified as a sort of meta-criterion by means of which all ontological theories can be evaluated; and hence it is not specific to the debate concerning the metaphysics of concrete particular objects. 22   See, for example, Shields (2010a), Section 2, for further discussion of how Aristotle’s matter–form distinction is motivated by the framework he adopts to address questions concerning causation, explanation, and change. In the Thomistic case as well, Brower (2014) provides powerful arguments that the commitment to hylomorphism directly emerges from Aquinas’ analysis of change. 23   For Aristotle’s position that, in the case of natural things which are subject to change, the formal cause, the final cause, and the efficient cause coincide, with all three causal roles taken up by the form of a matter–form compound, see, for example, Phy. II.2, 194a27–30, Phy. II.7, 198a22–31, and Phy. II.8, 199a30–2. For further discussion concerning the causal priority of form in Aristotle, see Koslicki (2014).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

22  Concrete Particular Objects defend the opposing view according to which the causal roles attributed to forms do not require that these entities must be interpreted as unique to each matter–form compound (see, e.g., Loux (1991), especially Ch. 6). We will discover that hylomorphists face equally challenging questions when it comes to the ontological categorization of the matter composing a matter–form compound as well.24 (3)  Essence and Accident. As noted earlier in connection with substratum theories, non-hylomorphic approaches to the metaphysics of concrete particular objects which allow for only a single form of property possession face difficulties accommodating the distinction between what is part of the essence of a concrete particular object and what is merely accidental to it. By contrast, it is central to Aristotle’s own hylomorphic analysis of concrete particular objects, and to that of other versions of hylomorphism as well, to take very seriously the distinction between essence and accident, as it applies to these entities. But hylomorphists have options when it comes to the question of how best to approach the essence/accident distinction: while many in fact follow Aristotle’s lead and Kit Fine’s pioneering work on essence in the 1990s, and accept a non-modal conception of essence, the doctrine of hylomorphism as such is also in principle compatible with a modal conception of essence.25 According to a modal conception of essence, an essential truth is just a modal truth of a certain kind (viz., one that is both necessary and de re, i.e., about a certain object); and an essential property is just a feature an object has necessarily, if it is to exist. The essential truths, according to this approach, are thus just a subset of the necessary truths; and the essential properties of objects are just a special kind of necessary property. Quine, for example, has such a modal conception of essence in mind when he argues that the view he calls “Aristotelian essentialism” is incoherent, e.g., because it requires quantification into intensional contexts (Quine (1953)). But the view Quine calls “Aristotelian essentialism” is for a variety of reasons not one Aristotle himself would have found congenial, since he does not subscribe to a modal 24   A further decision point arises for hylomorphists in connection with the question of how concrete particular objects are apparently able to persist through change over time. According to the position known as perdurantism or four-dimensionalism, concrete particular objects exist at a particular time by having a temporal part that exists at the time in question. The opposing position, known as endurantism or three-dimensionalism, denies this claim and holds that concrete particular objects which persist over time are, in some sense, “wholly present” at each time at which they exist. I will not enter into this debate here; but see Koslicki (2003) and (2008a), especially Chapter 2, for further discussion. 25   See especially Fine (1994a, 1995a, 1995b, 1995c). Other contemporary neo-Aristotelians who have gravitated toward a non-modal conception of essence include the following: Gorman (2005); Koslicki (2012a, 2012b, 2013a, 2013b); Lowe (1994a, 1998, 2006, 2007, 2008, 2012a, 2013); Oderberg (2007, 2011); Tahko and Lowe (2015). The conception of essence and necessity advanced in Hale (2015) is similar in many ways to Fine’s, though the label “non-modal,” strictly speaking, does not apply to it, since Hale takes as basic certain modal truths, viz., the essential truths, from which other necessary truths are supposed to be derivable. For the purposes of this discussion, however, I am happy to subsume Hale’s account under the label, “non-modal,” as it is used in the text, since I really have in mind approaches that reject the reduction of essence to modality and hold instead that, if any such reduction is possible at all, it would have to proceed in the opposite direction. The contrasting model, which does take essence to be reducible to modality, dominated the metaphysical landscape for many decades starting in the 1970s; see, for example, Forbes (1985); Lewis (1986); Mackie (2009); and Plantinga (1974), for some representative examples.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

desiderata and decision points for hylomorphists  23 conception of essence. For Aristotle, the characteristics of a concrete particular object can be divided into the following three kinds: (i) those that are part of the essence proper; (ii) those that follow from the essence (viz., the so-called “propria” or “necessary accidents”); and (iii) those that are merely compatible with an object’s nature, but not required by it. For example, for Aristotle, while it is part of the essence of planets that they are heavenly bodies that are near, it is merely a de re necessary (but nonessential) feature of planets that they do not twinkle; the latter follows from, but is not strictly speaking part of, the essence of planets. Thus, the definition whose job it is to state the essence of or what it is to be a planet would have to make reference to the fact that planets are heavenly bodies that are near. But the definition should not also make reference to the fact that planets do not twinkle, since this would wrongly represent a derivative feature of planets (namely, their not twinkling) as a basic or non-derivative feature of them (namely, their nearness to the Earth).26, 27 Finally, other characteristics (e.g., a planet’s closeness to a comet) are merely compatible with the nature of planets, but neither part of nor dictated by their essence. In these cases, it is equally compatible with the nature of the concrete particular object that the characteristic either does or does not belong to the entity in question. Decision Point for Hylomorphists. Is the essence of a matter–form compound identical to its form? Or does the matter composing a concrete particular object also figure into its essence in some way, if only generically? Although hylomorphists tend to agree that the distinction between essence and accident should be placed front and center in their analysis of concrete particular objects, they do not all opt for the same strategy when it comes to the implementation of this desideratum with respect to their central piece of machinery, viz., the distinction between matter and form. In particular, the question arises as to how we are to understand the relation between the essence of a concrete particular object, on the one hand, and its form, on the other. According to some versions of hylomorphism (e.g., Frede (1985, 1987a); Frede and Patzig (1988); Lowe (1999); Whiting (1984, 1986, 1991, 1992)), the essence of a concrete particular object just is its form.28 According to other versions of hylomorphism (e.g., Peramatzis 26   See, for example, Aristotle, Posterior Analytics, A.13 and B.16. Aristotle thought of the planets and other heavenly bodies as being arranged and moving in accordance with a series of fixed heavenly spheres, with the Earth lying at the center. In Aristotle’s view, this arrangement of the heavenly bodies, including their movements and their position relative to the Earth, is not subject to change, since it is eternal, necessary, and perfect. 27   The idea that something might be part of the essence or included in the definition (a statement of the essence) of something else is no doubt at first sight puzzling and stands in need of elucidation. See Koslicki (2012a) and (2012b), for some relevant discussion. In Fine’s view, which will be considered in more detail in Section 5.3.2, essences are propositions (or collections of propositions) and these propositions are taken to have constituents. According to this conception, one can understand the idea of something being part of the essence of something else in terms of the notion of being a constituent of a proposition. Even if one does not want to go along with Fine’s propositional conception of essence, there is still the option of understanding something being part of the essence of something else in terms of the former being a constituent in the real definition of the latter. 28   I cite Lowe (1999) here as an example of a hylomorphist who takes the essence of a concrete particular object to be exhausted by its form (see, e.g., Lowe (1998), Ch. 9, “Matter and Form,” especially pp. 190–203).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

24  Concrete Particular Objects (2011)), a statement of the essence of a matter–form compound in some way includes a reference to its matter as well, e.g., by generically specifying the type of matter of which matter–form compounds of a certain kind are composed. Either way, hylomorphists can avail themselves of various options to accommodate the idea that some of the characteristics of a concrete particular object are part of its essence, while others are accidental to it (either by following from its essence or by being merely compatible with the object’s nature).29 (4)  Identity and Indiscernibility. Earlier, we saw that approaches according to which concrete particular objects are bundles of universals run into difficulties with respect to scenarios in which we are apparently presented with numerically distinct but qualitatively indiscernible concrete particular objects. Other approaches (e.g., This ascription, however, needs to be qualified in certain respects. For one thing, Lowe argues that the essence of a concrete particular object should not be taken to be a further entity that is numerically distinct from the object itself (see Lowe (2008)). In addition, there is some question as to whether Lowe should even be counted as a hylomorphist at all. After all, we do find Lowe, in various of his writings, distancing himself from hylomorphism, which he interprets as giving rise to a particular type of constituent ontology. Instead, Lowe argues that his own view, viz., the “four-category ontology” (see Lowe (2006)), which is inspired by Aristotle’s Categories, fits neither the constituent nor the relational mold (Lowe (2012b)). Lowe is happy to take on board the hylomorphist’s notion of form, but only as long as forms are not interpreted as constituents of concrete particular objects. Forms, in Lowe’s view, can be either construed as universals or as particulars. In the first case (i.e., if forms are construed as universals), they can be viewed as substantial kinds instantiated by concrete particular objects. Lowe cautions us, however, that instantiation is not to be reified as a further entity, i.e., a relation which holds between a kind and its instances. In the second case (i.e., if forms are construed as particulars), they can be taken to be particular instances of substantial kinds, in which case a concrete particular object simply is (i.e., is numerically identical to) a particularized form (see Lowe (1999)). Either way, no sense can be made, in Lowe’s view, of taking forms to be constituents of concrete particular objects. Concerning matter, Lowe remarks: “I have no serious need for the hylemorphist’s category of matter. I might be prepared to say that the ‘matter’ of the hydrogen atom is or consists of its proton and electron, but just in the sense that these are its parts and serve to compose it” (see Lowe (2012b), p. 237). Lowe also allows that the parts composing a concrete particular object need not themselves be concrete particular objects, namely, in cases in which a concrete particular object (e.g., a rubber ball) is composed of stuff (viz., a portion of rubber), though it may turn out that the stuff in question is ultimately composed of further concrete particular objects, e.g., microscopic particles (see Lowe (1999), pp. 3–5). Overall, then, we find that Lowe’s main concern is to distinguish his own view from a hylomorphic constituent ontology. But as long as we allow for non-constituent readings of hylomorphism, which also resist the collapse into a relational ontology, I see no obstacle to classifying Lowe as a certain type of hylomorphist, since he does seem to hold that the distinction between matter and form, when properly interpreted, is of use within a metaphysical theory of concrete particular objects. 29   Relatedly, accounts of concrete particular objects which do not make room for a distinction between what is essential and what is merely accidental to a concrete particular object also have trouble capturing the distinction between substantial and non-substantial change. (This issue was already hinted at in connection with the previous set of desiderata, when we asked “Why are concrete particular objects capable of persisting through some changes but not others?”.) Concrete particular objects undergo a non-substantial change when they change with respect to one or more of those of their characteristics that are neither part of their essence nor follow from their essence. In contrast, a substantial change requires that a concrete particular object either comes into, or goes out of, existence. Since the distinction between substantial and non-substantial changes is already covered jointly by (2) and (3), I will not list it separately in my catalog of desiderata and decision points. I very much agree with Brower’s assessment that the importance of capturing the contrast between substantial and non-substantial changes is often overshadowed in contemporary discussions by the almost exclusive focus on the problem of temporary intrinsics which is solely concerned with instances of non-substantial change (see Brower (2014), pp. 174–83).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

desiderata and decision points for hylomorphists  25 trope-theoretic bundle theories or bare particularist substratum theories) can respond to such scenarios by pointing to some non-qualitative difference which obtains between any two numerically distinct concrete particular objects, regardless of whether or not we suppose these objects to be qualitatively indiscernible (e.g., a distinction between numerically distinct tropes or bare particulars). A scenario such as that presented by Max Black raises questions about how to differentiate between numerically distinct objects at a single time. In addition to these questions concerning synchronic distinctness, however, we must also consider how different accounts treat the following questions concerning synchronic, diachronic, and cross-world identity: what, if anything, makes a single concrete particular object identical to itself synchronically, i.e., at a single time? What, if anything, makes a single concrete particular object that exists at one time diachronically identical to the very same concrete particular object existing at a different time, despite the fact that the object in question may have undergone qualitative change? And what, if anything, makes a single concrete particular object which exists in one world cross-world identical to the very same object existing in another world? As in the case of synchronic distinctness, accounts which allow that these questions concerning numerical identity have an informative answer at all find themselves taking recourse to some non-qualitative factor(s) or particular item(s) (e.g., a bare particular, haecceity, or collection of tropes) that is in a certain way associated with the concrete particular object at issue and whose numerical identity in turn is accepted as not further explicable in more basic terms. Decision Point for Hylomorphists. Not surprisingly, hylomorphists can proceed in a similar fashion, but some of the options that are available to these theorists are distinctive to the hylomorphic analysis of concrete particular objects: (i) to take facts about the numerical identity or distinctness of the matter composing concrete particular objects as basic and explain facts about the numerical identity or distinctness of concrete particular objects in terms of facts about the numerical identity or distinctness of their matter; or (ii) to take facts about the numerical identity or distinctness of forms as basic and explain facts about the numerical identity or distinctness of concrete particular objects in terms of facts about the numerical identity or distinctness of their forms. In order for the first option to be viable, the matter composing a concrete particular object must be of a suitable ontological type if it is to serve as the principle of individuation for concrete particular objects; similarly for the form, according to the second strategy. Thus, the route taken by hylomorphists in response to questions concerning the numerical identity and distinctness of matter–form compounds can be expected to impact their position with respect to the ontological categorization of matter and form which was brought up in connection with the previous set of questions outlined in (3). (5)  Material Constitution. The Problem of Material Constitution concerns the relation between a concrete particular objects and what constitutes it, e.g., the relation between a clay statue and the clay of which it is made. The project of arriving at a plausible

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

26  Concrete Particular Objects understanding of this relationship has proven to be challenging for metaphysicians, in part because of the following reasons. For one thing, since the relation in question is obviously intimate, there is pressure to regard a concrete particular object and what constitutes it as numerically identical: they do, after all, occupy the same region of space-time and share many of their characteristics (e.g., their weight, shape, color, texture, etc.). At the same time, we also observe that some apparent differences obtain between a concrete particular object and what constitutes it (e.g., temporal or modal differences): for example, a clay statue and the clay that constitutes it can begin or cease to exist at different times and they appear to have different persistence conditions (e.g., the clay can survive smashing, while the clay statue cannot, etc.). The latter apparent differences which obtain between a concrete particular object and what constitutes it create pressure to regard these items as numerically distinct. Going this route, however, appears to saddle us with the consequence that regions of space-time which we would normally think of as being occupied by only a single concrete particular object in fact contain multiple numerically distinct but spatiotemporally coincident objects. The hylomorphic analysis of concrete particular objects seems well suited to yield a solution to the Problem of Material Constitution. After all, at the very heart of the doctrine lies the distinction between a concrete particular object, on the one hand, and its matter and form, on the other hand. Thus, hylomorphists are likely to understand the notion of constitution in terms of the relation between a concrete particular object and its matter. Since a concrete particular object, according to the hylomorphic analysis, is not exhausted by its matter, such approaches have a way of capturing the second set of pressures identified earlier, according to which a concrete particular object and what constitutes it ought to be viewed as numerically distinct. Some hylomorphists (e.g., Rea (1998)) have also tried to accommodate the first set of pressures by allowing for a notion of sameness that is to be distinguished from numerical identity. While hylomorphism thus delineates the general outlines of a response to the Problem of Material Constitution, different versions of the doctrine have developed the details of their respective proposed solution in different ways. Decision Point for Hylomorphists.  What sort of solution to the Problem of Material Constitution is adopted by a specific version of the hylomorphic analysis of concrete particular objects will therefore depend on how the theorist in question conceives, firstly, of the matter composing a concrete particular object and, secondly, of the relation which holds between a concrete particular object and its matter. Hylomorphists who take concrete particular objects and their constituting matter to be numerically distinct should ideally also have something to say in response to the concern that they are committed to a rampant multiplication of numerically distinct but spatiotemporally coincident objects. Some of the different directions which are available to hylomorphists in this connection will be considered in more detail later. (6)  Mereological Composition. As noted in the Introduction, one of the main motivations for the hylomorphic view I developed in Koslicki (2008a) was my desire

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

desiderata and decision points for hylomorphists  27 to formulate a more full-blooded conception of what it is to be an object than that associated with the still very popular theory of parts and wholes known as standard or classical extensional mereology (see, for example, Leonard and Goodman (1940); Lewis (1991); Simons (1987)). According to the latter, any plurality of objects, no matter how disparate and gerrymandered, itself composes a further object, even when the objects in question fail to exhibit interesting similarities, internal unity, cohesion, or causal interaction amongst each other. The resulting exceedingly deflationary conception of objecthood, parthood, and composition, however, is not only committed to all sorts of counterintuitive objects (such as Lewis’ infamous “trout-turkey”); it is also unable to draw such intuitively obvious distinctions as that between, say, a properly assembled motorcycle in running condition and a heap of disassembled motorcycle parts. In contrast, according to my own approach, composition is restricted and takes place only when the material parts which come to compose a concrete particular object satisfy certain conditions which are set by the formal components that are associated with the whole in question. In particular, in order for a plurality of objects to come to compose a further object, these objects must satisfy two kinds of constraints, viz.: (i) type constraints, and (ii) configuration constraints. To illustrate, in order for some objects to come to compose an H2O molecule, the objects in question must, first, be of the right kind (viz., in this case, two of the objects must be hydrogen atoms and one an oxygen atom); and, second, these objects must be configured in the right way (viz., the atoms in question must be arranged in the manner of chemical bonding, which requires them to share electrons). In addition, I argued that the organizing principle that is associated with structured wholes (viz., their form) is itself literally and strictly speaking part of the whole it organizes, so that structured wholes, on my conception, turn out to be compounds of matter and form. There is thus, on this view, an important difference between a properly assembled motorcycle in running condition and a heap of disassembled motorcycle parts. In the first scenario, the motorcycle parts in question compose a further object, viz., a motorcycle, while in the second the very same motorcycle parts do not compose a further object, since they do not satisfy the structural constraints associated with any genuine kind to which we want to be committed for independent reasons. As a result, the neo-Aristotelian conception of concrete particular objects to which I am sympathetic promises to yield a mereology that aligns more naturally with the scientifically informed common-sense ontology to which we are committed for reasons unrelated to the mereological framework we adopt. Decision Point for Hylomorphists. Should hylomorphists aim for a restricted notion of mereological composition? If so, how can the apparatus of matter and form be put to use to impose an appropriate restriction on mereological composition? It is certainly open to hylomorphists to endorse restricted notions of composition, though not all of them do (e.g., Fine (1982); Johnston (2002, 2006)). Even for those hylomorphists who are sympathetic to such mereologies, a further question arises as to how exactly

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

28  Concrete Particular Objects composition is supposed to be restricted. Presumably, form plays some role in this; beyond that, however, hylomorphists can avail themselves of different methods by which the relevant restrictions are to be imposed.30 (7)  Unity. Aristotle holds that some complex entities (viz., wholes) are more unified than others (e.g., heaps). In a celebrated regress argument which is given at the end of Met. Z.17, Aristotle illustrates the distinction he has in mind by citing the syllable, “BA,” and flesh as examples of unified wholes. We can see from considering these cases, so Aristotle reasons, that there must be more to a syllable or to flesh besides the matter that composes them: a “something else” (heteron ti). For it is possible for the syllable, “BA,” or flesh to be dissolved into its material parts, viz., the elements (stoicheia) of which they consist (the letters “B” and “A”; earth, air, fire, and water). In this case, the unified whole in question (the syllable; flesh) is gone, but the elements of which it previously consisted (the letters; earth, air, fire, and water) are still there. What could this “something else” be? Aristotle considers three possibilities: either (i) the “something else” is itself an element; or (ii) it is composed of elements; or (iii) it is neither an element nor composed of elements. The first and the second cases, Aristotle reasons, lead to a regress. As a result, he embraces the third possibility; the “something else” is neither itself an element nor composed of elements: The syllable, then, is something—not only its elements (the vowel and the consonant) but also something else; and the flesh is not only fire and earth or the hot and the cold, but also something else. [. . .] But it would seem that this is something [ti touto], and not an element [stoicheion], and that it is the cause [aition] which makes this thing flesh and that a syllable. And similarly in all other cases. And this is the substance of each thing [ousia de hekastou men touto]; for this is the primary cause of its being [touto gar aition prōton tou einai]; and since, while some things are not substances, as many as are substances are formed naturally and by nature, their substance would seem to be this nature [phusis], which is not an element but a principle [archē]. An element is that into which a thing is divided and which is present in it as matter [hulē], e.g. a and b are the elements of the syllable.  (Met. Z.17, 1041b16–33)

Based on this regress argument, Aristotle takes himself to have established by the end of Met. Z.17 that there is more to a unified whole than its matter. Assuming, as is natural, that the “something else” in question is the form, we therefore reach the conclusion that unified wholes must be conceived of as compounds of matter and form. By contrast, as the case of heaps illustrates, pluralities of elements which are not unified by the presence of form within them can be expected to lack the high degree of unity characteristic of matter–form compounds.31 Decision Point for Hylomorphists. But how exactly does the presence of form in a matter–form compound help to account for the degree of unity that is manifested by these 30   Given that I have already weighed in on these mereological questions in Koslicki (2008a) and in other previously published work, I will not devote much attention to these topics in this study. 31  I discuss my reading of Aristotle’s regress argument in Met. Z.17 in more detail in Koslicki (2006, 2008a, and 2014). See also Mann (2011).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

desiderata and decision points for hylomorphists  29 objects? Sadly, the notion of unity has up to this point not nearly received the attention it deserves from contemporary metaphysicians. Those theorists who have tried to offer accounts of unity tend to approach this subject matter from an Aristotelian perspective (see, for example, Hoffman and Rosenkrantz (1999)). It is common among hylomorphists to follow Aristotle’s lead in designating at least a certain privileged class of matter–form compounds as manifesting a high degree of unity: traditionally, the privileged class of unified wholes has been thought to include living organisms and perhaps other natural concrete particular objects, i.e., those which are not designed or created by intentional agents but are rather the results of natural processes. The high degree of unity which is ascribed to matter–form compounds is commonly traced in some way to the presence of form within these matter–form compounds. More remains to be said, however, about how exactly the form of a matter–form compound is thought to carry out its responsibility of tying together the material parts composing a matter–form compound into a unified whole. The topic of unity will be taken up again in more detail in Chapter 7. (8)  The Grounding Problem. The Grounding Problem poses a challenge especially for those who opt to respond to the Problem of Material Constitution by positing numerically distinct spatiotemporally coincident objects. For coincidence theorists, such scenarios are by no means a rare and exceptional metaphysical possibility; rather, we are presumably surrounded by vast numbers of actually coincident objects, everywhere we look. Given this approach, then, the question arises as to how these numerically distinct but spatiotemporally coincident objects can be different in various respects (e.g., with respect to their temporal or modal characteristics) despite the fact that they are similar in so many other respects. In particular, coincident objects are apparently composed of the same material parts; they share the same location in space and time; and they agree with respect to all those characteristics which we would normally trace to the fact that they share the same matter and spatiotemporal location (e.g., weight, shape, texture, color, etc.). The Grounding Problem thus prompts coincidence theorists to offer some account of what grounds the apparent differences between numerically distinct spatiotemporally coincident objects. The Grounding Problem certainly poses a serious challenge for coincidence theorists, whether they embrace hylomorphism about concrete particular objects or not. And while the doctrine of hylomorphism does not by itself generate a complete solution to the Grounding Problem, it does provide those who are sympathetic to this approach with some promising directions to pursue which may or may not be as easily accessible to other coincidence theorists who do not also subscribe to hylomorphism.32 In particular, coincidence theorists can hope to make some headway toward finding a solution to the Grounding Problem, if they combine their analysis of concrete particular objects with a non-modal conception of essence, mentioned earlier in connection 32   For arguments to the effect that hylomorphism does not contribute a distinctive solution to the Grounding Problem, see Sidelle (2014).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

30  Concrete Particular Objects with (3). According to such approaches to the essence/accident distinction, facts about the essences of concrete particular objects are taken as basic and facts about their modal properties (i.e., facts about what is necessary but non-essential to these objects) are regarded as, at least in part, derivable from these primitive essential facts. As a result, a hylomorphic analysis of concrete particular objects, when combined with a suitable approach to the essence/accident distinction, can respond to the challenge posed by the Grounding Problem. My own approach, as will be elaborated further in Chapter 4, proposes to explain the modal differences which apparently obtain between numerically distinct spatiotemporally coincident objects in terms of differences between their respective essences, together with whatever further considerations turn out to be relevant to the derivation of the non-essential features of hylomorphic compounds from facts about their essences. Decision Point for Hylomorphists. The questions posed by the Grounding Problem thus take us back to some of the decision points for hylomorphists which were brought up earlier in connection with the essence/accident distinction: should hylomorphists opt for a non-modal conception of essence or a modal conception of essence? If the former, how do the necessary accidents follow from the essence of a concrete particular object? If the latter, what is the proper hylomorphic response to the Grounding Problem? Hylomorphists who go the non-modal route can opt for the response to the Grounding Problem outlined earlier. For those who reject a non-modal conception of essence, it remains to be seen what sort of response to the Grounding Problem is in the offing.33

1.7 Conclusion In this chapter, I have reviewed existing approaches to the metaphysics of concrete particular objects and positioned the doctrine of hylomorphism, according to which concrete particular objects are to be analyzed as compounds of matter and form, with respect to these competing accounts. The literature is divided over whether concrete particular objects are or are not further analyzable into constituents which do not themselves belong to the ontological category of concrete particular objects and in terms of which the character of these latter entities is to be explained. Constituent ontologies (e.g., bundle theories or substratum theories) attribute such a constituent structure to concrete particular objects and appeal to this structure in explaining the character of concrete particular objects. By contrast, non-constituent ontologies (e.g., Platonism or austere nominalism) deny that concrete particular objects are further analyzable into constituents which do not themselves belong to the same ontological 33   In collecting together these desiderata and decision points, I have been very much helped and influenced by the discussion of hylomorphism in Brower (2014) and Rea (2014). In his discussion of Aquinas’ version of hylomorphism, Brower emphasizes many of the issues cataloged here, in particular: property possession, change, the essence/accident distinction, identity, and material constitution. Rea (2014) addresses many of these topics as well, in his exposition, and points out furthermore that hylomorphism is naturally combined with a restricted notion of mereological composition.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

conclusion  31 category as the wholes or complex entities to which they belong; such frameworks therefore do not attempt to explain the character of concrete particular objects by appeal to such a constituent structure. In our brief survey of these non-hylomorphic approaches to the metaphysics of concrete particular objects, we considered prominent objections which have been raised against these accounts, though without being able to enter into these debates in any detail. This overview in turn gave rise to a set of desiderata and decision points which hylomorphists should keep in mind as they formulate their own competing analysis of concrete particular objects. In particular, hylomorphists encounter various choices when they put to use their central piece of machinery, viz., the distinction between matter and form, with respect to the following problem areas: (1) Property Possession and the Problem of Universals: a concrete particular object’s relation to its own properties and those it apparently shares with objectively similar objects; (2) Causation, Explanation, and Change: a concrete particular object’s character at a time, how it comes into and goes out of existence, and its ability to persist through certain kinds of changes; (3) Essence/Accident: what is part of the essence of a concrete particular object and what belongs to it merely accidentally; (4) Identity and Indiscernibility: the source of facts concerning the numerical identity and distinctness of concrete particular objects; (5) Material Constitution: the relation between a concrete particular object and what constitutes it; (6) Mereological Composition: the conditions which must be met in order for a plurality of concrete particular objects to compose a single such object; (7)  Unity: how a unified whole can emerge from a plurality of material parts; (8) The Grounding Problem: how numerically distinct spatiotemporally coincident objects can nevertheless differ with respect to their temporal or modal profile. The explanatory goals identified here will continue to guide us in our development of a hylomorphic analysis of concrete particular objects in the chapters to come.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

2 Matter 2.1  Introductory Remarks In response to the various desiderata and decision points discussed in Chapter 1, it emerged that a hylomorphic analysis of concrete particular objects can be developed in a variety of ways, depending on how hylomorphists conceive of the matter composing a concrete particular object: the form that is associated with a concrete particular object; and the relation(s) which hold between a concrete particular object, its matter, and its form. In this chapter, our focus is on the first of these questions, viz., how should hylomorphists conceive of the matter composing matter–form compounds? The three main options, as I see it, which are open to hylomorphists in connection with this question, are as follows: (i.a) Matter as Prime Matter: The matter of which matter–form compounds are composed is prime matter. (i.b) Matter as Stuff: The matter of which matter–form compounds are composed is stuff. (i.c) Matter as Hylomorphic Compounds: The material parts of matter–form compounds are themselves matter–form compounds. In what follows, I examine representative versions of the first two approaches—(i.a) and (i.b)—in detail and argue that both give rise to serious difficulties. As a result, I opt for the third approach—(i.c)—which conceives of the matter composing a concrete particular object as belonging to the same ontological type as the concrete particular object itself which it composes. A concrete particular object, on this approach, is ­analyzed as a matter–form compound; the matter composing it, in cases in which an object is composed of more than a single material part, as a plurality of matter–form compounds.1 This approach does not immediately put to rest all of the difficult questions 1   Option (i.c) is compatible with the possibility of infinitely descending chains of matter–form compounds whose material parts are themselves matter–form compounds. In the face of other possibilities, however, which we will consider in more detail later in this chapter, I adopt a more precise formulation of the third approach which contains the following important additional proviso: “ . . . unless or until we reach an empirically confirmed level in the compositional hierarchy at which the material parts of matter–form compounds are not themselves structured wholes.” If we ever were in possession of empirical evidence which indicates that, at some level in the compositional hierarchy, the material parts of a concrete particular object are not themselves structured wholes, then proponents of (i.c) should concede that, at that point, hylomorphism no longer applies. Such a possibility should not be interpreted as a counterexample to the

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

matter as prime matter  33 hylomorphists face as they weigh up different possible responses to the question of how best to conceive of the matter composing matter–form compounds. Moreover, as will emerge, the hylomorphic approach to matter also gives rise to several explanatory challenges of its own. Overall, however, I will argue that (i.c) should nevertheless strike hylomorphists as the most promising option for the purposes of developing an analysis of concrete particular objects that is compatible with our best scientific theories concerning the natural world.2

2.2  Matter as Prime Matter 2.2.1  Prime Matter in Aristotle There is some controversy in the scholarly literature as to whether Aristotle should be interpreted as being committed to prime matter.3 According to those who read him in this way, the commitment to prime matter emerges primarily from the following positions Aristotle endorses in connection with his analysis of change. Firstly, as was noted in the previous chapter (see Section 1.5), Aristotle proposes to solve the Parmenidean puzzle concerning change by positing that something always persists through any change that occurs (viz., the matter or subject underlying the change) and something else does not persist through the change (viz., the form or privation). Secondly, Aristotle subscribes to the Empedoclean doctrine that everything in the sublunary sphere is ultimately composed of the four elements (viz., earth, air, fire, and water) in different doctrine, but rather as a reason to restrict its scope. Other reasons for restricting the scope of the hylomorphic analysis are already known to us, e.g., in the case of domains (e.g., that of set theory) consisting of complex entities which are not structured wholes. Even so, the doctrine of hylomorphism still finds plenty of application with respect to complex entities that are independently recognized as structured wholes. If the proviso ever did come into effect, however, we should not immediately conclude that, in these circumstances, either of the alternative conceptions of matter— (i.a) or (i.b)—should be favored. In fact, we will encounter reasons for thinking that the representative formulations of (i.a) and (i.b), which will be examined in this chapter, should in any case be rejected on independent grounds. For reasons of simplicity, I will, for the time being, use the simpler formulation of (i.c) given above, without the important qualification just specified. 2   In this chapter, I address some, but not all, of the desiderata outlined in the previous chapter. This selective attention is to be expected, since some of the issues in question more naturally fall under the ­purview of the responsibilities assigned to form or the relation(s) between a concrete particular object, its matter, and its form, while others are more likely to be settled by the contributions of matter. In particular, the following will figure prominently in this chapter: (2) “Causation, Explanation, and Change”; (4) “Identity and Indiscernibility”; (5) “Material Constitution”; and, perhaps somewhat surprisingly, (7) “Unity” and (8) “The Grounding Problem.” Many of these topics will continue to be of concern to us in subsequent chapters as well. The remaining desiderata will receive closer attention when we turn to the discussion of form and the hylomorphic relations in the next two chapters: (1) “Property Possession and the Problem of Universals”; and (3) “Essence and Accident.” (6) “Mereological Composition” will mostly stay in the background in the present study, since I have already dealt with these issues extensively in Koslicki (2008a). 3  The attribution of prime matter to Aristotle is challenged, for example, by King (1956) and in W. Charlton’s commentary on Aristotle’s Physics, Books I and II (see especially the Appendix in Charlton (1970), pp. 129–45). For rebuttals of King’s and Charlton’s arguments, see Robinson (1974) and Solmsen (1958). Further discussion of the controversy, as it pertains to On Generation and Corruption, can also be found in the Appendix to C. J. F. Williams’ commentary on this work (see Williams 1982, pp. 211–19). Williams sides with Solmsen and Robinson, over King and Charlton.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

34  matter ratios. Third, he held that it is possible for the four elements to transform into one another and that, when they do, these processes constitute substantial changes, rather than mere alterations (i.e., qualitative, non-substantial changes). To illustrate, air comes into being and water ceases to exist, in Aristotle’s view, when the combination of wet and cold (the so-called “primary qualities” or “contrarieties” which are characteristic of water) is replaced with the combination of wet and hot (the primary qualities characteristic of air). These three positions together seem to require that something must also underlie these elemental transformations and, furthermore, that it cannot be part of the essence of the subject persisting through these substantial changes to exhibit any one of the four pairs of primary qualities in particular (viz., wet/hot, wet/dry, moist/ hot, moist/dry). Since Aristotle also believes that all the other perceptible qualities can in some way be explained by appeal to these primary qualities, it follows that the ­subject which persists through these elemental transformations has only accidentally whatever other perceptible qualities are associated with it (if any). The combination of views Aristotle endorses thus seems to lead to the result that he is committed to what is traditionally identified as prime matter: a kind of matter which, among other things, is capable of persisting through substantial change even at the most basic elemental level and which has no perceptible qualities essentially. Certain passages in Aristotle’s Metaphysics can also be interpreted as referring to prime matter, such as the following remarks from Met. Z.3:4 By matter I mean that which in itself is neither a particular thing nor of a certain quantity nor assigned to any other of the categories by which being is determined. [. . .] Therefore the ultimate substratum is of itself neither a particular thing nor of a particular quantity nor otherwise positively characterized; nor yet negatively, for negations also will belong to it only by accident. (Met. Z.3, 1029a20–6)5

According to these remarks, matter in the sense of the ultimate substratum is that which underlies all other determinations, while none of them apply to it essentially. Immediately following the passage just cited, the kind of matter under discussion is also characterized as not separable (chōriston) and not a this (tode ti) (Met. Z.3, 1029a26–30).6 4   Bostock (1994), for example, reads the so-called “striptease” argument in Met. Z.3 as beginning with the perceptible matter of a concrete particular object, e.g., the bronze which constitutes a statue (Bostock (1994), pp. 76ff), and then proceeding to “strip away” conceptually all of the attributes, including the dimensions (viz., length, breadth, and depth), which are associated with whatever entity forms the starting point of the thought experiment in question. What remains, in thought, according to Bostock, is prime matter which is not perceptible. Only prime matter, so Bostock argues, can survive the “striptease” Aristotle performs in his Met. Z.3 thought experiment, since nothing belongs to it essentially. Other commentators, however, read the “striptease” argument in Met. Z.3 differently. Gill (1989), for example, takes Aristotle to be objecting to a competing conception of matter which he rejects (most likely the conception of matter that is operative in Plato’s Timaeus). 5   This and all subsequent quotations from Aristotle’s Metaphysics come from the translation by W. D. Ross (see Barnes (1984)). 6   An engagement with the exceedingly difficult question of how to make sense of Aristotle’s notions of separability and thisness would take us too far afield for present purposes. If the kind of matter under ­discussion in passages such as those from Met. Z.3 just cited is in fact prime matter, then presumably the

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

matter as prime matter  35 Furthermore, in Met. Z.10, Aristotle speaks of matter as in itself unknowable (agnōstos), since nothing can be said of it kath’ hauto, i.e., in virtue of itself or essentially (Met. Z.10, 1036a8–9). In order to know each thing fully, in Aristotle’s view, we must grasp its essence, as stated by the definition (see, for example, Met. Z.1, 1028a36–1028b7). In sum, Aristotelian prime matter (if there such a thing) can be described as follows: it is what underlies the elemental transformations; it is in itself neither a particular thing nor of a particular quantity, nor assigned to any other of the categories by which being is determined; it lacks an essence, since neither positive nor negative characterizations apply to it kath’ hauto; it cannot exist on its own; and, since it lacks an essence, it is in itself unknowable. Prime matter, so characterized, is thus even more “bare” than the substratum posited by bare particularists (see Section 1.4): for, in addition to ­lacking an essence or intrinsic nature, prime matter is also said to be devoid of all particularity.7

2.2.2  Thomistic Prime Matter While commentators disagree strongly over whether a commitment to prime matter should even be ascribed to Aristotle, there is much less scholarly controversy in the case of Aquinas over whether he was committed to prime matter, though the question of how best to interpret Aquinas’ notion of prime matter has generated heated discussion among Thomists (see, for example, Pasnau (2014)). In what follows, I will consider the Thomistic version of hylomorphism defended in Oderberg (2007) in detail, as a representative formulation of the view stated earlier in (i.a) according to which the matter composing a hylomorphic compound is prime matter. Oderberg, moreover, is engaged in a project very similar to my own, since he takes himself to be putting forward a position which is intended both as a contribution to the historical scholarship and as a contribution to the contemporary debate concerning the metaphysics of c­ oncrete particular objects. In Section 2.3 we turn to a different understanding of the Thomistic doctrine of prime matter, viz., that developed in Brower (2014), whose hylomorphic account is also put forward in the spirit of contributing a denial of separability can be understood at least to entail a modal existential claim, viz., that prime matter cannot exist on its own, without being affiliated with some particular matter–form compound which it composes. The denial of thisness can be construed, for example, in terms of either particularity or determinateness, i.e., as claiming that prime matter lacks a particular or determinate nature. 7   Is Aristotelian prime matter (if there is such a thing) a type of entity? That depends on what we mean by “entity.” If the term is used very broadly, then I suppose the answer would be “yes” (according to those who read Aristotle as being committed to prime matter), in the sense that prime matter is a type of being, though one to which all the qualifications cited in the text apply. But if the term is used more narrowly, e.g., to apply to those types of being which belong to one of Aristotle’s categories (i.e., substance, quality, quantity, etc.), then the answer would be “no.” If by “entity” we mean “thing,” then the question is difficult to evaluate in Aristotle’s case, since he does not employ a separate term which straightforwardly corresponds to our term, “thing.” Both Oderberg and Brower would answer the latter question negatively for Thomistic prime matter; in Brower’s case, the reason for this assessment is that he assigns Thomistic prime matter to the ontological category of stuff, which he takes to be fundamentally different from that of things. Both, however, would agree that the first two answers correctly describe Thomistic prime matter as well: it is a type of being, broadly construed, but not one which belongs to any of the Aristotelian categories.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

36  matter position that is viable both on historical grounds as an interpretation of Aquinas, and on philosophical grounds as a plausible contribution to the contemporary debate concerning the metaphysics of concrete particular objects. Brower’s interpretation of Thomistic prime matter is inspired by a conception of stuff developed by Ned Markosian, which will be examined here as well as a representative version of (i.b), viz., the matter-as-stuff hypothesis. According to Oderberg’s Thomistic hylomorphism, those concrete particular objects which he calls “material substances” are compounds of prime matter and substantial form:8 In its primary application, the distinction between form and matter applies in a quite literal way to material substances. Every material substance is a literal compound of two elements— prime matter and substantial form. Prime matter is the underlying substrate, itself wholly undifferentiated, which form actualizes to produce a material substance, be it inorganic, such as a lump of rock or a molecule of water, or organic, such as a bacterium, a tree, or a cow. (Oderberg (2014), p. 164)9

Oderberg conceives of the matter composing a matter–form compound as something that is not already itself a matter–form compound or composed of matter–form compounds: . . . [T]he soul is the form of something else, something not itself shot through by the very soul to which it is united—and this is what the hylemorphist calls primordial matter, or prime matter.  (Oderberg (2007), pp. 71–2) [Prime matter] is a pure passive potentiality, without any form whatsoever, nor subject to any privation (i.e., it does not lack some form that it needs, in the way that a blind person is deprived of sight), but it is wholly receptive of any form whatsoever. It is the completely undifferentiated basic material of the physical universe. It is not something, in the sense of something or other, but it is not nothing either. It is the closest thing there is in the universe to nothingness without being nothingness, since it has no features of its own but for the potential to receive substantial 8   Following Aquinas, Oderberg holds that only a single substantial form is associated with each matter–form compound, a position that is known as “unitarianism” about substantial forms. 9   In Oderberg’s view, not all concrete particular objects belong to the privileged category for which he reserves the label, “material substances.” Among concrete particular objects, Oderberg also recognizes a class he calls “accidental unities” which includes, for example, such objects as fists (i.e., hands that are clenched) and artifacts. Accidental unities, in Oderberg’s view, are not compounds of prime matter and substantial form; rather, they are matter–form compounds (or the actual parts of matter–form compounds) “taken together” with some of their accidental modes; e.g., a fist is a hand “taken together” with its accidental mode of being clenched. For the time being, when I speak of “concrete particular objects,” I am restricting myself to the privileged class of concrete particular objects which, in Oderberg’s view, would count as genuine matter–form compounds (or as actual parts of genuine matter–form compounds). I purposefully do not refer to the privileged class of concrete particular objects in question as “material substances,” since the question of whether, and in what sense, these entities should be regarded as substances raises difficult issues which will be examined in detail later. For now, I simply refer to the entities in question as “concrete particular objects” or “matter–form compounds,” while leaving open exactly what kinds of entities are to be included in or excluded from this category. Oderberg accepts as examples of the privileged category in question living organisms as well as members of other natural but non-living kinds (e.g., chemical compounds).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

matter as prime matter  37 forms. It is changeless, but is the support of all substantial change, and as such is subject to numerical identity, so that prime matter is conserved throughout substantial change. (Oderberg (2007), p. 72)

As these passages indicate, among Oderberg’s main motivations for positing prime matter is to help explain how matter–form compounds can come into existence and cease to exist, as stated in the second set of desiderata, “Causation, Explanation, and Change,” identified in Section 1.6. He argues that prime matter is needed to account for the fact that something always persists through any kind of substantial change; and this something, in his view, cannot itself be a matter–form compound or composed of matter–form compounds. In addition to the role ascribed to prime matter as the ultimate subject of substantial change, Oderberg allows that “loosely speaking” we can ascribe a certain nature to prime matter, though “strictly speaking” prime matter has no essence: [T]here is nothing wrong with speaking of the nature of prime matter as pure passive potency, as long as we take ‘nature’ loosely and not as meaning essence in the strict sense. Strictly, prime matter has no essence. Loosely, it has the nature of being pure potentiality unmixed with any determining form, substantial or accidental. [. . .] [P]rime matter . . . has no appearance and does not of itself come in arrangements. It is [. . .] radically disposed to dimensionality, but this is manifested wholly through the forms that prime matter takes on.  (Oderberg (2007), p. 76)

In sum, Oderberg ascribes the following characteristics to prime matter: it is wholly undifferentiated; pure passive potentiality; without any form whatsoever; not subject to any privation (since it does not lack a form that it requires); wholly capable of receiving any accidental or substantial form whatsoever; without features of its own, except for its capacity to receive forms; in itself changeless, but the support of all substantial change; and subject to numerical identity, since prime matter is conserved while ­matter–form compounds come into and go out of existence.

2.2.3  Difficulties for Oderberg’s Thomistic Prime Matter Hypothesis In what follows, I argue that the Thomistic doctrine of prime matter, as it is developed by Oderberg, is subject to serious difficulties, among them at least the following six: (1)  Potential Metaphysical Overreaching: The doctrine of prime matter, even though it is not intended to be put forward as an empirical hypothesis, has the potential to lead to overreaching on the part of the metaphysician into the scientific terrain of fundamental physics. (2)  Non-hierarchical Conception of Matter: The Thomistic conception of matter is problematically non-hierarchical: prime matter can receive any substantial form directly, without requiring any actual matter–form compounds acting as compositional intermediaries. (3)  Threat of Contradictions: Like bare particularism, the doctrine of prime matter threatens to give rise to contradictions; prime matter is said to lack an essence and yet

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

38  matter various characterizations are attributed to it which do not appear to be merely accidental to prime matter. (4)  Other Possibilities: Alternative possibilities for what might underlie substantial change even at the most basic micro-physical levels have not yet been ruled out and therefore should be considered. (5)  Non-particularity: Given the non-particularity of prime matter, we cannot appeal to it to help us resolve difficult questions concerning the numerical identity or distinctness of matter–form compounds. (6)  Material Constitution, Unity and the Grounding Problem: The Thomistic doctrine of prime matter allows for an eliminativist solution to the Problem of Material Constitution, the Problem of Unity, and the Grounding Problem, but only when combined with unitarianism about substantial forms (see n. 8) and a distinction between actual and so-called “virtual” parts. If both of these positions are rejected, as I think they should be, then no solution to these problems is, as of yet, readily available. Due to these difficulties, I take the Thomistic doctrine of prime matter, as it is developed by Oderberg, to be not sufficiently well-motivated to warrant taking it on board over competing conceptions of matter. Moreover, in the course of our evaluation, we will find that alternative conceptions of matter prove to be more successful in meeting the desiderata identified in Chapter 1 than Oderberg’s doctrine of prime matter. (1)  Potential Metaphysical Overreaching. Oderberg’s doctrine of prime matter is intended to be put forward as a metaphysical postulate, rather than as an empirical claim. Nevertheless, one might be concerned that, in advancing the prime-matter hypothesis, Oderberg is guilty of metaphysical overreaching into the scientific terrain of fundamental physics. For one thing, Oderberg’s account assumes that all concrete particular objects are ultimately composed of a single type of matter, which is not in turn composed of anything more basic that is also material.10 Secondly, Oderberg’s approach requires that the single type of ultimate matter in question, which allegedly constitutes the bottom layer of the compositional hierarchy, has the characteristics 10   Oderberg is comfortable using the language of parthood and composition when he describes the r­ elationship between prime matter, substantial form, and the concrete particular object which results from their combination. However, he does want to distinguish between hylomorphic composition (i.e., the sense in which a concrete particular object is a compound of prime matter and substantial form) and mereological composition (i.e., the sense in which a whole is composed of its parts), as the following passage for example makes clear:

Now substantial form is intrinsic since it is a constituent solely of the substance. It is a constituent because it is a real part or element of it, though not on the same level as a substance’s natural parts such as the branch of a tree or the leg of a dog. Rather, substantial form (or ‘form’ for short) is a radical or fundamental part of the substance in the sense of constituting it as the kind of substance it is. It is a principle in the sense of being that from which the identity of the substance is derived . . . And form actualizes the potencies of matter in the sense of being the principle that unites with matter to produce a finite individual with limited powers and an existence circumscribed by space and time. (Oderberg (2007), p. 66) See also Oderberg (2012), especially pp. 21–2, for further elaboration.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

matter as prime matter  39 he ascribes to prime matter. For example, as noted earlier, the single type of ultimate matter in question is described as “wholly undifferentiated;” “pure passive potentiality” or “potency;” “wholly receptive of any form whatsoever;” “not something, in the sense of something or other,” but “not nothing either;” “in itself changeless,” but “conserved throughout substantial change;” “subject to numerical identity;” something which “has no appearance and does not of itself come in arrangements;” and is “radically disposed to dimensionality.” Some of these characteristics are formulated in terms that would be considered alien from the point of view of contemporary physics. This disconnect in itself should give us pause, since it stands in the way of our ability to adjudicate or even recognize potential conflicts between what metaphysicians and scientists, respectively, are saying about processes of generation and destruction which take place at micro-physical levels. In those instances, however, in which Oderberg’s ascriptions are interpretable from the point of view of contemporary physics, his commitments would need to be checked against the predictions of our best current scientific theories of matter. Since Oderberg regards his doctrine of prime matter as a contribution to metaphysics rather than science, he does not attempt to provide empirical support for the claim that the material universe ultimately bottoms out in a single type of basic substratum that answers to the characteristics he ascribes to prime matter. As the following passage indicates, however, Oderberg nevertheless at least leaves open the possibility that the metaphysical prime matter hypothesis might converge with the postulates of fundamental physics: [. . .] [M]ight prime matter be energy? It is an intriguing question that I cannot pursue here. One problem is that the hylemorphist has a better grasp of what prime matter is than the physicist has of what energy is, and since metaphysics has to be informed by science there will be severe limits to what the former can say about the possible identification of prime matter with energy. If there are substantial energy transformations (e.g. heat to sound, chemical to light) by which a wholly new kind of thing comes into existence, there will have to be prime matter distinct from energy as a support [. . .]. But if such transformations are but phases of an underlying pure energy that has no determinate form in itself, then perhaps one might venture the thought that they are one and the same.  (Oderberg (2007), p. 76)

(2)  Non-hierarchical conception of matter. Our current empirically supported conception of matter is (or at least appears to be) hierarchical: atoms are said to be made up of protons, neutrons, electrons, etc.; molecules are said to be made up of atoms; proteins are said to be made up of molecules; etc. According to Oderberg, however, the very same type of matter, viz., prime matter, which is capable of receiving, say, the form of a quark (thereby giving rise to a quark), is also capable of receiving the form of a human being (thereby giving rise to a human being). In order to reconstruct the appearance of a compositional hierarchy, Oderberg supplements his non-hierarchical notion of prime matter with a distinction between actual and merely virtual parts  of  a matter–form compound. To illustrate, quantities of water, in Oderberg’s view, contain H2O molecules as actual parts, while hydrogen and oxygen atoms are

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

40  matter present within them merely as virtual parts (as are the parts of hydrogen and oxygen atoms and their parts, viz., electrons, neutrons, protons, quarks, neutrinos, etc.). This distinction between actual and virtual parts allows Oderberg to hold that less complex concrete particular objects (e.g., atoms) maintain at least some kind of presence, viz., by way of their surviving features, within more complex concrete particular objects (e.g., molecules) into which they are “absorbed.” Strictly speaking, however, Oderberg is committed to claims for which we appear to lack independent empirical evidence, e.g., that atoms (or their parts) cease to exist when they enter into chemical bonds with other atoms to form molecules. I take it that, other things being equal, a more straightforwardly hierarchical conception of matter which allows that less complex matter–form compounds can maintain an actual presence within the more com­ plex matter–form compounds they compose is preferable to a non-hierarchical doctrine of prime matter, supplemented with a distinction between actual and virtual parts. Brief Interlude: Metaphysics and Science. Both of the objections just raised against Oderberg’s conception of prime matter rely on the invocation of empirical evidence, which raises difficult questions concerning the relationship between metaphysics and science. A proper discussion of these complex issues would take us too far afield for present purposes, but a few clarifying comments are nevertheless in order. Firstly, one might worry, among other things, that it is dialectically inappropriate for me to appeal to empirical evidence against competing conceptions of matter, unless I also address apparent ways in which my own views might be taken to conflict with contemporary science. A second concern, which arises specifically with respect to (2), is that it is not obvious whether a hierarchical conception of matter is even supported by contemporary science, since (so one might think) such a picture is in fact incompatible with quantum mechanics. Thirdly, when Thomistic prime matter is properly interpreted as a metaphysical postulate, rather than as an empirical claim, then (as in other cases in which competing metaphysical postulates are intended to capture the same set of phenomena) one might expect that this position need not be directly supported by, but only consistent with, empirical evidence from contemporary science. All three of these concerns raise serious and important general questions about how metaphysicians should respond to challenges from contemporary science, and my replies in what follows will unfortunately have to be brief. In response to the first point (viz., the dialectical appropriateness of invoking evidence from contemporary science against competing conceptions of matter), I take it that my own account does in fact fare better than the alternative approaches considered in this chapter in at least the following respects. As will become evident , the hylomorphic conception of matter, in its amended formulation offered later in this chapter, explicitly leaves open (a) whether there is a fundamental level in the compositional hierarchy; and (b) if so, what matter is like at this fundamental level. In this way, my own account incorporates an element of openness and deference to science, familiar from Kripke/Putnam-style accounts of the semantics of natural kind terms, which I find to be absent from the competing conceptions of matter considered here. Moreover, as I argue in Section 2.4, I take the hylomorphic approach to be more readily compatible with a hierarchical conception

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

matter as prime matter  41 of matter. Of course, whether this feature is in fact an advantage of my account over competing approaches depends on how we respond to the second concern, to which I now turn. This next concern calls into question precisely whether a hierarchical conception of matter is even compatible with contemporary science, in particular, quantum mechanics.11 In response to this point, I want to make clear that, in my earlier appeal to such a conception of matter, I had in mind not only apparently hierarchical compositional claims which concern the domain of entities posited by a single science, but also those which appear to span across the domains of entities posited by distinct sciences, e.g., physics, chemistry, biology, and so on. There are lively and ongoing debates over whether the special sciences should be recognized as autonomous or whether the entities posited by them should be regarded as, in some way, reducible to those invoked by fundamental physics. But these are highly controversial issues with respect to which I would like to remain as neutral as possible for present purposes. I take it, though, that even those who do adopt a hard-line reductionist approach will need some account of why it seems to be the case that less complex entities compose more complex entities not only within a particular scientific domain, but also across such domains. Such apparently hierarchical compositional claims surface even within the domain of fundamental physics, as when, for example, atoms are said to be composed of electrons, protons, and neutrons. Finally, according to the third concern raised above, the Thomistic prime matter hypothesis, when properly interpreted as a metaphysical postulate rather than as an empirical claim, should only be expected to be consistent with, but not directly supported by, empirical evidence from contemporary science. When the prime matter hypothesis is regarded in this light, the consistency requirement demands, at the very least, that the Thomistic prime matter hypothesis be formulated in such a way that it can be evaluated for possible conflicts with scientific postulates. In my earlier comments, however, I voiced some reservations as to whether, in its present formulation, the consistency requirement could even be enforced, given the characteristics Oderberg ascribes to Thomistic prime matter. Moreover, some of Oderberg’s remarks certainly suggest that he has in mind a stronger reading of the prime matter hypothesis than mere consistency with scientific evidence: otherwise, it is difficult to see how Thomistic metaphysics and fundamental physics could somehow converge on the very same ultimate substratum by different routes, as is suggested by the possible identification of prime matter with energy.

11   I should note here that I am not currently concerned with the more general question of how an Aristotelian substance ontology might be compatible with contemporary physics. (But see, for example, Ismael and Schaffer (2016) for recent work on quantum holism, and Tahko (2017) on the compatibility of an Aristotelian substance ontology with quantum holism; in a similar vein, I discuss criticisms raised by Simons (1998) concerning the compatibility of an Aristotelian substance ontology with contemporary ­science in Koslicki (2015b).) This more general worry would affect hylomorphists of all stripes to the same extent (including those, such as Oderberg and Brower, who favor a Thomistic conception of prime matter), since all of us are in some way committed to the reality of macroscopic concrete particular objects.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

42  matter (3)  Threat of Contradictions. In addition to the empirical observations discussed in (1) and (2), Oderberg’s conception of prime matter also appears to be subject to many of the same difficulties as the notion of a bare particular. Specifically, prime matter is said to lack an essence, and yet various characterizations are attributed to prime matter which do not appear to be merely accidental to it. For example, Oderberg describes prime matter as something that is capable of receiving any substantial form whatsoever, and as that which is conserved throughout substantial change. But this characterization, together with the other conditions cited earlier, does seem to amount to at least a partial positive description of what prime matter is in itself, i.e., essentially: given the theoretical role played by prime matter, it cannot be assumed that it is merely accidental to prime matter that it has the capacity to persist through substantial change or the capacity to receive any substantial form whatsoever. Thus, the attempt to characterize prime matter in positive or negative terms, while at the same time denying that any such characterization applies to prime matter in its own right, appears to give rise to contradictions.12 (4)  Other Possibilities. Next, it seems that other possibilities for what might underlie substantial change even at micro-physical levels have as of yet not been ruled out and therefore should be considered. These alternative possibilities, moreover, do not lead to the conclusion that the compositional hierarchy bottoms out with a single type of ultimate substratum which meets the characteristics Oderberg ascribes to prime matter. Firstly, we might countenance the possibility of infinitely descending chains consisting of ever smaller matter–form compounds whose material parts are themselves matter–form compounds. Secondly, we might eventually reach a point at which matter–form compounds are composed of material parts which are not them­ selves matter–form compounds.13 At this point, we have reached a level in the compositional hierarchy to which the doctrine of hylomorphism no longer applies. The second type of possibility might be realized, for example, in a scenario in which the material parts composing matter–form compounds are (ii.a) simples, i.e., material objects that do not have any material parts; or (ii.b) portions of stuff. (Both of these options will be discussed in more detail later in this chapter.) What matters most for present purposes 12   For Oderberg, prime matter is not a thing and hence not of the right ontological category to have an essence. He does, however, allow that other entities besides matter–form compounds have essences as well (e.g., accidents which inhere in matter–form compounds). Moreover, in order to perform the roles ascribed to it, prime matter has to be at least sufficiently “thing-like” to be able to persist through substantial change: it must be possible, for example, for the same prime matter which was around before a quark, say, came into existence still to be around after the quark has come into existence. Given the roles ascribed to prime matter and the relatively broad application of the notion of essence beyond the category of matter–form compounds, it strikes me as ad hoc not to recognize the characterizations ascribed to prime matter as amounting to at least a partial specification of its essence. 13   Different versions of hylomorphism will describe the conditions which would have to obtain in order for the second possibility to be realized differently. According to my own view, as developed in Koslicki (2008a), this possibility would obtain in a scenario in which a structured whole is composed of material parts and these material parts are not themselves structured wholes, i.e., they are not themselves composed of any further material parts which must meet the type constraints and configuration constraints imposed by the kind to which the whole they compose belongs.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

matter as prime matter  43 is that none of these possibilities, as far as I can see, has been explicitly excluded either on empirical or on conceptual grounds, by the considerations Oderberg brings to the table; and none of them requires embracing his doctrine of prime matter. (5)  Non-particularity. As outlined in Section 1.6, hylomorphists must confront the question of whether, and how, their central piece of apparatus, viz., the distinction between matter and form, can be used to help resolve questions concerning the numerical (synchronic, diachronic, and cross-world) identity and distinctness of matter–form compounds. To illustrate, these include questions of the form: what, if anything, makes Socrates the very concrete particular object that he is at each time at which he exists? What, if anything, makes Socrates at a time, t1, numerically identical to Socrates at a distinct time, t2? What, if anything, makes Socrates in a world, w1, numerically identical to Socrates in a distinct world, w2? What, if anything, makes Socrates and Callias numerically distinct at a single time at which they both exist? Contrary to what is often assumed by Thomists, Oderberg does not take prime matter to be a principle of individuation for matter–form compounds: [. . .] although prime matter is one of the basic constituents of material substances, it is—­ perhaps surprisingly, one might think, given hylemorphism—not the principle of individuation. The reasons are: (1) it is common, i.e. multiply instantiable (wherever there is actuation by a substantial form), and it is a hallmark of individuality, including that of material substances, that it is, to use the traditional term, incommunicable. [. . .] (2) Prime matter is indivisible, being mere potentiality, so it cannot serve as the basis of the division of a species or nature into individuals. We cannot say “Here is some prime matter, and there is some more,” but we can say, “Here is Socrates, and there is Callias,” or in other words “Here is prime matter informed by the nature of Socrates, and there is prime matter informed by the nature of Callias.” (We can call these ‘Socrateity’ and ‘Calliaeity’, but must not confuse them with haecceities or individual essences as postulated by Duns Scotus; these I reject as at least unwarranted. Socrateity is just general human nature as particularized in Socrates. The particularizing is done by matter, not by “thisness.”)  (Oderberg (2007), p. 109)

Given Oderberg’s conception of prime matter, it cannot play any role in his theory in  settling questions concerning the numerical identity or distinctness of matter– form compounds. We cannot say, for example, that Socrates owes his synchronic, diachronic, or cross-world identity to the prime matter which composes him at any given time, across times, or across worlds, since Socrates’ prime matter cannot be identified independently of Socrates. For similar reasons, the fact that Socrates and Callias are numerically distinct at a single time cannot be traced to the numerical distinctness of the prime matter which composes them at that time.14 14   An additional puzzle concerning individuation which arises for Oderberg’s approach is that it becomes difficult to understand how something non-particular (viz., prime matter) together with something else that is also non-particular (viz., substantial form) can result in something that is particular (viz., a particular matter–form compound). Since the prime matter composing Socrates cannot be identified independently of Socrates, and Socrates’ substantial form is just “general human nature particularized in Socrates,” one wonders why prime matter as combined with general human nature on one occasion results in Socrates,

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

44  matter (6)  Material Constitution, Unity, and the Grounding Problem. The Problem of Material Constitution challenges hylomorphists to explain how a matter–form compound can be so similar to, and yet nevertheless numerically distinct from, its constituting matter. The Grounding Problem poses a challenge in particular for those who opt for a coincidence– theoretic response to the Problem of Material Constitution, since these philosophers then face the question of what grounds the apparent modal differences between numerically distinct but spatiotemporally coincident objects. Finally, the Problem of Unity raises the question of how a plurality of parts can give rise to a unified whole. The following remarks suggest that Oderberg might be sympathetic towards an eliminativist solution to all three of these problems: Another way of putting the point is to say that substantial form permeates the entirety of the substance that possesses it, not merely horizontally in its parts—there is as much dogginess in Fido’s nose and tail as in Fido as a whole—but also vertically, down to the very chemical elements that constitute Fido’s living flesh. To use the traditional Scholastic terminology, the chemical elements exist virtually in Fido, not as compounds in their own right but as elements fully harnessed to the operations of the organism in which they exist, via the compounds they constitute and the further compounds the latter constitute, through levels of compounds—DNA, the proteins coded for by that DNA, the organelles that make up the cells, the organs made up of the cells, and so on.  (Oderberg (2007), pp. 70–1)

According to the proposal outlined here, both the Problem of Material Constitution and the Grounding Problem simply go away, since each region of space-time (including its subregions) can be occupied by at most one matter–form compound. The region of space-time that is occupied by the dog Fido, for example, does not contain an additional matter–form compound, viz., Fido’s body, that is numerically distinct from, but spatiotemporally coincident with, the dog it constitutes. Rather, only a single substantial form, viz., that associated with the dog Fido, brings with it a single (non-derivative) essence, which “permeates” the whole region of space-time that is occupied by Fido and all of Fido’s actual parts (e.g., Fido’s nose and tail). All other “parts” (viz., the cells, proteins, organs, etc.) maintain a merely virtual presence within Fido and hence do not require the recognition of a separate substantial form or (non-derivative) essence within the region of space-time (or its subregions) that is occupied by Fido. The proposal at hand furthermore generates a response to the Problem of Unity: since none of Fido’s parts (actual or virtual) are regarded as matter–form compounds in their own right, no account is needed for how a plurality of matter–form compounds, each with its own substantial form, could ever give rise to a single unified whole. The eliminativist solution to the three problems just outlined requires the acceptance of the following two positions which Oderberg endorses: unitarianism about and on another occasion results in Callias. The following remark suggests that Oderberg takes these facts not to be open to further explanation: “ . . . [E]very material substance is the principle of its individuation by its own proper entity. It is the very union of prime matter to substantial form that constitutes the individual substance . . . ” (Oderberg (2007), p. 111).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

matter as stuff  45 substantial forms, according to which only a single substantial form is ever present in any region of space-time (including its subregions) that is occupied by a matter–form compound; and the doctrine of virtual parts, according to which matter–form compounds (e.g., atoms) cannot survive the process of becoming “absorbed” into other matter–form compounds (e.g., molecules). It already emerged above in connection with (2), however, that the doctrine of virtual parts appears to conflict with our current empirically supported hierarchical conception of matter. Once the doctrine of virtual parts is abandoned, any motivation to accept unitarianism about substantial forms disappears as well. Assuming that an eliminativist solution along the lines indicated here is not available, the difficulties in question will have to be confronted in some other way. As I will argue in what follows, however, this price is worth paying, since alternative responses to these challenges are available and the objections we have identified in this section in connection with Oderberg’s doctrine of prime matter can thereby be avoided.

2.3  Matter as Stuff 2.3.1  Thomistic Prime Matter as Stuff I now turn to an examination of (i.b), the second conception of matter mentioned in the beginning of this chapter, according to which the matter composing a hylomorphic compound should be construed as stuff. As a representative formulation of this approach, I consider the approach to matter developed by Jeffrey Brower, in his recent exposition of Aquinas’ hylomorphism (see Brower (2014)). Brower uses as a backdrop for his interpretation of the Thomistic doctrine of prime matter the category of stuff as described by Ned Markosian (see Markosian (1998, 2004a, 2004b, 2015)). Although Brower does not take the two conceptions of matter to coincide in every respect, he nevertheless finds Markosian’s notion of stuff to be a useful interpretive tool with which to approach Aquinas’ doctrine of prime matter. Thomistic prime matter, according to Brower’s characterization, has the following five main characteristics. Firstly, prime matter is a type of being which can be compounded and divided. Thus, portions of prime matter can enter into part–whole relations with other portions of prime matter; the portions and subportions into which prime matter can be divided are its parts (Brower (2014), pp. 10–11, 115). In addition, portions of prime matter, along with substantial forms, constitute matter–form compounds; but the relation of constitution or hylomorphic composition here is a distinct relation from that of mereological composition (i.e., the relation between a whole and its parts), although Brower (following Aquinas) uses the language of parthood and composition to describe both relations. For example, the prime matter associated with a given concrete particular object, e.g., a bronze sphere, is a certain portion of prime matter which partly constitutes (i.e., hylomorphically composes) the bronze sphere in question; and

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

46  matter this portion of prime matter is itself (mereologically) composed of smaller portions of prime matter, which are (mereological) subportions of the original portion of prime matter.15 Every portion or subportion of prime matter is a part of all the prime matter that exists in the world. Secondly, portions of Thomistic prime matter, according to Brower’s reading, can be identified and reidentified over time as well as distinguished from other portions of prime matter, independently of the matter–form compounds with which they are associated throughout their careers. For example, in Brower’s view, we can make sense of claims of the form “a’s prime matter at t1 is identical to b’s prime matter at t2,” where a and b are matter–form compounds which may or may not be numerically distinct.16 Brower argues that the possibility of such reidentifications follows directly from the role that Aquinas assigns to prime matter in substantial change: this role requires us to be able to say, for example, that the portion of prime matter composing Socrates immediately after he comes into existence is identical to the portion of prime matter composing the zygote from which Socrates is generated, even if the zygote in question is taken to be numerically distinct from Socrates (Brower (2014), pp. 20, 103–5, 113–14). Thirdly, Brower interprets Aquinas as being committed to unrestricted mereological composition for portions of prime matter (Brower (2014), pp. 116–17). Thus, any plurality of portions of prime matter itself composes a portion of prime matter. Brower proposes to make sense of Aquinas’ puzzling claim that “prime matter is said to be numerically one in all things” (De Principiis Naturae, 2.97–108) by interpreting him as making reference to the mereological sum of all portions of prime matter in the ­universe. Brower construes Aquinas’ remark concerning the numerical oneness of prime matter in all things as a claim about composition, rather than identity: “ . . . that there exists a single sum or portion of prime matter composed of all the distinct, smaller portions existing in the world” (Brower (2014), p. 117).17 Fourthly, Thomistic prime matter, in Brower’s view, is infinitely divisible as well as atomless and hence comparable to what contemporary metaphysicians call “atomless gunk” (Brower (2014), p. 118). There are no restrictions concerning the ways in which portions of prime matter can be divided into subportions of prime matter, 15   For Brower’s distinction between constitution, i.e., hylomorphic composition, and mereological composition, see Brower (2014), p. 6, n. 9; p. 10, n. 20; p. 11, n. 21; as well as Sections 5.4 and 8.1. 16   In this respect, Brower’s interpretation of the Thomistic doctrine of prime matter contrasts starkly with that developed by Oderberg, as emerged in our discussion of (5), “Non-Particularity.” According to Oderberg, the prime matter composing, say, Socrates cannot be identified independently of Socrates (see Oderberg (2007), p. 109). 17   The assumption that there is a single such all-encompassing mereological sum composed of all the other portions of prime matter in the universe also requires a commitment to the uniqueness of mereological composition, which Brower accepts for Thomistic prime matter as well. Since in addition he takes parthood for portions of prime matter to be transitive, Thomistic prime matter, according to Brower’s reading, satisfies all three of the axioms that are associated with the system known as “standard mereology” or “classical extensional mereology” (see Leonard and Goodman (1940); Lewis (1991); Simons (1987)): unrestricted composition, uniqueness of composition, and the transitivity of parthood.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

matter as stuff  47 and, given Aquinas’ rejection of atomism, no portions of prime matter are composed of indivisible parts.18 Finally, when Aquinas says of prime matter that it is “a being which exists only in pure potentiality” (Brower (2014), p. 119), Brower interprets him as having in mind the non-individuality of prime matter: portions of prime matter, in Brower’s view, are best construed not as things, but as portions of stuff (Brower (2014) pp. 120–2). Thus, when Aquinas describes prime matter as a type of being which lacks actuality, Brower reads him not as saying anything against the reality or existence of prime matter; nor, according to Brower, does Aquinas mean to be denying that prime matter has a distinctive character or nature of its own.19 Rather, the best way to think of Thomistic prime matter, so Brower argues, is along the lines of the category of stuff, particularly as it has been described in the recent literature by Ned Markosian (Brower (2014), pp. 125ff).20 Markosian’s notion of stuff is similar to Thomistic prime matter, according to Brower’s reading, in the following five main respects (Brower (2014), p. 126). Firstly, like Thomistic prime matter, stuff, in Markosian’s view, comes in only one fundamental type for all material things.21 Secondly, stuff is divisible into portions, which can enter into part–whole relations with one another. Thirdly, the domain of stuff satisfies the axioms of classical mereology (see n. 17). Fourthly, Markosian denies that portions of stuff can enter into part–whole relations with things; rather, portions of stuff constitute things, where constitution here is to be understood as a primitive sui generis relation distinct from the part–whole relations governing portions of stuff. Fifth, portions of stuff cannot exist apart from the things which are constituted by them.22 And, finally, 18   But Brower’s interpretation requires only that Thomistic prime matter is potentially infinitely divisible and hence does not conflict with the Aristotelian prohibition against actual infinities (see Brower (2014), p. 118, n. 17). 19   Though when the term, “essence” is used in its strict sense, then, so Brower emphasizes, Aquinas would want to deny that prime matter has an essence (Brower (2014), p. 20). In this respect, Brower’s and Oderberg’s readings of the Thomistic doctrine of prime matter agree. 20   Other theorists besides Markosian have also argued for the acceptance of an irreducible ontological category of stuff (see, for example, Jubien (1993); Kleinschmidt (2007); Laycock (2006)). In what follows, I focus on Markosian’s defense in particular. For one thing, Markosian’s conception of stuff is the most ­pertinent for present purposes, since his approach underwrites Brower’s interpretation of the Thomistic doctrine of prime matter. Secondly, to my knowledge, Markosian’s case in favor of accepting stuff as an irreducible category is the most detailed and explicit defense and characterization of this position in the literature. For critical discussions of Jubien (1993) and Laycock (2006), however, see Sider (1999a) and Koslicki (2007a), respectively. 21   According to Markosian, stuff is not divided up into distinct subcategories which have any ontological significance, as it would be according to an Anaxagorean picture, just as, in Markosian’s view, there are also no ontologically significant subdivisions among things. Rather, the category of stuff contains only one kind of stuff, “generic stuff,” and all the more familiar “stuff-kinds” (e.g., water, bronze, etc.) turn out to be generic stuff arranged in a certain manner. (We will return to these issues in Section 2.3.3.) 22   According to Markosian’s “Principle of Constitution” (PC), whenever there is a thing, then there is also a portion of stuff which constitutes the thing in question; and, by Leibniz’s Law, the portion of stuff and the thing it constitutes are never identical. In addition, Markosian holds a kind of qualified reverse version of PC as well, which rules out what we might call “free-floating” stuff. The principle in question can be stated roughly as follows: “Every portion of stuff can be partitioned into non-overlapping subportions, such that each of these subportions of stuff constitutes a thing” (see, e.g., Markosian (2004b), p. 409). In particular, Markosian’s conception of simples, which will be discussed in more detail in Section 2.3.3, leads him to

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

48  matter since portions of stuff can be identified and distinguished from other portions of stuff, they are countable, and not merely measurable. In other respects, however, the category of stuff, as characterized by Markosian, behaves differently from Thomistic prime matter, according to Brower’s reading. For one thing, Brower takes Markosian’s stuff to be ultimately composed of point-sized portions of stuff, i.e., simple indivisible but non-individual “atoms” of stuff.23 For another, as we will see shortly, the motivations behind Markosian’s recognition of a distinct ontological category of stuff are, not surprisingly, quite different from those which guide Brower’s interpretation of Thomistic prime matter.

2.3.2  Difficulties for Brower’s Thomistic Prime Matter as Stuff Hypothesis Among the main advantages Brower claims for his approach is its ability to yield an account both of substantial change and of the distinction between substantial and nonsubstantial change (Brower (2014), pp. 174–83). In addition, Brower’s Thomistic hylomorphism generates solutions to the Problem of Material Constitution, the Grounding Problem, and the Problem of Unity (Brower (2014), pp. 165–74).24 In contrast to Oderberg’s approach, however, Brower’s construal of the Thomistic doctrine of prime matter also promises to be of help in settling at least some of the difficult questions concerning the numerical identity and distinctness of matter–form compounds, since adopt a specific version of this reverse-constitution principle, according to which every portion of stuff which occupies a maximally continuous region of space constitutes a simple (viz., an object which has no proper parts). 23   It is not entirely clear, however, whether Brower is correct in ascribing this last commitment to Markosian. In his earlier work, Markosian claims that it is an advantage of his conception of simples (according to which simples may be spatially extended objects which occupy maximally continuous regions of space) that it is nevertheless compatible with a rival conception of simples (according to which they are non-extended pointy objects) as well as with the possibility of atomless gunk. According to Markosian, if any of these alternative possibilities obtained, it would do so as a matter of contingent fact (see Markosian (1998), pp. 215, 227). Given these earlier views, we should thus expect Markosian similarly to want to leave it open whether stuff is gunky or non-gunky. See Markosian (2015), p. 678, for similar remarks concerning neutrality which suggest that he still agrees with his earlier self on this point. 24  According to Brower’s exposition, two different types of solutions to the Problem of Material Constitution are open to the Thomistic hylomorphist. The first solution follows the eliminativist line sketched earlier and therefore requires a commitment to both unitarianism about substantial forms and the doctrine of virtual parts (Brower (2014), p. 169, n. 31). (In Brower’s terminology, virtual parts are referred to as “integral” or “quantitative” parts; see, for example, pp. 15–16.) The second solution proceeds by drawing a distinction between sameness and numerical identity (pp. 168–9). (See also the solution embraced in Rea (1998) cited in Section 1.6.) According to this second solution, a statue and the clay of which it is made, for example, are regarded as the same material object, without thereby also being identified. Brower’s distinction between sameness and numerical identity raises many difficult questions of its own. In addition, however, the second solution to the Problem of Material Constitution does not by itself yield a solution to the Grounding Problem and the Problem of Unity. For suppose that the clay counts as a material object that is the same as, but numerically distinct from, the statue. In this case, there is still a further question as to whether, first, the clay and the statue have distinct modal profiles and, if so, what grounds these differences. Secondly, it still remains to be explained how the many statue parts can together manage to give rise to a unified thing, viz., the statue. Thus, unless some other strategy can be found, it would seem to be more methodologically economical for the Thomistic hylomorphist to embrace the eliminativist solution, since it solves all three problems at once.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

matter as stuff  49 Brower takes portions of prime matter to be identifiable and reidentifiable independently of the matter–form compounds with which they are associated. Brower’s Thomistic doctrine of prime matter thus speaks directly to the following sets of desiderata identified in Section 1.6: (2) “Causation, Explanation, Change”; (4) “Identity and Indiscernibility”; (5) “Material Constitution”; (7) “Unity,” and (8) “The Grounding Problem.”25 From the point of view of the considerations raised in Section 2.2.3, then, the main question which arises when comparing the conceptions of matter proposed by Oderberg and Brower is how well the latter fares when it comes to the fifth objection raised earlier, which asks to what extent prime matter can help resolve difficult questions concerning the numerical identity or distinctness of matter–form compounds. In Brower’s view, questions concerning the numerical identity and distinctness of matter–form compounds (and their forms) can be settled derivatively by appeal to the portions of prime matter which hylomorphically compose them. But in order for portions of prime matter to be able to play this role, so Brower reasons, they in turn must have their identity and distinctness primitively: Indeed, on [Aquinas’] view we might say that, whereas the portions of prime matter associated with distinct material substances are primitively distinct, the substances themselves, as well as their substantial forms, are derivatively distinct—that is to say, distinct solely in virtue of their relationship to the primitively distinct portions of prime matter that serve as their substrata or individuators.  (Brower (2017), p. 141)

Thus, in contrast to Oderberg, Brower assumes not only that it makes sense to speak of the identity and distinctness of portions of prime matter at a particular time or over time independently of the matter–form compounds with which they are associated at any given time, but, in addition to that, he takes facts about the numerical identity and distinctness of portions of prime matter to be the primitive source from which matter–form compounds (and their forms) derive their identity and distinctness. We might thus describe portions of prime matter, given their role in Brower’s Thomistic ontology, as themselves particulars or individuals in their own right, although Brower himself does not characterize them in this way.26 Importantly, though, if portions of prime matter are recognized as particulars or individuals in their own right, the ontological category to which they belong must be sharply distinguished from that occupied by other entities which count as particulars or individuals within Brower’s Thomistic ontology. For unless portions of prime matter are awarded a status that is 25   Other aspects of Brower’s approach, which will not concern us here, are designed to address the remaining sets of desiderata: (1) “Property Possession and the Problem of Universals”; (3) “Essence and Accident”; and (6) “Mereological Composition.” 26   Although I am using the terms, “particular” and “individual,” interchangeably, Brower importantly does not. A “particular,” according to Brower, is “a being that is both subsistent and individual” (Brower (2014), p. 23); “subsistence” here contrasts with “inherence,” where “inherence is best understood as a type of dependency built into the nature of forms and properties” (p. 8). A being is an “individual” if it satisfies two conditions: first, “it must have actuality through itself; and second, it must belong to some natural kind or species. [. . .] Only forms, substances, and compounds have actuality through themselves” (pp. 19–20).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

50  matter irreducibly distinct from that of other particular or individual entities, two commitments which, on Brower’s reading, are absolutely central to Aquinas’ metaphysics would be threatened: the ability to distinguish between substantial and accidental unity; and the ability to distinguish between substantial and non-substantial change. To see why, suppose for a moment that both Socrates and Socrates’ matter are ­particulars of the same type. In that case, so Brower reasons, Socrates would lose his status as a “basic particular” in Aquinas’ system and would instead be demoted to an “accidental unity,” i.e., a “non-basic particular” which is composed of other particulars (Brower (2014), pp. 24–6).27 Secondly, if Socrates and Socrates’ matter were both particulars of the same type, then one would expect Socrates’ substantial form to bear the same relationship to both Socrates and Socrates’ matter (see especially Brower (2014), Sections 4.1, 5.5, 7.2). But Socrates’ substantial form does not bear the same relationship to both Socrates and Socrates’ matter: Socrates is essentially human, but Socrates’ matter is not human at all, neither essentially (since portions of prime matter function as the substratum for substantial change), nor accidentally (since nothing is ever accidentally human). To capture these different ways in which a substantial form or property can be related to an entity, Brower introduces a distinction between mere “possession” and “characterization”: while Socrates’ matter “possesses” but is not “characterized by” Socrates’ substantial form, Socrates not only “possesses” but is also “characterized by” Socrates’ substantial form (see especially Brower (2014), pp. 76–8). The best explanation for this difference in how Socrates’ substantial form is related to Socrates and Socrates’ matter, in Brower’s view, is that Socrates and Socrates’ matter belong to two irreducibly distinct ontological categories, only one of which contains entities of the right type to be “characterized by” Socrates’ substantial form. Socrates, as a matter–form compound, is of the right type to be a human being, i.e., to be “characterized by” the substantial form of being human, while Socrates’ matter, as a portion of prime matter or stuff, is not of the right ontological type to be a human being, i.e., to be “characterized by” the substantial form of being human (see especially p. 127). As the preceding remarks indicate, Brower’s account treats portions of prime matter as the primitive source by means of which questions concerning the numerical identity and distinctness of matter–form compounds are to be settled. But Brower’s ability to do so comes at a steep price: it is methodologically and ontologically costly in that it requires a commitment to an irreducibly distinct ontological category of stuff, occupied by portions of prime matter, accompanied by a distinction between two different ways in which substantial forms or properties may be exhibited by entities (viz., mere “possession” vs. “characterization”). Other things being equal, then, if the same range of phenomena can be explained by an account which carries fewer ontological and methodological commitments than Brower’s, such an alternative account should be 27   A “basic particular,” according to Brower, is “[a] particular that not itself composed of any particulars,” while a “non-basic particular” is “[a] particular that is composed of at least one particular” (Brower (2014), p. 24).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

matter as stuff  51 preferred on grounds of parsimony. We will see in Section 2.4, that the hylomorphic conception of matter is in fact more economical in this respect, since it matches or exceeds the explanatory power of its competitors without having to take on board an irreducible thing/stuff distinction. In addition, as we will discuss in detail in Chapter 3 (see Section 3.4.3), a theory like Brower’s which treats matter as the primitively ­individuated source of facts concerning the numerical identity and distinctness of matter–form compounds is in fact not able to match the explanatory of a theory which locates the source of these facts elsewhere. In the end, then, I will conclude that my own package of views is preferable to those of my competitors on grounds of both economy and explanatory power. At this point, however, until this stronger position can be properly justified, we are licensed to conclude only that, other things being equal, a conception of matter which does not require a commitment to an irreducible ontological category of stuff, along with its accompanying apparatus of relations, should be preferred over its competitors on grounds of ontological and methodological parsimony. Given the particular historical setting in which Brower’s account is embedded, we may wonder, in addition, whether an irreducible thing/stuff distinction can be motivated by appeal to considerations not related to Brower’s particular interpretive goals of offering a plausible reading of Aquinas. To this end, I now turn to an examination of the conception of stuff developed by Ned Markosian.

2.3.3  Difficulties for Markosian’s Matter as Stuff Hypothesis Markosian (2015) cites a variety of reasons in support of his “mixed” or “dualist” ontology, which includes distinct and irreducible ontological categories for both stuff and things.28 Some of Markosian’s reasons derive from what he would classify as pretheoretic intuition; others, in contrast, are based on philosophical analysis. Markosian’s endorsement of stuff as an irreducible ontological category is not supported by empirical considerations. In fact, as will emerge in what follows, it is doubtful whether we could find out anything about stuff, as Markosian conceives of it, by consulting a chemist or a physicist. In what follows, I argue that, since Markosian’s pre-theoretic or philosophical arguments fail to create sufficient pressure to take on board an irreducible category of stuff, and his position is furthermore not corroborated by empirical evidence, his case for a mixed ontology, as it stands, can and should be resisted. In Markosian’s view, our common-sense way of talking and thinking supports a mixed ontology. In fact, judging from his examples, Markosian is committed to the claim that we typically make reference to stuff when (i) we either use nouns in their mass-occurrences (e.g., “more ice-cream,” “two liters of water,” etc.) or when (ii) the expression, “stuff ” (or “matter”), explicitly occurs in a statement we utter (as when we use phrases like “the stuff,” “this stuff,” “some stuff,” and so on). I assume he would accept 28   For ease of exposition, I will, for the remainder of this section, move back and forth freely between the terminology Markosian uses to state his own views (e.g., “stuff ” and “thing”) and the terminology hylomorphists of different stripes would use to express Markosian’s views (e.g., “matter” or “prime matter” and “matter–form compound”).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

52  matter an analogous claim for reference to things, viz., that such reference typically takes place when (i) we either use nouns in their count-occurrences (e.g., “many chairs,” “two bicycles,” etc.) or when (ii) the expression, “thing” (or “object”), explicitly occurs in a statement we utter. But there are good reasons for thinking that our pre-theoretic commitments and our ordinary ways of talking are not philosophically fine-grained and stable enough to support the philosophical weight Markosian wishes to place on them. In particular, if we want to avoid being thrown into a Whorfian relativism, according to which different language communities conceptualize the world in radically different and possibly incommensurable ways, we must give up the hope of finding a direct link between the mass/count distinction and any single metaphysical distinction, such as Markosian’s distinction between things and stuff.29 Firstly, given ordinary usage, our terms, “stuff ” and “things,” are often interchangeable without any significant distortion in meaning, as in “the stuff/things you’ve written,” “the things/stuff in your attic,” and so on. (Similarly for related terms, such as “matter” and “object.”) Secondly, the classes of nouns under discussion are strikingly heterogeneous. For example, “asparagus,” “traffic,” “jewelry,” and “poetry” are like “mud” in that they are standardly used primarily in mass occurrences; “cloud,” “virus,” “hurricane,” and “mistake” are like “tree” in that they are standardly used primarily in count occurrences; “carrot,” “pain,” and “justification” are like “hair” in that they are standardly used in both ways. Thirdly, actual language use is  extremely flexible in permitting existing nouns to acquire new uses (“A BMW 300-series is not much car for the money”; “I just sent you another email”). Fourthly, there is significant cross-linguistic variation in how or even whether different languages mark the mass/count distinction grammatically at all. For example, the English noun, “hair,” standardly has both mass and count occurrences, while the German noun, “Haar,” standardly has count occurrences and has a mass use only in poetic contexts (as in “Rapunzel, Rapunzel, lass mir Dein Haar herunter”). Furthermore, Asian classifier languages, such as Mandarin Chinese and Japanese, have been thought to represent all nouns as having mass occurrences, since all nouns in these languages must be preceded by a classifier phrase, as in “a long/thin unit of pencil/cigarette/ . . . ” . I take these and other similar observations to indicate that language and our use of it are much too tolerant and varied to take either occurrences of the terms, “stuff ” (“matter,” etc.) and “thing” (“object,” etc.), or the mass/count distinction in general as a reliable indicator of a deep ontological distinction of the sort Markosian is trying to motivate. As a result, the best way to proceed, for the purposes of the present discussion, is to leave aside our ordinary ways of talking and thinking, and to focus instead on those of Markosian’s arguments which derive from philosophical analysis. Suppose for a moment, however, that Markosian was right in thinking that our ordinary ways of talking and thinking support a thing/stuff distinction. This in itself would establish very little of philosophical significance, for we would still need to   I have discussed the mass/count distinction in more detail in Koslicki (1995, 1997, 1999a, and 1999b).

29

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

matter as stuff  53 know whether we are correct in being pre-theoretically committed to such a distinction. Since we are (or were at one time) pre-theoretically committed to all sorts of bizarre entities and distinctions, a special story would be required to establish that this ­particular piece of alleged folklore should be taken so seriously by ontologists. As will emerge below, however, it is not at all obvious, given Markosian’s account, how we would be able to confirm whether a thing/stuff distinction, which he sees as being part of our common-sense ontology, is on the right track. I turn next to Markosian’s philosophical reasons in favor of accepting a thing/stuff distinction, viz., in particular those connected to his argument from the possibility of spatially extended simples.30, 31 Markosian considers and rejects a rival view of simples he calls “The Pointy View of Simples” (PVS): according to this approach, an object is a simple just in case it occupies exactly one point in space (Markosian (1998), pp. 216ff). Markosian accepts the right-to-left half of PVS, viz., that if something is a pointy object, then it is a simple; but he does not accept the left-to-right half of PVS, viz., that if something is a simple, then it is an extensionless point-sized object. Markosian’s main argument against PVS is based on the following scenario: Imagine a possible world in which there is only one physical object, a perfectly solid sphere made of some homogeneous substance, floating in otherwise empty space. If you can imagine such a world—and I think you can—then the Pointy View of Simples is false. For the Pointy View of Simples entails that any extended object that occupies a continuous region of space must be composed of an infinite number of parts.  (Markosian (1998), p. 218)

According to PVS, any world which contains at least one spatially extended physical object in fact contains an infinite number of objects, viz., the simples which occupy the extensionless points included in the region of space in question. This entailment strikes Markosian as unacceptable, since he believes that, in the scenario specified above, only a single physical object exists and this physical object is spatially extended (Markosian (1998), p. 219).32 30   Markosian (2015) cites nine philosophical reasons for believing in stuff as an irreducible category: some of them are reasons he himself finds convincing, while others are associated with philosophical positions Markosian rejects. In what follows, I focus on Markosian’s argument from the possibility of spatially extended simples, since this argument, if it were successful, would in fact provide independent reasons in favor of accepting stuff as a distinct and irreducible ontological category. The remaining philosophical considerations to which Markosian appeals and to which he himself is sympathetic (viz., those concerning constitution and mereology) are only applicable once we have already been persuaded to take on board a category of stuff which is irreducibly different from the category of things. Unless we feel inclined to follow Markosian in adopting a mixed ontology, it is still a live question whether Markosian’s stipulated use for the term, “constitution,” as well as the axioms of mereology he proposes for the domain of stuff, in fact, apply to anything. 31   For a more detailed examination of Markosian’s approach to simples than what I can provide here, see McDaniel (2003); responses to McDaniel’s objections can be found in Markosian (2004a); a further development of Markosian’s position concerning simples is given in Markosian (2004b). 32   In addition to PVS, Markosian also considers and rejects a further alternative conception of simples, according to which simples are objects which are either physically or metaphysically indivisible. Markosian finds this rival approach implausible for the following reasons: an object might be physically indivisible without being simple (e.g., a chain which consists of links which are in fact physically unbreakable); and

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

54  matter According to the approach Markosian favors, “MaxCon,” simples are maximally continuous objects, i.e., objects which occupy maximally continuous regions of space.33 The intuitive idea behind MaxCon, as described by Markosian, is that “simples are objects that occupy the largest matter-filled, continuous regions of space around” (Markosian (1998), p. 222). MaxCon allows for a description of the scenario cited above as one which contains only a single spatially extended physical object. By MaxCon’s lights, the perfectly solid homogeneous sphere we were asked to imagine would count as a simple, provided that it occupies a maximally continuous region of space. In addition, however, given Markosian’s views, it is also revealed to us as a result of philosophical analysis that, when we were asked to imagine a world which contains a single perfectly solid homogeneous sphere floating in otherwise empty space, we were in fact imagining a world which contains a single sphere and the stuff that constitutes the sphere and the parts (i.e., the subportions) of the stuff that constitutes the sphere. Thus, while it is true to say, according to Markosian’s account, that the scenario in question contains only a single spatially extended physical object, it is not true to say that it contains nothing but a single spatially extended physical object. For Markosian is also committed to holding that the world in question contains additional physical content besides the single sphere we thought we were imagining; only the additional physical content in question belongs to the category of stuff, rather than to the category of things.34 Markosian’s description of the scenario in question thus not only relies heavily on certain modal intuitions (viz., that the world we are asked to imagine is in fact possible); it also requires a rather fine-grained interpretation of the modal intuitions which are supposed to be elicited by his thought experiment.35 But now consider the following anything that is physically indivisible might still be metaphysically divisible (e.g., an object which is in fact pointy but might become spatially extended). For these reasons Markosian concludes that the simplesas-indivisible approach ultimately collapses into PVS (see Markosian (1998), p. 221). 33   A maximally continuous object, x, satisfies the following conditions: x is a spatially continuous object and there is no continuous region of space, R, such that (i) the region occupied by x is a proper subset of R, and (ii) every point in R falls within some object or other (Markosian (1998), p. 221). 34   The connection between MaxCon and Markosian’s acceptance of a dualist ontology can be brought out, for example, by considering Markosian’s response to McDaniel’s “Argument from Spatial Intrinsics” (McDaniel (2003), p. 274). Consider an object, “Multi,” which by hypothesis occupies a maximally continuous region of space. Hence, according to MaxCon, Multi is a simple; i.e., it has no proper parts. Suppose further that Multi is blue at region R1 and gold at region R2. Since, by hypothesis, Multi is a simple, it is not open to the proponent of MaxCon to describe Multi as having a part which is blue and a part which is gold. Markosian agrees that we should resist the idea that relations such as being blue and being gold should be indexed to regions of space. Instead, his preferred analysis of the scenario is also the one McDaniel recommends: namely that, even though Multi itself does not have parts, the portion of stuff which constitutes Multi has parts, viz., its subportions; and some of these subportions are blue, while others are gold (Markosian (2004a), pp. 338–40). This response to McDaniel’s objection, however, requires that MaxCon be supplemented with a thing/stuff distinction. 35   Notice also the following further odd feature of Markosian’s description of the scenario in question: given MaxCon, the subportions of the portion of stuff which constitutes the sphere cannot themselves constitute anything, since otherwise the sphere in question would have proper parts and therefore would no longer count as a simple (see McDaniel (2003) for further discussion). In the case of a clay statue, for example, we might feel inclined to say that the nose-shaped piece of clay, which is a subportion of the clay

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

matter as stuff  55 alternative description of the scenario outlined earlier: when we were asked to imagine a world that contains a single sphere, we were in fact imagining a world that includes not only the sphere, but also the proper parts of the sphere. Since the sphere, by hypothesis, is spatially extended, and any spatially extended physical object has proper parts (or so the competing interpretation maintains), we could not help but imagine the sphere along with its proper parts, whether we were aware of it or not. The proper parts of the sphere were not explicitly mentioned in the description of the scenario provided earlier, but neither was the portion of stuff which, if Markosian has his way, constitutes the sphere, or its subportions. While the alternative interpretation of the scenario just provided is incompatible with Markosian’s position, it is compatible with other rival ontologies (e.g., ones that include a commitment to PVS). Given the availability of such alternative interpretations, we should therefore conclude that Markosian’s account, which favors MaxCon, provides merely one, but by no means the only possible, coherent description of the thought experiment in question. Consequently, insofar as MaxCon is intended to provide support for Markosian’s dualist ontology, we need not feel compelled to embrace a distinct and irreducible ontological category of stuff on the basis of the argument for the possibility of spatially extended simples.36 So far, my aim has been to argue that Markosian does not succeed in providing pressing pre-theoretic or philosophical reasons for taking on board a distinct and ­irreducible category of stuff. I want to end my examination of Markosian’s approach by bringing out why I am deeply puzzled by what stuff is really supposed to be, given that constitutes the statue, itself constitutes a thing, viz., the statue nose. To preserve consistency with MaxCon, however, Markosian is forced to deny that such a possibility could ever obtain in the case of spatially extended maximally continuous physical objects. This raises the question of how, in general, constitution “facts” are determined on Markosian’s view: when is it the case that a portion of stuff constitutes a thing and when is it not the case that a portion of stuff constitutes a thing? Markosian’s writings suggest, firstly, that he takes constitution “facts” to be brute; and, secondly, that it is up to the proponent of the mixed ontology to decide how he wants these constitution “facts” to play out. I take both of these commitments concerning constitution to be unappealing. Similar remarks apply to scenarios in which Markosian considers the possibility of atomless gunk instead of simples (see Markosian (2015), Fourth Reason). For one wonders why a world in which there is stuff-arranged-chairwise is not also a world which contains chairs. What prevents the stuff-arranged-chairwise, in such a world, from constituting a chair? It appears that these consequences are merely the result of certain terminological constraints Markosian imposes on his particular use of the term “constitution.” In this case, however, an appeal to alleged “facts” about constitution cannot be used to provide independent motivation for accepting a thing/stuff distinction. 36   Perhaps the most serious objection against Markosian’s account of simples is the argument from touching, especially in its application to the case of simple people. Suppose (as Markosian is willing to allow) that it is metaphysically possible for people to occupy maximally continuous regions of space; then, by MaxCon, there could be simple people. Suppose further that two such simple people, A and B, come into contact with one another. According to MaxCon, the result of A and B touching is a single object, C, which is itself a simple. Assuming that the scenario under discussion is described in such a way that neither A nor B has a better claim than the other to be identified with C after the contact has taken place, the proponent of MaxCon would presumably want to hold that, under these circumstances, both A and B cease to exist as a result of their contact and C is therefore numerically distinct from both A and B. (Given that C, according to MaxCon, is a simple, it furthermore cannot be described as having either A or B as proper parts.) By Markosian’s own admission, however, the consequence that two people can cease to exist merely as a result of touching seems quite counterintuitive. For more discussion, see Markosian (1998,  2004a, 2004b); McDaniel (2003).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

56  matter Markosian’s characterization, and how we could ever find out anything about it. According to Markosian’s particular brand of stuff theory, there are no ontologically significant subdivisions within the category of stuff; rather, in his view, stuff comes in only one fundamental type, viz., “generic stuff,” and all the other more familiar “stuff kinds” (e.g., water, bronze, etc.) are just “generic stuff ” arranged in a particular way (see, for example, Markosian (2015), p. 2). To illustrate, quarks, in Markosian’s view, are made of “quark stuff,” which is described as “the kind of matter that quarks are made of, and not, say, collections of quarks” (Markosian (2015), p. 2, n. 2). But “quark stuff,” for Markosian, is not a fundamentally different kind of stuff from, say, water, since he describes the latter as “generic stuff arranged in subportions that constitute H2O molecules” (p. 22, n. 47). Thus, the difference between, say, “quark stuff ” and water, according to Markosian’s account, is just that the former is “generic stuff ” arranged in subportions which constitute quarks, while the latter is “generic stuff ” arranged in subportions which constitute H2O molecules. As far as I can see, Markosian’s remarks concerning “generic stuff ” and its relation to the more familiar “stuff kinds,” such as water, can be interpreted in two ways; neither interpretation, however, leaves him in a good place. Consider Markosian’s characterization of water cited above as “generic stuff arranged in subportions that constitute H2O molecules.” Presumably, the “generic stuff ” in question, which is arranged in H2O molecule-constituting subportions, includes “generic stuff ” arranged in subportions which constitute hydrogen atoms. We can assume, further, that this “generic stuff ” in turn includes “generic stuff ” arranged in subportions which constitute electrons, protons, and neutrons, and so on. Suppose we live in a world in which there is a bottom level of subatomic stuff or particles, which are not composed of anything more basic; and let’s call these “boson stuff ” and “bosons,” respectively. “Boson stuff,” I take it, would be characterized by Markosian as “generic stuff ” arranged in subportions which constitute bosons. Is this “generic stuff ” different in any way from the “generic stuff ” which we assumed at any other level in the constitutional hierarchy? If not (and this seems to be the reading preferred by Markosian), then it seems that his view at this point collapses into either a “bare substratum” or a “prime matter” approach of the sort discussed in Sections 1.4–5 and 2.2. According to this reading of Markosian’s “generic stuff,” the very same kind of stuff can simultaneously give rise to something as different as a boson, a quark, an electron, a proton, a neutron, and an H2O molecule, merely by virtue of having its subportions arranged in different ways. I have already voiced my concerns that this conception of matter appears to lack empirical support. According to the second interpretation, the stuff which constitutes, say, hydrogen atoms, is viewed as belonging to a different kind from the stuff which constitutes, say, electrons. Let’s refer to the former as “hydrogen stuff ” and to the latter as “electron stuff ”; and let’s assume further that neither one of them would warrant the title, “generic stuff.” This, I take it, is what Markosian has in mind when he speaks of the Anaxagorean picture he opposes: a view which conceives of the world as consisting of different kinds of stuff and which does not regard these different stuff kinds as reducible to a single

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

the hylomorphic conception of matter  57 fundamental type of “generic stuff ” whose subportions can be arranged in different ways. The challenge that now arises for the Anaxagorean stuff theorist, however, is to tell us what this supposed “hydrogen stuff ” really is other than just pluralities of hydrogen atoms. Certainly, we do not encounter any empirical evidence from chemistry that creates pressure to posit a separate kind of “hydrogen stuff,” in addition to pluralities of hydrogen atoms. Rather, chemistry tells us that hydrogen atoms are entities whose nucleus contains a single proton and a single electron in the orbit of the nucleus (ignoring the isotopes). Thus, as far as our empirical evidence is concerned, we have reason to believe that hydrogen atoms are composed of a single proton and a single electron; but we have no reason to believe that hydrogen atoms are constituted of a special kind of “hydrogen stuff.” No mention of this special kind of “hydrogen stuff ” is required in order to account for the behavior of hydrogen in chemical reactions and the like. When all is said and done, then, I find myself in the same position I was in to begin with, before being presented with Markosian’s characterization and defense of stuff as a distinct and irreducible category: I still fail to understand what stuff is supposed to be, if it is not simply things or pluralities of things; how we could find out anything about it; and why we are supposed to need this alleged ontological category in the first place. Our ordinary ways of thinking and speaking, in particular our use of terms such as “stuff,” “matter,” “thing,” or “object,” and the mass/count distinction more generally, do not support anything as philosophically weighty as the thing/stuff distinction Markosian is trying to motivate. Physicists and chemists do not seem to find it necessary to appeal to anything like Markosian’s category of “generic stuff,” or to a more Anaxagorean picture, in order to account for the behavior of those entities with which they are concerned. Since Markosian’s version of the stuff theory in any case has some troubling features (e.g., his mysterious notion of “generic stuff ” as well as his, to my mind, unappealing views concerning constitution), I propose that, unless other considerations surface, we should stick with a thing ontology, which is really all we need to  satisfy our philosophical and empirical needs. The hylomorphic conception of ­matter, with which we will be concerned in the final section of this chapter, serves this purpose very nicely.

2.4  The Hylomorphic Conception of Matter The third approach to matter, (i.c), outlined at the beginning of this chapter, holds that matter–form compounds and their matter belong to the same ontological type: the matter of a concrete particular object, according to the hylomorphic conception, is nothing other than the material parts which compose it. So long as these material parts are themselves structured wholes, the hylomorphic analysis applies to them to the same extent as to the concrete particular object whose material parts they are. As I will illustrate, the third approach preserves many of the advantages of alternative conceptions of matter as prime matter or stuff, while avoiding their main disadvantages. This conception of matter does leave us with several challenges, but these can be addressed

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

58  matter by clarifying further in subsequent chapters, how hylomorphists should conceive of form and the relation(s) between a concrete particular object, its matter, and its form. Like its competitors, the hylomorphic conception assigns an important role to ­matter in explaining substantial change, viz., the coming to be and ceasing to be of matter–form compounds (see Section 1.6: (2) “Causation, Explanation, Change”). In a typical case, a newly created concrete particular object (e.g., an H2O molecule) comes to be from some pre-existing ingredients (viz., two hydrogen atoms and an oxygen atom) which are themselves structured wholes, i.e., matter–form compounds, according to the hylomorphic analysis. In cases in which the pre-existing ingredients persist through the substantial change in question, these matter–form compounds become the material parts of the newly created concrete particular object. In other cases, in which the preexisting ingredients do not persist through the substantial change in question (e.g., when some eggs are combined with other ingredients to bake a cake), whatever lower-level material parts do persist through the substantial change in question (e.g., the proteins which were previously among the material parts composing the eggs) in turn take on the role of serving as the material parts of the newly created concrete particular object. (Cases of substantial change which appear to fit neither of these two models will be considered shortly.) Either way, there is something which persists through the substantial changes in question, viz., the pre-existing ingredients or some of their lower-level material parts, and these persisting subjects become the material parts or matter of the newly created concrete particular object. The hylomorphic conception of matter does not require that the matter which underlies substantial change at micro-physical levels lacks an essence (regardless of whether the term, “essence,” is used here in a strict or in a loose sense), since both a matter–form compound and its material parts are taken to be concrete particular objects, and hence belong to the same ontological type. Whatever reasons motivate us to ascribe an essence to a composite concrete particular object will therefore be equally relevant to the material parts which compose the object in question, since both are regarded as matter–form compounds by the hylomorphic approach. In this way, proponents of the hylomorphic conception are able to avoid the threat of contradiction which arises, as noted earlier in connection with the doctrine of prime matter (see Section 2.2.3: (3) “Threat of Contradictions”), from the practice of ascribing certain characteristics to the ultimate substratum underlying substantial change at microphysical levels which do not appear to belong to it accidentally, while at the same time denying that this ultimate substratum is of the right ontological type to have an essence. The hylomorphic approach is compatible with an empirically supported hierarchical conception of matter (see Section 2.2.3: (1) “Potential Metaphysical Overreaching” and (2) “Non-Hierarchical Conception of Matter”). A quantity of water, for example, according to the present approach, is conceived of as a plurality of H2O molecules (ignoring the impurities that may or may not be present in any given quantity of water) and, as a structured whole, each individual H2O molecule counts as a matter–form compound; so do its material parts (i.e., the individual hydrogen and oxygen atoms),

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

the hylomorphic conception of matter  59 and so on down, unless or until we are faced with something that is material, but in itself unstructured. At that point, if such a possibility ever were to obtain, we would have reached a level in the compositional hierarchy at which the hylomorphic analysis no longer applies. In order to maintain consistency with an empirically confirmed hierarchical conception of matter, the hylomorphic approach must allow that, as a matter of contingent fact, any number of alternative possibilities for what might underlie substantial change at micro-physical levels are, as far as we know, still live options (see Section 2.2.3: (4) “Other Possibilities”): in particular, (i) the possibility of infinitely descending chains consisting of ever smaller matter–form compounds whose material parts are themselves matter–form compounds; as well as (ii) the possibility that we might eventually reach a point in the compositional hierarchy at which matter–form compounds are composed of material parts which are not themselves structured wholes.37 In order for the hylomorphic conception of matter to be compatible with such alternative possibilities for what might underlie substantial change at micro-physical levels, I amend the official formulation of the third approach with the following proviso: (i.c*) Matter as Hylomorphic Compounds. The material parts of matter–form compounds are themselves matter–form compounds, unless or until we reach an empirically confirmed level in the compositional hierarchy at which the material parts of matter–form compounds are not themselves structured wholes. The hylomorphic conception of matter, as stated in (i.c*), is thus committed to the position that the material parts of concrete particular objects are themselves matter–form compounds, as long as we have empirical confirmation that these material parts are themselves structured wholes. The hylomorphic conception of matter foregoes any commitment to a puzzling category of non-particulars or non-individuals (see Section 2.2.3: (5) “Non-particularity”). Since this approach classifies both concrete particular objects and their material parts as matter–form compounds, so long as the entities in question are structured wholes, these material parts are therefore eligible to figure in an account of the numerical identity and distinctness of the matter–form compounds they help to compose. As we will have occasion to observe in the next chapter, a reference to facts concerning the material parts of a matter–form compound cannot by itself be expected to resolve all 37   The second type of possibility might be realized, for example, in a scenario in which the material parts composing matter–form compounds are simples (i.e., concrete particular objects which have no proper parts) or portions of stuff. We saw earlier that there are empirical and philosophical reasons for resisting the particular conception of simples and stuff developed by Markosian. However, other conceptions of simples (e.g., “The Pointy View of Simples”) or stuff (e.g., what contemporary metaphysicians call “atomless gunk”) might nevertheless turn out to be defensible on empirical and/or philosophical grounds. As a proponent of the third approach to matter, I am not committed to the position that any of the possibilities, in fact, obtains; rather, I simply want to leave it open, provided that our empirical evidence points in this direction, that some such possibility might obtain, as a matter of contingent fact, and therefore should not be ruled out on a priori grounds.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

60  matter of the difficult questions surrounding the numerical identity and distinctness of matter–form compounds, since some of these responsibilities will also have to be borne by the remaining constituent of a matter–form compound, viz., its form. In addition, suppose an account of the numerical identity and distinctness of a concrete particular object, A, appeals to facts concerning the numerical identity and distinctness of some further concrete particular objects, B and C, which materially compose A. In that case, a further question arises as to how (if at all) facts concerning the numerical identity and distinctness of B and C are determined, and so on down, unless or until we reach a point at which we find ourselves appealing to facts concerning numerical identity and distinctness which can be assumed to be basic, i.e., not open to further explanation. In this way, even if the material parts composing matter–form compounds are of the right ontological type to supply us with some of the facts that are needed for an account of the numerical identity and distinctness of the wholes they compose, further questions concerning the numerical identity and distinctness of matter–form compounds will still be left open. Nevertheless, the hylomorphic conception of matter at the very least constitutes a step in the right direction, by assigning the material parts of matter–form compounds to an ontological category which makes these entities suitable to serve as a partial source for facts concerning numerical identity and distinctness of the wholes they help to compose.38 In response to the Problem of Material Constitution, proponents of the hylomorphic conception of matter will presumably point out that the relation of material constitution is just the relation which holds between a matter–form compound and its material parts, i.e., a type of mereological composition, since this approach is committed to the thesis that the matter of a concrete particular object is nothing other than its material parts.39 And while there is little motivation for proponents of the third approach to endorse a distinction between “actual” and “virtual” parts or unitarianism about forms, the resulting position thereby also creates the need to propose a non-eliminativist solution to the Problem of Unity and the Grounding Problem (see Section 2.2.3: (6) “Material Constitution, Unity, and the Grounding Problem”). For suppose, on the one hand, that a concrete particular object is composed of a plurality of material parts which, 38  In this connection, the hylomorphic approach also avoids a puzzle which arises for competing approaches (see n. 14), viz., how something which is non-particular/non-individual (prime matter or stuff) can give rise to something that is particular/individual (a matter–form compound), especially if the remaining constituent (form) is also taken to be non-particular/non-individual. Since a matter–form compound and its material parts, according to the hylomorphic conception, are of the same ontological type, there is no mystery as to how a plurality of concrete particular objects (viz., the preexisting ingredients, or their material parts, which come to serve as the material parts of a newly created concrete particular object) can give rise to another object of the same type. 39   See Koslicki (2008a), Ch. 7, for a defense of a mereological approach to the Problem of Material Constitution along these lines. Those who assume that constitution is to be conceived of as a one–one relation will no doubt find the response I just outlined surprising, since this strategy would allow that constitution may be a many–one relation, viz., when multiple material parts constitute a single concrete particular object. My reply to this concern is that we should not take it to be obvious that constitution must be conceived of as a one–one relation; rather, this is a question on which competing accounts of constitution may reasonably differ.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

conclusion  61 according to the hylomorphic conception, are themselves matter–form compounds and actually present in the whole they help to compose. In that case, one wonders how a plurality of such actually present matter–form compounds can together manage to give rise to a unified whole. Suppose, on the other hand, that a concrete particular object is composed of only a single material part which again, following the hylomorphic approach, will itself count as a matter–form compound that is actually present in the whole it helps to compose. In this case, the hylomorphic conception appears to be committed to the possibility of numerically distinct but spatiotemporally coincident objects and therefore owes us an account of what grounds their apparently different modal profiles. A proper treatment of the Problem of Unity and the Grounding Problem, both of which will concern us in more detail in subsequent chapters, requires additional apparatus in addition to what is supplied by the hylomorphic conception of matter developed in this chapter.

2.5 Conclusion Our focus in the current chapter was on the question of how hylomorphists should conceive of the matter composing a concrete particular object. Three main options were considered: (i.a) the doctrine of prime matter as developed by David Oderberg; (i.b) the matter-as-stuff hypothesis defended by Jeffrey Brower and Ned Markosian; and (i.c) the hylomorphic conception of matter, which I argued should be the preferred choice for hylomorphists. The doctrine of prime matter, which is popular among Thomists, comes out of a certain interpretation of Aristotle’s views concerning substantial change and the possibility of elemental transformations, though whether Aristotle in fact held this doctrine is a matter of scholarly controversy. We found that the representative competing approaches to matter we examined give rise to a number of difficulties. As far as our evidence is concerned, then, hylomorphists should opt for the third approach to matter, (i.c), according to which the matter composing concrete particular objects is nothing other than their material parts and these are themselves construed as matter– form compounds, unless or until we reach an empirically confirmed level in the compositional hierarchy at which the hylomorphic analysis no longer applies.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

3 Form 3.1  Introductory Remarks The doctrine of hylomorphism, as well as its application to the specific case of concrete particular objects, leaves open whether forms are (i) to be construed as universal or general entities or (ii) as particular or individual entities. In what follows, I refer to (i) as the “universal forms hypothesis” and to (ii) as the “individual forms hypothesis.” For present purposes, I understand the distinction between universal or general ­entities, on the one hand, and particular or individual entities, on the other hand, as follows: entities of the former type are in principle repeatable, i.e., they can be shared among multiple distinct entities by being wholly present in each of them at a single time; entities of the latter type, in contrast, by their very nature are not repeatable or sharable among multiple distinct entities by being wholly present in each of them at a single time. In what follows, we will encounter a range of options proposed in the literature which illustrate (i) and (ii), respectively. Yet a third option, (iii), rejects the dichotomy suggested by (i) and (ii) altogether and argues that forms do not neatly fit into either category, since they have features distinctive of both categories (the “hybrid position”). I will argue in this chapter that certain of the desiderata and decision points outlined in Chapter 1, which are of particular relevance to the ontological status of forms, are neutral between (i), (ii), and (iii): these are, in particular, the second and third set of desiderata and decision points outlined in Section 1.6 (see “Causation, Explanation, and Change” and “Essence and Accident”). By contrast, the fourth set of desiderata and decision points (see “Identity and Indiscernibility”), as I will go on to indicate, favors (ii), the individual forms hypothesis, over (i), the universal forms hypothesis. In addition, in Chapter 4, we will encounter further considerations (viz., in particular, the eighth set of desiderata and decision points, viz., the “Grounding Problem”) by means of which we can advance further in our goal of arriving at a clarified understanding of what sorts of entities forms are. Once all of these results are properly assembled, we arrive at a version of (ii) according to which forms are to be construed as robust particulars, i.e., as non-repeatable, non-sharable entities which, by their very nature, do not simultaneously belong to the matter–form compound (essentially) and to the matter composing it (accidentally).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

the ontological category of form  63

3.2  The Ontological Category of Form 3.2.1  Areas of Agreement and Disagreement In what follows, I give a brief, and by no means exhaustive, survey of positions with the intention of outlining at least roughly the range of options that are open to hylomorphists when it comes to the assignment of forms to a particular ontological category. Given space limitations, unfortunately I cannot devote the attention each undoubtedly deserves to the representative views I cite along the way, although some approaches will be discussed in more detail later on. Those readers who might find it frustrating to be presented with a plethora of different options, without being able to give each alternative close consideration, are advised to move on to Section 3.3. Despite the many interesting differences between the various positions I cite here, we can point to several significant areas of agreement that will emerge from even the cursory glance I provide in what follows. Whatever their particular commitments, hylomorphists tend to agree that the form, rather than the matter, is primarily responsible for explaining the following features of a matter–form compound: (i) Kind-membership: the membership of a matter–form compound in the most specific kind or species to which it belongs; (ii) Structure: the structure (arrangement, organization, configuration) exhibited by a matter–form compound or its material parts; (iii) Unity: the relatively high degree of unity that is exhibited by matter–form compounds, as compared to that displayed by other composite entities (e.g., heaps); and (iv) Characteristic Activities: the range of behaviors in which things of the type in question are characteristically found to engage. Although hylomorphists do not always agree on exactly how the form of a matter– form compound does the work that is assigned to it according to (i)–(iv), they do tend to hold that it is the form of a matter–form compound, rather than its matter or the matter–form compound itself, which takes on these explanatory responsibilities. At the same time, we will also come across several noteworthy areas of disagreement among the various theorists canvased here, not only with respect to the ontological category to which forms are assigned, but also (and relatedly) with respect to the explanatory tasks which are supposed to be settled by appeal to the form of a matter– form compound. To illustrate, some hylomorphists (e.g., Jeffrey Brower and Eleonore Stump), in addition to their philosophical commitments, have tailored their conceptions of form to meet certain theological requirements as well, e.g., those which in their view, arise in connection with Aquinas’ take on the doctrines of transubstantation or personal immortality. In addition, perhaps the greatest distance between those who take forms to be universal or general entities of some kind and those who subscribe to

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

64  form some version of the individual forms hypothesis lies in how they expect to settle questions concerning the (synchronic, diachronic, or cross-world) numerical identity or distinctness of matter–form compounds. We return to this important topic in more detail in Section 3.4.

3.2.2  The Universal Forms Hypothesis Properties or Relations. Some theorists take forms to be properties (attributes, features, characteristics, modes, or ways of being) or relations (or collections thereof), where properties and relations, according to this option, are conceived of as general or universal entities that are shared among members of a single kind. According to the interpretation of Aristotle developed in Peramatzis (2011), for example, forms are properties of a certain kind, conceived of as universal “modes or ways of being” which are essential to the types of objects possessing them (Peramatzis (2011), p. 4). For example, Socrates’ essence or form is to be a human being, while Bucephalus’ essence or form is to be a horse. Forms themselves, in Peramatzis’ view, have essences, which are also modes or ways of being; but, in this case, the essence of the thing and the thing whose essence it is (viz., the form) are identical (p. 5). According to Michael Loux’s reading of Aristotle, a form is a universal of a certain kind which is present in a matter–form compound as a constituent and predicable of its matter, where predication here is conceived of as a non-linguistic relation among non-linguistic items (cf., e.g., Loux (2005), pp. 94–5). For every lowest level natural kind (or infima species), S, there is a “primitive or unanalyzable universal that is necessarily such that it is a proper constituent in all and only the members of S and it is the primary substance of each of the things whose proper constituent it is” (pp. 118–19). In the case of living beings, for example, the form is the arrangement or organization which has to be exhibited by the matter in order for a particular matter–form compound to be a member of the species to which it belongs (e.g., geranium, dog, human being, etc.). Johnston (2006) describes forms as principles of unity which hold together the parts of a matter–form compound. A principle of unity, in Johnston’s view, is a relation of a certain kind, namely “a relation holding of some other items, such that (origins aside) what it is for the given item to be is for the relation to hold among those items” (Johnston (2006), p. 653). For example, if the matter–form compound in question is a model airplane, then the form of the model airplane is roughly the relation which holds among the parts of the model airplane (viz., the wings, the tail, the fuselage, etc.), when these “hang together in the modeled shape of an airplane in such a way as to resist separation in the face of the range of forces to which we usually subject such models” (p. 653). The form of a hydrogen chloride molecule is the relation, bipolar bonding: for there to be a given hydrogen chloride molecule is for there to be a hydrogen ion and a chlorine ion together in a bipolar bond (p. 653). The same idea, in Johnston’s view, applies to abstract items as well (e.g., conjunctive properties, sets, sentences or propositions): the form or principle of unity of a subject–predicate sentence token, for example, is the relation, predication, which holds between the reference of the subject term and the

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

the ontological category of form  65 denotation of the predicate term and which makes it the case that the sentence token is capable of being true or false (p. 657). The relations in question are conceived of by Johnston as universals, so that different exemplars belonging to the same type, e.g., numerically distinct token subject–predicate sentences, share the same form (p. 659). Even mereological sums, for Johnston, can be considered matter–form compounds, only their form or principle of unity is the least demanding one imaginable, since it requires nothing more than that the parts in question exist (pp. 688–9). Powers. Alternatively, forms are conceived of by some to be powers, capacities, dispositions, or propensities (or collections thereof), where these are taken to be universal or general entities shared by objects of the same type. Anna Marmodoro, for example, follows the general outlines of a reading of Aristotle first championed in Scaltsas (1994): “The substantial form according to Aristotle is an operation on the elements of a substance, stripping them of their distinctness, rather than being an item in the ontology” (Marmodoro (2013), p. 17). Forms, on this reading, are universals, since they are common to all concrete particular objects of a particular kind. They are ways in which previously existing concrete particular objects (e.g., bread and water) can combine when they enter into a new matter–form compound (e.g., flesh): “Bread and water are re-formed, re-shaped qualitatively and functionally, to make up flesh” (p. 11). On Marmodoro’s reading, Aristotle holds that “the ultimate level of reality is the fundamental powers (hot, cold, wet, and dry)” (p. 11). Activities. Others take the form of a matter–form compound to be a certain type of activity in which it characteristically engages. Kosman (2013), for example, stressing Aristotle’s proposed solution to the problem concerning the unity of matter and form (see Met. H.6, 1043a23–6), conceives of the matter–form distinction in terms of Aristotle’s distinction between dunamis (capacity or potentiality) and energeia (activity or actuality) as follows: Matter and form are not linked together by the possibilities of becoming. They are present together in the being that is nothing other than the active essence—that is, the essential activity—of the one being that both are. It is in this sense that matter is the locus of ability or capacity—dunamis—and form, as the principle of that ability’s exercise, is the principle of its active being—its energeia. Form, we might say, is the principle of matter in operation; in instances of substantial being, it is the operation of the being that is substance. (Kosman (2013), p. 81)

The body or matter of a living organism, for example, is understood, on this view, as a “locus of ability”; its soul or form as “the principle of that ability’s active exercise” (Kosman (2013), p. 81). For a horse, the activity in question is just being a horse or carrying out the ergon or work that is characteristic of horses. Elaborating the details of what this activity consists in is the task of biology (p. 85; p. 101). Acts of Creation. According to this approach, what it is to be a screwdriver, for example, is to be an object that is produced from some suitable matter (e.g., some metal and plastic) in an act of creation that is guided (and partly individuated) by the maker’s intention to create an object which can serve a certain function (viz., to tighten or

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

66  form loosen screws). The form of a screwdriver, on this view, is identified with a certain kind  of historical essence, viz., the object’s having been brought into existence by means of a creative intentional act of a particular type. This approach is a universalist version of the particularist account developed by Evnine (2016a) (see Section 3.2.3 and Chapter 8). To extend this approach to the natural realm, a substitute (e.g., the evolutionary history of a biological species) has to be found which can take over the role played by intentional acts of creation in the case of artifacts. Sui Generis Universal Entities. Finally, proponents of the universal forms hypothesis also have the option of assigning forms to some hitherto unrecognized ontological category of sui generis universal or general entities which do not neatly fit into any of the options outlined here. According to the proposal under consideration, forms are to be construed as entities which are in principle repeatable or sharable, i.e., entities which can be wholly present in multiple distinct entities at a single time; and yet they are also taken to lack the characteristic features associated with the ontological categories already cited (viz., properties, relations, powers, activities, or act types), or others which are already familiar to us from non-hylomorphic contexts.

3.2.3  The Individual Forms Hypothesis Objects. Some proponents of the individual forms hypothesis take forms to be themselves objects of some kind. Michael Frede, for example, reads Aristotle as postulating in the Metaphysics that forms are primary substances (cf., e.g., Frede (1985)), diverging from the earlier Categories ontology in which this role is assigned to concrete particular objects. The soul of a living organism, according to Frede, is that of a particular or individual entity which “has to stay the same as long as a particular animate object exists,” while the organism’s matter and properties may change from one time to another (p. 76). Individual forms, in Frede’s view, are “the real individuals in the category of substance which are to explain the individuality of ordinary individual objects” (Frede (1987a), p. 64). The form’s role in settling facts about the numerical identity of concrete particular objects, so Frede maintains, “may give some plausibility to the assumption that it is really the form which is the thing we are talking about when we at different times say different things about an object” (Frede (1985), p. 76, my emphasis). Although our specifications of individual forms will be exactly the same across all members of a single species and will not reveal any “intrinsic essential distinguishing mark” differentiating one concrete particular object from another, the forms themselves are nevertheless particular (p. 78).1 A view such as Frede’s need not collapse into full-blown Platonic or 1   Frede also refers to the individual form of a living organism as “its organization, structure, and disposition,” viz., that in virtue of which the organism has the capacity to behave in a certain characteristic way and live the kind of life that is characteristic of the kind of organism it is (see, e.g., Frede (1987a), p. 66). This aspect of Frede’s view makes it sound like a version of the forms-as-particular-powers approach which will be reviewed shortly. At the same time, Frede is also explicit about wanting to assign individual forms to the category of substance and Aristotle presumably would classify the organization or disposition of a thing to behave in a certain way as a dunamis (capacity, potentiality) of a substance, and not as itself a substance (using “substance” here in the non-relational sense of “substance simpliciter,” as opposed to the

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

the ontological category of form  67 Cartesian dualism, however, since forms, despite their status as primary substances, are also taken to be inseparable from the matter in which they are found. In connection with their comments on Met. Z.7, Frede and Patzig (1988) remark: Richtiger ist der Text [1032b1–2] aber wohl so zu verstehen, daß der Mensch eine Seele ist, die außerdem noch dadurch charakterisiert ist, daß sie sich an einer bestimmten Art von Körper findet. Denn ohne diese Art von Körper könnte sie nicht existieren; und ohne diese Art von Körper wäre die Seele kein Mensch.  (Frede and Patzig (1988), p. 113)2

Properties or Relations. According to this approach, forms are taken to be properties or relations (or collections thereof), conceived of as particulars, e.g., tropes, moments, modes, or ways. According to the interpretation of Aquinas’ hylomorphism offered in Brower (2014), for example, matter, form, and the compound could in principle be identified with any type of being capable of playing the relevant functional roles. In fact, however, Brower reads Aquinas as identifying these entities, at least in some cases, with concrete particulars (or substances), immanent properties, and complexes (or ­concrete states of affairs). For Brower (p. 22), Aquinas conceives of forms/properties as individuals, and does not see the need to posit universal forms/properties to explain kind membership, as is illustrated in the following passage from Aquinas’ treatise, Scriptum Super Libros Sententiarum: Even if this individual [say, Socrates] is a human being and that individual [say, Plato] is a human being, it is not necessary that both have numerically the same humanity—any more than it is necessary for two white things to have numerically the same whiteness. On the contrary, it is [only] necessary that the one resemble the other in having an [individual] humanity just as the other does.  (In Sent. 2.17.1.1, Brower’s translation)

But Thomistic properties, according to Brower’s reading, are not exactly like tropes, since Aquinas only accepts the following version of non-transferability: in all possible worlds in which a certain property exists and God does not miraculously intervene, the property, once it exists and is possessed by a certain substance, continues to exist and is possessed by the same substance. With supernatural intervention, however, an accident can exist apart from the substance which individuated it and once possessed it (e.g., as in the case of the Eucharist). In his account of transubstantiation, Aquinas allows that accidents can have other accidents as substrata (Brower (2014), pp. 247–8).3 relational sense of “the substance of a thing”). At most, then, we can chalk up Frede’s tendency to speak of individual forms in this way to a potential tension within his own view: depending on which direction is emphasized, we could read him as subscribing either to a version of the forms-as-particular-objects approach or to a version of the forms-as-particular-powers approach. 2   “But this text is more correctly understood as saying that a human being is a soul, which is also characterized as being found in a certain kind of body. For without this kind of body the soul could not exist; and without this kind of body the soul would not be a human being.” (My translation.) 3   See also Brower (2016) for further discussion. The special case of the human soul is discussed in Brower (2014), ch. 11, especially Sections 11.4–11.5. In this case, substantial forms and properties come

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

68  form States. According to the reading of Aquinas offered in Stump (1995), the form or soul of a person is a state of a certain kind, namely, a configurational state. This approach, in Stump’s view, allows Aquinas to occupy a middle ground between Cartesian dualism and eliminative materialism, since the soul is thought of as essentially immaterial and yet nonetheless as realized in material components (Stump (1995), pp. 505–6). In particular, this reading, so Stump argues (pp. 506ff), allows one to make sense of the theological doctrine of the afterlife, as interpreted by Aquinas. The challenge is to reconcile two apparently conflicting aspects of Christian teaching: that, on the one hand, a human being is a material object (“dust thou art, and unto dust shalt thou return.” Genesis 3:19), while, on the other hand, a human person is also said to survive death (“Then shall the dust return to the earth as it was, and the spirit shall return unto God who gave it.” Ecclesiastes 12:7). In Stump’s view, the best way to accommodate both of these biblical injunctions within the framework of Aquinas’ hylomorphism is as follows: . . . [T]he soul is an essentially configurational state which is immaterial and subsistent, able to exist on its own apart from the body. On the other hand, the soul is the form that makes the living human body what it is. While it is possible with divine help for the soul to exist and exercise cognitive function on its own, apart from the body, that state is unnatural to it. In the natural condition, human cognitive functions are to be attributed to the whole composite and not to the soul alone, although the composite exercises cognitive functions by means of the soul.  (Stump (1995), p. 519)

It is easy to see how Stump’s interpretation can account for the idea, captured by the first biblical doctrine, that human beings are material objects which disintegrate upon death. To allow for the possibility that the disembodied soul might persist after death, however, Stump adopts a conception of Thomistic substantial forms according to which human souls are contingently configurers which are also themselves configured: they are configurers of their matter during certain phases of their existence, namely, prior to the separation of the soul from the body; after death and prior to resurrection, they are themselves configured without configuring anything, like the souls of angels (Stump (1995), pp. 514–15). One of the ways in which this conception of the human soul is distinct from Cartesian dualism, Stump maintains, is that the human soul is not itself conceived of as a complete substance, even though it is capable of continued existence, with the help of God’s intervention, in an incomplete separated state (p. 517). Functions. According to the neo-Aristotelian mereology developed in Fine (1999), which elaborates on and refines the theory of “qua objects” given in Fine (1982), a composite material object can be thought of as a “variable embodiment,” /F/, whose formal aspect is a principle, F, and whose matter or material parts are whatever objects satisfy the principle in question at any time at which the complex material object in question exists (pp. 68–9). For example, the water which fills a certain river, in Fine’s view, is a variable embodiment, /F/, whose principle, F, selects the different quantities apart: although the human soul is a form, it is not a property; rather, it is an incorporeal substance (Brower (2014), pp. 251–4).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

the ontological category of form  69 of water which constitute the river at each time at which it exists. The principle, F, of a variable embodiment, /F/, is described by Fine as “any suitable function” from times to objects (p. 69), though no clear constraints are imposed on what might make a ­function from times to objects “suitable” for the formation of a variable embodiment. In fact, Fine suggests that principles of embodiment are “intensional or conceptual in nature,” and that the resulting ontological commitments need not be regarded as “ultimate” (p. 73).4 Powers. Some theorists take the forms of concrete particular objects to be powers, capacities, dispositions, or propensities (or collections thereof), conceived of as particulars. Corcilius and Gregoric (2010), for example, regard the capacity for theorizing, in the case of Aristotle’s unmoved mover, as identical to the divine soul (p. 91); and the nutritive capacity, in the case of plants, as identical to the soul of a plant (p. 92). In the case of non-human and human animals, in their view, the soul is a complex entity whose parts are the various capacities that are essential to or constitutive of what it is to be the type of living organism in question. William Jaworski has also argued for a version of the individual forms hypothesis, according to which forms (or “structures,” as he calls them) are powers of a certain kind, construed as particulars, viz., the power of a structured whole to organize or configure its material parts in a certain way characteristic of the kind to which the whole in question belongs: I argue that structures are powers to organize or configure things—powers that structured wholes are essentially engaged in manifesting. You and I are essentially engaged in configuring the materials that compose us; we impose a human-wise organization on them, and we persist exactly as long as we do so. These particular configurings—yours and mine—are particularized properties or tropes. They are numerically different properties that nevertheless resemble each other rather closely—more closely than, say, either resembles Fido’s configuring the materials that compose him or the oak tree’s configuring the materials that compose it. (Jaworski (2016b), p. 4)

In the case of concrete particular objects, Jaworski takes the relevant powers to be “individual-making structures”; in the case of activities (e.g., thinking or perceiving), they are “activity-making structures.” The latter idea, viz., that activities themselves have a hylomorphic structure, turns out to be crucial for Jaworski’s proposed contributions in the philosophy of mind: The key to understanding the hylomorphic approach to mind–body problems is the notion of an activity-making structure. The structures [. . .] that make individuals what they are [are] the kinds of things traditional hylomorphists called “substantial forms.” But individual-making structures are not the only structures that exist on the hylomorphic view. The activities in which structured individuals engage have structures as well; they are activity-making structures. The idea that there are activity-making structures is based on the observation that the activities of   For a more detailed discussion of Fine (1982 and 1999), see Koslicki (2007b and 2008a), especially Ch. 4.

4

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

70  form structured individuals involve coordinated manifestations of the powers of their parts. When we walk, talk, sing, dance, reach, grasp, run, jump, throw, breathe, and engage in the various other activities we do, we are imposing an order on the ways our parts manifest their powers. On the hylomorphic view, these structured manifestations of powers include thinking, feeling, and perceiving.  (Jaworski (2016b), p. 5)

Activities. The position taken by Kosman (2013), reviewed earlier, has a correlate which is compatible with the individual forms hypothesis. According to this approach, the individual forms of concrete particular objects are activities of a certain kind (or collections of such activities), where the activities in question are conceived of as particulars. Thus, if there is such as thing as a characteristic activity in which Bucephalus engages, not only qua horse but qua the particular horse that Bucephalus is, then Bucephalus’ individual form, on this approach, is identified with this activity of being the horse Bucephalus, or with carrying out the ergon or work that is characteristic of the particular horse, Bucephalus.5 Facts. According to Sattig (2015), forms are facts of a certain kind. An ordinary object, in Sattig’s view, is a “double-layered compound,” consisting of a material object (viz., its matter) and a complex fact (viz., its individual form). The fact in question “contains properties that realize an ordinary kind, such as person or table” (Sattig (2015), p. viii). In Sattig’s view, a sortal or kind, such as person or table, has a certain qualitative content by means of which material objects, which he takes to be just the mereological sums of classical mereology, can be traced through time (Sattig (2015), p. 16). Since the identity conditions of mereological sums can, and often do, diverge from those associated with the sortal concepts satisfied by ordinary objects at particular times during their careers, these objects (i.e., compounds of mereological sums and facts), for Sattig, lead a “double life” of sorts, depending on whether they are considered from the perspective of their material or their formal aspect. Acts of Creation. Evnine (2016a) defends a version of hylomorphism according to which this doctrine applies most centrally to the case of artifacts. In Evnine’s view, an artifact is essentially the product of a making, i.e., an intentional act of creation which (in typical cases) involves a maker working on some matter with the intention of ­creating a certain type of object (e.g., a screwdriver). Often, artifacts are also associated with functions (e.g., screwdrivers are for tightening or loosening screws) which they acquire from the intentions governing the creative acts which bring these artifacts into existence. In this sense, Evnine’s account strives to preserve the Aristotelian doctrine that, in central cases, the formal cause (or essence), the efficient cause (or origin), and the final cause (the purpose or end) of a concrete particular object coincide. In contrast to other versions of hylomorphism, however, Evnine objects to the idea that “hylomorphically complex objects” (as he calls them) are really compounds of their matter and some other entity, viz., their form. Instead, Evnine holds that, when a hylomorphically complex object is brought into existence through an intentional act of creation,   See also Koons (2014) for an approach according to which forms are construed as processes.

5

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

the ontological category of form  71 the process in question requires only that the matter which comes to compose the newly created artifact be worked on in a certain way, but not that the matter be combined with a further entity, a form. As a result, the matter acquires a certain history, namely, the history of having been worked on by the maker in a particular act of ­creation, and this historical origin in turn comes to be associated essentially with the artifact that is created as a result of the imposition of the maker’s labor and intentions on the matter in question. Although Evnine takes artifacts to be the central case for his doctrine of hylomorphism, he allows that certain natural objects, viz., in particular living organisms, can also be regarded as hylomorphically complex objects. In this case, the evolutionary history of these objects takes the place of intentional acts of ­creation in the case of artifacts. However, for large swaths of cases (viz., everything that is neither an artifact nor a natural living organism), Evnine recommends a version of eliminativism, supplemented by a fictionalist interpretation of our discourse about mountains, rivers, stars, and the like.6 Sui Generis Particular Entities. Finally, proponents of the individual forms hypothesis also have the option of assigning forms to some hitherto unrecognized ontological category of sui generis particular or individual entities which do not fit neatly into any of the options outlined here. According to the proposal under consideration, forms are to be construed as entities which, by their very nature, are non-repeatable or non-sharable, i.e., entities which cannot be wholly present in multiple distinct m ­ atter– form compounds at a single time. At the same time, the sui generis approach also has it that forms lack the characteristic features associated with any of the ontological categories already cited (viz., objects, properties, relations, states, functions, powers, activities, facts, or acts of creation) as well as those of other categories which are already familiar to us from non-hylomorphic contexts. This approach, like its universalist counterpart, brings with it the significant cost of requiring the recognition of an entirely new ontological category which is not already motivated on other, non-hylomorphic, grounds. Those who object to hylomorphism because they find Aristotelian forms mysterious will likely feel strengthened in their opposition to this type of account, if it turns out that forms cannot be accommodated by our array of traditionally recognized ontological categories.

3.2.4  The Hybrid Position According to the hybrid position, the particular/universal dichotomy is considered to be neither exclusive nor exhaustive. Forms are taken to be neither fully particular nor fully universal; rather, they are regarded as having some of the features of particulars and some of the features of universals. The hybrid position is nevertheless compatible with the assignment of forms to one of the ontological categories listed earlier; the

6   We will have occasion to examine the conception of artifacts proposed in Evnine (2016a) in more detail in Ch. 8.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

72  form claim would then be that the category in question itself must be recognized as a hybrid of universal and particular. According to Code (1984), we should read Aristotle as assigning to forms some of the features of a particular and some of the features of a universal. For one thing, forms, in their role as primary substances in the Metaphysics, have to be regarded as strictly speaking particular, since they are predicable only of matter and not of a plurality of primary substances. On the other hand, forms cannot be considered to be separable particulars, in the way in which matter–form compounds are, since they cannot exist independently of the matter–form compounds with which they are associated. At the same time, forms are also knowable and definable, and in that respect they resemble universals. In addition, forms are general entities, since members of a single species are found to exhibit the same range of capacities. Mary Louise Gill, in her reading of Aristotle’s Metaphysics, takes forms to be fully determinate, repeatable, and definable: they are individual in the sense of being fully determinate, but universal in the sense of being repeatable and definable (Gill (1989), p. 34, n. 47). As for separability, Gill distinguishes between two types: separability in account or definition and what she calls “simple separability,” separability in existence. While forms are separable in the first sense, the forms of material objects lack the ­second kind of separability, since they must be present in matter–form compounds in order to exist. Other forms, however, e.g., that associated with Aristotle’s unmoved mover, can exhibit both kinds of separability. In contrast, matter–form compounds, in Gill’s view, are characterized by the second kind of separability, but not the first: they are separable in existence, but not separable in account or definition. Both forms and matter–form compounds, in Gill’s view, qualify as primary substances, but in ­different ways. Among contemporary defenders of hylomorphism, Michael Rea also adopts the hybrid position (see Rea (2011)). In Rea’s view, hylomorphism, as it is traditionally understood, carries with it a commitment to several assumptions which he finds ­controversial and wishes to avoid, among them the following: (i) a distinction between universals and particulars;7 (ii) some form of realism about either universals or tropes (since forms are typically taken to be kind properties, construed either as universals or as tropes); and (iii) some notion of “constituency” according to which matter and form can be said to be “in” a concrete particular object. Rea’s goal is to develop a version of hylomorphism which avoids these commitments, but nevertheless manages to express everything hylomorphists intend to be able to say. He does so by taking on board as 7   In connection with (i), Rea notes, for one thing, that existing attempts to draw a universal/particular distinction all face problems of various sorts (see MacBride (2005); Ramsey (1925)); he has in mind here, for example, attempts which are formulated in terms of (a) a subject–predicate distinction; or the idea that universals, but not particulars, are (b) multiply locatable, (c) obey the law of the identity of indiscernibles, or (d) are instantiable or non-inherent (Rea (2011), p. 343). In addition, Rea believes that a universal/particular distinction is not needed, since all the necessary philosophical work can be accomplished without such a distinction.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

the individual vs. universal forms debate  73 primitives the concepts of power, location, ground, and numerical sameness (Rea (2011), p. 345).8 Although powers, in Rea’s view, are to be understood neither as universals nor as particulars, they behave in some ways like universals (according to those who believe in them), since a single power can be simultaneously present in distinct regions of space–time and we can say of an object that it “has” a power. In other ways, however, powers behave like particulars (according to those who believe in them): e.g., they are able to enter into causal relations and are located in regions of space–time (p. 346). The role of forms, in Rea’s power-based hylomorphic theory, is played by natures, which in turn are powers of a certain kind: the natures of substances, according to Rea, are fundamental powers which act as principles of unity, unifying the other natures that are present in the region of space–time that is occupied by the object in question.9 The role of matter, in Rea’s theory, is played by whatever individuates the substantial nature in question: either regions of space–time, in the case of simple objects that lack parts; or collections of parts, in the case of complex objects that have parts. Thus, a simple object, according to Rea, is a hylomorphic compound of a substantial nature (its form) and the space–time region it occupies (its matter), while a complex object is a hylomorphic compound of a substantial nature (its form) and a collection of parts (its matter), viz., those objects whose natures are unified by the complex object’s nature.

3.3  The Individual vs. Universal Forms Debate In this section, I briefly highlight some of the considerations which have figured prominently in the debate among ancient scholars over whether Aristotelian forms should be construed as universals, as individuals, or as hybrid entities of some kind. Since historians of philosophy have devoted much careful attention to the question of how best to conceive of Aristotelian forms, even those of us who are mostly in the business of proposing a hylomorphic analysis of concrete particular objects for contemporary consumption would be foolish not to take note of the sorts of factors that have moved these scholars in one direction or another in their reading of the relevant texts. We will find that many of these motivations are transferable to the contemporary debate concerning the metaphysics of concrete particular objects, even once we leave primarily interpretive questions behind. 8   As already noted, Rea regards numerical sameness as a genus of which numerical identity is merely a species, since non-identical objects, in his view, can be the same material object, e.g., in cases of constitution (see Rea (1998)). Rea’s “sameness without numerical identity” view is not required to make his powerbased hylomorphic theory work, though he formulates the latter in such a way that it is compatible with the former. 9   Substance natures, in Rea’s view, are “fundamental” powers, in the sense that (i) they are perfectly natural; (ii) they are not reducible to other powers; and (iii) they ground non-natural powers (Rea (1998), p. 347). A power P (e.g., humanity) “unites” some other powers (e.g., the capacity for rational thought, growth, running, dancing, etc.) just in case: “P is so connected to the other powers that its manifestation depends upon the cooperative manifestation of the united powers and, furthermore, the latter do not confer any powers on the object that has P that are both intrinsic to the object and independent of P” (pp. 348–9).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

74  form The following issues are often invoked by participants in the debate over the nature of Aristotelian forms to bolster their interpretation of the relevant texts. I will not, in the present context, try to take a stand on what I take to be the best overall interpretation of Aristotle. In what follows, I indicate only when a particular consideration seems, or has been taken, to count as prima facie evidence in favor of one interpretation over another. It should be understood, however, that, in every single case, there is always much more that can be (and has been) said by the opposing camp to attempt to make the point in question consistent with their reading of Aristotle.10 (a)  Forms as the proper objects of knowledge/definition. Aristotle takes forms to be the proper objects of knowledge and definition. He also holds, however, that knowledge (at least in the sense of epistēmē, i.e., the demonstrative or scientific knowledge that is at issue in the Posterior Analytics), is of universals and that only universals are definable. The status of forms as the proper objects of knowledge and definition thus seems to provide at least prima facie support for the universal forms hypothesis and against the individual forms hypothesis. (b)  The existence of forms as presupposed in the generation of particular compounds. Met. Z. 7–9 argues that the existence of both form and matter must be presupposed in an explanation of how particular matter–form compounds come into being. This commitment on Aristotle’s part is sometimes read as providing evidence in favor of the universal forms hypothesis and against the individual forms hypothesis. For if forms are construed as particulars and they are nevertheless taken to be capable of existing without the matter–form compounds with which they come to be associated, the resulting conception of forms might strike us as more Platonic than Aristotelian. (c)  The ultimate subject criterion. One of the criteria for primary substancehood Aristotle invokes in various texts (e.g., Categories, Met. Z.3) is the ultimate subject ­criterion, according to which the primary substances are those entities which are themselves not said of anything, while other things are said of them. At the same time, Aristotle also on occasion seems to speak of form as predicable of matter (e.g., arguably at Met. Z.3, 1029a21–4). The first of these positions is, at least at first sight, easier to make sense of for proponents of the individual forms hypothesis than for their opponents, since Aristotle also holds that particulars cannot be predicated of anything, while other things are predicated of them. For example, to say of something that it is Socrates, in Aristotle’s view, is to identify the entity in question with Socrates, rather than to predicate anything of it. The second doctrine (viz., that form can be predicated of matter), however, is difficult to square with the individual forms hypothesis: for if forms are the sorts of things that can be predicated of something (viz., matter), it seems that they would have to be universals, rather than particulars, since only universals, in Aristotle’s view, are predicable of many things, viz., the particulars which exemplify them. 10   In what follows, in order to keep an already very complex dialectical situation at least somewhat manageable, I will bracket the hybrid position and characterize the debate over Aristotelian forms as confronting us with just two main choices, viz., the universal forms and the individual forms hypothesis.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

the individual vs. universal forms debate  75 (d)  Individuation. According to one traditional way of interpreting Aristotle, matter, rather than form, is taken to be the principle of individuation for matter–form compounds. At Met. Z.8, 1034a5–8, for example, Aristotle famously says that Socrates and Callias are the same in form but different on account of their matter. This reading of Aristotle is more naturally combined with the universal forms hypothesis, while proponents of the individual forms hypothesis tend to shift the work associated with individuating matter–form compounds away from their matter and onto their form. (e)  The alleged identification of form and essence. According to some texts (e.g., Met. Z.4–6), the substance of a thing is taken to be the same as, or at least not different from, its essence. Assuming that here Aristotle has in mind the form of a matter–form compound when he speaks of the substance of a thing, these passages may lead us to believe, or at least provide preliminary evidence to the effect that Aristotle wants to identify the essence of a thing with its form. Other texts (e.g., Met. Z.11), however, have been interpreted as providing support for the conclusion that the essence of a thing in some way also includes its matter, at least in a generic fashion. These considerations by themselves do not seem to point in favor of either the universal forms hypothesis or the individual forms hypothesis. Whatever conception of form is adopted, it can be matched by an appropriate notion of essence: (non-derivative) individual essences are an option only for proponents of the individual forms hypothesis, while proponents of the universal forms hypothesis will opt for (non-derivative) kind-essences instead. Either way, however, Aristotelian essences can be understood to be either purely formal or as at least partly material, regardless of whether they are taken to apply primarily at the level of kinds or at the level of individuals. (f)  Apparent arguments against the substance status of universals. In some texts (e.g., Met. Z.13–16), Aristotle endorses the following claims: (i) universals are not substances; and (ii) substances are not composed of substances which are present in them actually. But Aristotle’s arguments in Met. Z, overall, are also generally read as leading to the conclusion that forms deserve substance status most of all, i.e., more so than their most likely competitors, viz., matter and the matter–form compound. Aristotle’s apparent case against the substance status of universals in Met. Z.13–16 has given proponents of the universal forms hypothesis a run for their money, whereas proponents of the individual forms hypothesis have generally pointed to this piece of the overall puzzle as a major consideration in favor of their interpretation. (g)  Forms as causes and the causal priority of forms. Aristotle not only takes forms to be causes, he also assigns them priority over the matter and the matter–form compound in their role as causes and principles (see especially Met. Z.17). At least in the case of living things, forms are said to function as the formal, efficient, and final causes of matter–form compounds, while the matter composing them only plays a subsidiary role as the material cause of the matter–form compound. In addition, we are told in certain texts (e.g., in Met. Λ.5) that there is never any need to invoke general or universal causes and principles to explain particular changes undergone by objects. Rather, according to this line of reasoning, it is only ever necessary to refer to particular causes

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

76  form and principles in a general way, since particular objects which belong to a single kind will exhibit the same types of particular causes and principles. These considerations seem to play into the hands of individual forms theorists who find it difficult to see how a universal which is common to all members of a given species could act as a cause of a particular matter–form compound’s engaging in some particular activity at a particular time and place. Summary. The foregoing remarks have taken us through some of the main issues which are discussed by Aristotle scholars on either side of the universal vs. individual forms debate. We can also now see, I think, why this dispute has proven to be so intractable, since the prima facie evidence which can be brought to bear by one side in the debate can be matched by apparently equally weighty considerations on the opposing side. Several of the points mentioned here (viz., in particular, (d), (e), and (g)) will be of interest to us again, when we return to the desiderata and decision points identified in Chapter 1 which are particularly relevant to how hylomorphists should conceive of the form that is present in a matter–form compound. I will unfortunately have to leave aside the remaining points (viz., (a), (b), (c), and (f)) for the time being, though their central importance for the task of developing a credible reading of Aristotle’s conception of form is undeniable.

3.4  In Defense of Individual Forms In this section, I defend my version of the thesis that forms are best construed as ­particular or individual entities of some sort by arguing that only individual forms can help us settle tricky questions about the cross-world identity and distinctness of matter–form compounds.11 In Sections  3.4.1 and  3.4.2, I consider the causal roles ascribed to form and the relation between form and essence and argue that these considerations by themselves do not tip the scales in favor of either the individual forms or the universal forms hypothesis. In Chapter 4, I go on to supplement my version of the individual forms hypothesis by arguing that forms should be construed as “robust” particulars, i.e., as non-repeatable, non-sharable entities which, by their very nature, do not simultaneously belong to the matter–form compound (essentially) and to the matter composing it (accidentally).12 11   The arguments advanced in the remainder of this chapter will not directly concern the fifth, seventh, and eighth sets of desiderata and decision points (see Section 1.6, “Material Constitution,” “Unity,” and “The Grounding Problem”), since these sets of issues either have been, or will be treated, in other chapters. How particular hylomorphists approach the Problem of Material Constitution depends on how the theorists in question conceive, firstly, of the matter composing a concrete particular object (which was treated in Chapter 2) and, secondly, of the relation between a hylomorphic compound and its matter. The Grounding Problem will be considered in connection with our discussion of the compound-form relation in Chapter 4. The question of unity, or how (if at all) the presence of form in a matter–form compound helps to account for the degree of unity that is manifested by these objects, is taken up separately in Chapter 7. 12   The eighth set of desiderata and decision points (see Section 1.6: “The Grounding Problem”) invites hylomorphists to clarify their stance concerning the explanatory work that is to be done by their two central

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  77

3.4.1  Causation, Explanation, and Change In Chapter 1, we noted in connection with the second set of desiderata and decision points (see Section 1.6: “Causation, Explanation, and Change”) that the doctrine of hylomorphism is very much designed to respond to questions concerning causation, explanation, and change involving concrete particular objects, such as the following: how do concrete particular objects come into and go out of existence? How do they persist through change? Why are concrete particular objects capable of persisting through some changes but not others? What causes concrete particular objects to be as they are, once they have come into existence? In this section, I want to ask whether we can conclude anything about the ontological category to which the form of a hylomorphic compound belongs from the causal roles which are assigned to it. In particular, we saw earlier (see item (g), Section 3.3) that, in the debate over the nature of Aristotelian forms, causal considerations are used by proponents of the individual forms hypothesis to argue that their position is superior to that of their opponents, viz., theorists who take Aristotelian forms to be universal or general entities of some kind. Although I will side with a particular version of the individual forms hypothesis for other reasons (viz., reasons having to do with cross-world identity), I am not convinced that the role of forms as causes can be used as a tie-breaker in the debate between individual and universal forms theorists. For the sake of concreteness, I will illustrate this point by means of the causal roles ascribed to form by Aristotle; contemporary hylomorphists who wish to depart from the Aristotelian account are invited to adjust the details accordingly.13 The relationship between the soul and the body of a living organism, as it is described in De Anima and elsewhere, provides us with a detailed and instructive example of how Aristotle distributes the causal responsibilities within the compound between the matter and the form. For a living organism to be what it is, so Aristotle tells us (see DA II.4, 415b13), is for it to live a certain kind of life, namely, one that is characteristic of the kind of organism at issue.14 Whatever is the cause of an organism’s living as it characteristically pieces of apparatus, viz., form and matter, in accounting for the modal profile of matter–form compounds. We will find, in connection with our discussion of the compound-form relation in Chapter  4, that the Grounding Problem provides additional considerations by means of which to adjudicate between the range of conceptions of form outlined earlier in Section 3.2. Furthermore, the Grounding Problem connects up with items (g) and (e), from Section 3.3, viz., the causal roles Aristotle assigns to forms and their relation to essence. Our discussion of the Grounding Problem will also shed some light on the issues raised by the first set of desiderata and decision points (“Property Possession and the Problem of Universals”), namely, the question of whether the relation(s) between a concrete particular object, its matter and its form should be conceived of as the same as, or different from, the relation of property possession which holds between a property-bearer and its properties. 13   The following paragraphs are extracted from Koslicki (2014), where I examine in more detail the causal priority Aristotle attributes to form, in its role as cause and principle, over the matter and the compound. 14   It follows immediately from this correlation between an organism’s being what it is and its living a certain characteristic life that there can be no such thing, for example, as a dead human being. For nothing that is dead could have the same essence as a human being, in Aristotle’s view, since the definition which states the essence of, or what it is to be, a human being requires that organisms of this kind live a certain kind of life, viz., one that is characteristically human.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

78  form does therefore is also the cause of its being what it is. The majority of this causal work, Aristotle argues, is shouldered by the form of a living organism, viz., its soul, which acts as the formal, final, and efficient cause of the living organism (see DA II.4, 415b8–28). The matter of the living organism, in contrast, viz., its body, contributes only by way of its role as the material cause of the living organism. We can appreciate the soul’s role as the formal cause of the living organism by inspecting Aristotle’s method of defining what it is to belong to the broad categories into which he divides living organisms (viz., plants, non-human animals, and human beings). In each case, we notice that the corresponding definitions, which state the essence of, or what it is to be, a living organism of the kind in question, make reference to the type of soul the living organism has. More specifically, when Aristotle defines what it is to be a plant, a non-human animal, or a human being, he does so by referring to a certain range of activities, and the capacities corresponding to them, which he regards as being characteristically associated with the living organism in question in virtue of the type of soul the living organism has. In the case of plants, Aristotle appeals to nutrition (broadly conceived to include the capacities of reproduction, growth, and decrease); in the case of non-human animals, to perception;15 and, in the case of human beings, to thought.16 To be sure, in each case (with the possible exception of thought), Aristotle holds that the activity in question requires a material basis as well.17 However, insofar as the body enters into these activities, it is relegated to an instrumental role which serves the purpose of carrying out the tasks involved in the organism’s characteristic way of life. The only requirements that are imposed on the matter of which a living organism is composed is that it must be of a suitable type to perform whatever activities are initiated by, and defined by reference to the form, i.e., the soul of the living organism (see DA II.1, 412a28ff; II.4, 415b15–21). To illustrate, to be a human being, according to this conception, is to be a living organism composed of a suitable body in which a human soul is present. What it is for the matter in question to count as “suitable” for the purpose at hand is determined by reference to the relevant form, viz., the human soul. What it is for a soul to be a human soul (as opposed to a soul belonging 15   I here follow the reading offered in Johansen (2012), according to which an animal’s ability to engage in locomotion follows from, but is not included in, the definition which states what it is to be an animal (Johansen (2012), Ch. 12). 16   Aristotle thinks of these capacities and their corresponding activities as arranged in a series, in which each successive member presupposes the member preceding it, so that whatever is able to engage in perception also possesses the capacity for nutrition, while whatever is able to engage in thought also possesses both the capacity for nutrition and the capacity for perception (see DA II.3, 414b28–415a12). 17   In the case of thought, Aristotle agonizes, for theological and other reasons, over the question of whether and to what extent the body is involved as well in carrying out the activities relating to the intellect. If, in the case of human beings, thinking requires imagination [phantasia], then this activity as well involves a bodily component at least indirectly, since imagination, in Aristotle’s view, presupposes perception, which in turn clearly has a material basis. If Aristotle in the end does take the intellect to be separable from the body, then the corresponding definitions would make no reference to matter. Otherwise, the same model applies to thought as to the other activities and capacities which Aristotle clearly recognizes as being realized in matter (viz., nutrition, perception as well as all the other capacities and activities which follow from them, e.g. locomotion, imagination, desire, sleeping, dreaming, etc.).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  79 to a non-human animal or plant) is defined in terms of the characteristic activity in which human beings can engage in virtue of having the kind of soul that they do (viz., thought and all that is presupposed by the organism’s ability to engage in this characteristic activity).18 Next, to understand the soul’s role as the efficient cause of the living organism, it helps to consider first an example from the realm of artifacts. A builder, who in Aristotle’s view acts as the efficient cause in the construction of a house, initiates a series of changes which (if successful) lead to the end result that a house form comes to be present in some suitable materials (e.g., bricks and stones). In this way, something that is actually a house (viz., the newly created house) is produced by the efficient cause (viz., the builder) from something that is potentially a house (viz., the matter that is suitable to compose a house). Throughout this process, however, the builder is guided by a house form which is already present in her mind to begin with and which is like the house form that comes to be present in the bricks and stones as a result of the changes implemented by the builder. Thus, in a successful case of efficient causation involving the production of an artifact, before a new matter–form compound can be produced, the efficient cause which produces it must already be in possession of a form that is of the same kind as the form which comes to be present in the matter as a result of the changes initiated and implemented by the efficient cause. Aristotle conceives of the soul as the efficient cause of the living organism in an analogous manner. To illustrate, consider a case in which a living organism grows by adding more flesh to its body as a result of ingesting and digesting food. The role played by the soul in this case, in Aristotle’s view, is comparable to that played by the builder in the construction of a house. As the efficient cause, the soul initiates and implements a series of changes both in the living organism and in the food that is ingested and digested; these changes, if successful, lead to the outcome that something that is actually flesh (viz., the flesh that is added to the living organism’s body) comes to be from something that is potentially flesh (viz., the food that is ingested and digested) by the agency of the soul. In the process, the soul makes it the case that a flesh form which is like the flesh form that is already present in the living organism to begin with comes to  be present in the properly transformed matter (viz., the concocted food). Here as well, a successful case of efficient causation presupposes a form which is already present in the living organism and which acts as a “blueprint,” so to speak, for the soul’s implementation of whatever changes are involved in a living organism’s exercise of its capacity for nutrition.19 18   The question of whether or to what extent the matter composing a living organism should be mentioned in its definition, i.e., the statement of its essence, has been discussed extensively among Aristotle scholars (see, for example, Lennox (2008); Peramatzis (2011)). We will return briefly to this complex set of issues in Section 3.4.2. 19   I have illustrated the soul’s role as the efficient cause of the living organism here by means of a particular case, viz., an exercise of the living organism’s capacity for nutrition. When we attempt to work out how the soul is supposed to act as the efficient cause (or at least an efficient cause) in other cases, e.g., perception and thought, the details become quite tricky (see Johansen (2012), Ch. 7, for useful discussion).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

80  form Lastly, in order to articulate Aristotle’s reasons for classifying the soul as the final cause of the living organism, it is useful to bear in mind the role of the human soul in Aristotle’s specification of the good life for humans, as it is set out in his ethical treatises. Aristotle argues in the context of his well-known “function argument” (NE I.7) that the telos or good for human beings resides in living a certain kind of life, viz., one in which the characteristic activity or function [ergon] for human beings is carried out well. In response to the question of what the function or characteristic activity for human beings might be, Aristotle reasons that it cannot consist in the exercise of a capacity that is exhibited by other living organisms as well: nutrition is common to all living organisms and perception is found in non-human animals as well. Rather, he proposes that the good life for humans must involve the exercise of a capacity that is unique to humans, viz., the capacity for thinking.20 Aristotle’s specification of the good life for humans thus proceeds directly by way of the human soul and its associated characteristic capacities and activities. In this way, the human soul sets the telos for human beings and acts as their final cause. The human body, in contrast, in its instrumental role, carries out whatever tasks are involved in living a characteristically human life. In the preceding paragraphs, I have described the soul’s causal responsibilities as the formal, efficient, and final cause of the living organism in completely general terms. On the basis of what has been said up to this point, the universal forms theorist might thus be inclined to reason that apparently no reference to an individual or particular explanatory principle is needed in order to capture the causal work that is assigned to the form of a hylomorphic compound according to the Aristotelian picture. Proponents of the individual forms hypothesis, however, will object that they have trouble seeing how a form, if construed as a universal, could act as a cause of a particular matter–form compound’s engaging in some particular activity at a particular time and place. Shields (2010c), for example, gives voice to this concern in connection with the soul’s role as the efficient cause of a living organism: If the soul is, for instance, the efficient cause of motion (De Anima i 1, 403a24–b9; ii 4, 415b8, Partibus Animalium i 1, 641a17–b10), then it seems responsible for the motion of an individual body in a specific place and time. Much the same can be said for perception and thinking, where the perceptual and intellectual activities of Socrates seem to be the activities of an Nevertheless, Aristotle’s claim that the soul is the efficient cause of the living organism is meant to be completely general. 20   For further exploration of this crucially important and difficult material, see, for example, the detailed examination of Aristotle’s function argument in NE I.7 in Barney (2008). For the role of thinking as a ­characteristic activity which distinguishes human beings from other living organisms, see Frede (2008). Relevant also in this context is Aristotle’s conception of the divine intellect in Met. Λ as well as his discussion of the active intellect in DA III. Such texts as Met. A.1–2, as well as the emphasis on theoretical contemplation in NE X.7–10, suggest that human beings, for Aristotle, are at their best when the activities on which their life centers most closely resemble those of the divine intellect. For further discussion of the divine intellect, see, for example, Burnyeat (2008).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  81 individual, situated in a peculiar place and time, and explained not by the presence of a uniform universal, but by dint of there being individual psychological faculties deployed on particular occasions. In these ways, it seems natural to understand Socrates’ soul to be numerically and qualitatively distinct from Callias’ soul. So it also seems natural to think of their souls as particulars, rather than as a single universal twice betokened.  (Shields (2010c))

Shields is of course correct to point out that, if we are wondering what the efficient cause is for some particular living organism’s engaging in some particular life activity at a particular place and time, an adequate answer to the corresponding “why” questions would not simply appeal to whatever is common to all members of the kind at issue. Suppose, for example, that we are interested in the efficient cause or causes which help to bring about the particular event consisting in Socrates’ perceiving the cup filled with hemlock when it is handed to him by his executioner. An adequate response to the question of what sets into motion the series of changes required for this specific act of perception would have to take into account the particular features present in this situation, and could not simply rely on the fact that Socrates, like all members of the species to which he belongs, is able to exercise the capacity of perception. At the same time, regardless of how the universal forms theorist decides to spell out the details of his account, we would expect him to appeal at the very least to the following two features: firstly, that it is the particular instances of the universals exemplified in the situation at hand which are causally active; and, secondly, that these particular instances interact causally in the way that they do because they instantiate some general causal pattern, connection, or law which is exemplified in the situation at hand. To illustrate, if we are wondering, for example why, on a particular occasion, a quantity of salt dissolves in water, it would be natural to make reference to the general fact that salt is soluble in water, together with the particular features instantiated in the situation at hand, e.g., that some particular quantity of salt was immersed in some particular quantity of water and that nothing prevented the general pattern at issue from being exemplified in the situation at hand. Similarly, it is open to a universal forms theorist to point to the universal features and causal connections exemplified by the particular interaction between Socrates’ well-functioning perceptual apparatus, the salient objects within his visual field, and the favorable conditions which obtain in the situation at hand. Thus, unless we bring to bear principled objections to an approach to causation of the style appealed to just now, it does strike me that the universal forms theorist can avail himself of the apparatus necessary to account for the idea that universals are causally active in a particular situation by way of their instances, together with the general causal patterns, connections, or laws which are exemplified in the particular situation at hand. And although much more would need to be said in a fuller treatment of the issues at hand, I have nevertheless tried to indicate briefly why I do not regard the causal roles and priority assigned to the form of a hylomorphic compound in the Aristotelian account as a decisive consideration in favor of the individual forms hypothesis.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

82  form

3.4.2  Essence and Accident What is part of the essence of a concrete particular object and what is merely accidental to it? And how is this contrast to be drawn? Hylomorphists tend to place high priority on their ability to capture the contrast between essence and accident, but they do not all approach this distinction in the same way. As noted earlier in connection with the third set of desiderata and decision points (see Section 1.6: “Essence and Accident”), one of the central questions hylomorphists encounter in this area is whether to adopt a modal or a non-modal conception of essence. In what follows, I will be recommending that it is in the interest of hylomorphists to opt for a non-modal conception of essence, especially for the purposes of providing a response to the Grounding Problem, among other things. In addition, however, hylomorphists face a second set of questions which also surfaced earlier in connection with Aristotle’s alleged identification of form and essence (see item (e), Section 3.3): Is the essence of a matter–form compound identical to its form? Or does the matter composing a concrete particular object also figure into its essence in some way, if only generically? Given the overall aims of this chapter, I am particularly interested here in the question of whether a clarified understanding of the relationship between form and essence can help us advance further in the debate over the ontological category to which forms should be assigned. As in the case of the causal-explanatory considerations explored in Section 3.4.1, it is my view that the pressures exerted on hylomorphists in their attempt to develop a particular stance concerning the relationship between form and essence can be resolved in a variety of ways, with the result that this set of desiderata and decision points is neutral with respect to the dispute between the universal forms and the individual forms hypothesis. Four possible combinations of views concerning the relationship between form and essence are available to hylomorphists, which I will abbreviate as follows: (i) Universal Form = Essence; (ii) Universal Form ≠ Essence; (iii) Individual Form = Essence; and (iv) Individual Form ≠ Essence. The first two positions proceed on the assumption that forms are universal or general entities, while the second two positions hold some version of the individual forms hypothesis. In both cases, it is possible either to identify the form of a hylomorphic compound with its essence (as in (i) and (iii)) or to reject the alleged identification of form with essence (as in (ii) and (iv)), e.g., on the grounds that the matter also in some way figures into the essence of a hylomorphic compound, if only generically.21 Since all four combinations of views are at least in principle coherent, the 21   In this section, for ease of exposition, I will focus only on the possibility that the definition which states the essence of or what it is to be a certain kind of hylomorphic compound may need to make reference to the type of matter or material parts of which entities of the kind in question are (or must be) composed. In addition, however, those who hold that the matter composing a hylomorphic compound in some way figures into its essence, and who therefore reject the identification of essence with form, may also consider the possibility that the particular matter or material parts of which a hylomorphic compound is composed (e.g., the original matter which composed it when it first came into existence) ought to be mentioned in its definition. I will argue against this second position in Section 3.4.3 on the grounds that sameness of original matter does not yield a sufficient condition for the cross-world identity of hylomorphic compounds. Other factors, besides the entity’s form and possibly its matter, may also turn out to be candidates for inclusion

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  83 adoption of a specific commitment concerning the relationship between form and essence does not by itself help us to adjudicate the dispute between the universal and the individual forms hypothesis. Universal Form = Essence. The first position is characterized by the following combination of views: forms are universal or general entities of some sort; the essence of a hylomorphic compound is exhausted by its form; and the matter composing a hylomorphic compound does not directly figure into its essence.22 This position leaves open, but does not require, that constraints concerning the material makeup of the type of entity in question nevertheless in some way follow from its essence, even though the definition (viz., the statement of the essence) does not itself include an explicit reference to matter. To illustrate, a proponent of this first position might hold, for example, that the essence of, or what it is to be, a human being, is correctly stated by the definition: “To be a human being is to be a living organism with a human soul,” where we may assume that what it is for a soul to be a human soul (as opposed to, say, a plant soul or a non-human animal soul) can itself be defined in a non-circular way without mentioning human beings, viz., in terms of the characteristic activities and capacities in which living organisms with a human soul are able to engage. Proponents of the first position would interpret the definition just cited as describing a kind-level universal form/essence that is common to, and instantiated by, all members of the human species. Although this definition makes no overt mention of the types of bodies of which human beings are (or must be) composed, it is still open to the proponent of the first position to hold, for example, that any living organism which is capable of the type of rational thought that is characteristic of human beings (as opposed to, say, angels or God) must also have access to sense perception and that the latter requires a body equipped with sense organs. Universal Form ≠ Essence. The second position (like the first position) takes forms to be universal or general entities of some sort; but (unlike the first position) it holds that the essences of hylomorphic compounds are not exhausted by their universal forms, e.g., on the grounds that the matter composing these hylomorphic compounds also in some way figures into their essences, if only in a generic way. A version of the second position can be found, for example, in Peramatzis (2011), who argues that a sample definition for a natural hylomorphic compound follows the general pattern: “FC =def F1, F2, . . . , Fm enmattered in M1, M2, . . . Mn”, where FC is a natural form of a hylomorphic compound, and F1, F2, . . . , Fm and M1, M2, . . . , Mn are the formal and material parts of the essence stated by the definition (Peramatzis (2011), p. 96). To illustrate, the definition into its essence, depending on the specific case under consideration. For example, in the case of artifacts, as we will see in Chapter 8, some theorists propose that the specific creative intention guiding an artifact’s maker or the specific act of creation resulting in the artifact’s production should also be included in the artifact’s essence. To keep our current discussion at the most general level, however, I will focus here only on form and matter as the most obvious candidates for what might be taken to be part of the essence of a hylomorphic compound.   Versions of this position can be found, for example, in Lewis (1984) and Cohen (2009).

22

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

84  form which states what it is to be human, according to this conception, might take the form, “Human-kind =def rational soul enmattered in a certain kind of organic body” (p. 35), where the type of matter (“a certain kind of organic body”) which composes hylomorphic compounds of the kind in question (viz., the kind, human being) is explicitly mentioned in a statement of their kind level essence.23, 24 Individual Form = Essence. According to the third position, forms are particular or individual entities of some sort; the individual form of a hylomorphic compound is identical to its essence; and the matter does not directly figure into its essence. Like its universalist counterpart, this position allows, but does not require, that constraints concerning the material makeup of the type of hylomorphic compound in question in some way follow from a statement of its essence, even though the definition itself does not explicitly refer to the matter or material parts composing it.25 Given that definitions are presumably formulated using general terms, but forms/essences according to this approach are individual or particular entities, the third position must acknowledge that there is a certain level of mismatch between the (general) content of a definition and the (particular) explanatory principle that is in fact present within each hylomorphic compound.26 As we noted already in connection with the individual formsas-objects approach adopted by Michael Frede and Günther Patzig (see Section 3.2.3), the specification of the form/essence that is given by a definition can be expected to be 23   Peramatzis’ development of the second position has some idiosyncratic features which need not concern us further here. In particular, he holds that a definition which takes the form, “FC =def F1, F2, . . . , Fm enmattered in M1, M2, . . . , Mn,” states the essence of a natural form, FC, i.e., the form of a certain type of natural hylomorphic compound (e.g., the kind, human being). In Peramatzis’ view, this natural form, FC (i.e., the definiendum of the definition in question) itself has a hylomorphic structure, since its essence is given by the definiens which states both essential formal constituents, F1, F2, . . . , Fm, and essential material constituents, M1, M2, . . . , Mn. Thus, Peramatzis is committed to holding that the form of a natural hylomorphic compound, confusingly, is itself “made of matter” (Peramatzis (2011), pp. 53–4). 24   The combination of views characteristic of the second position is also congenial to the neo-Aristotelian mereology developed in Koslicki (2008a). According to the position defended there, mereologically complex material objects to which we are committed as part of our scientifically informed common-sense ontology are unified structured wholes. To state what it is to be a mereologically complex material object of a certain kind, K, one must appeal both (i) to the type of matter or material parts of which objects of kind, K, are composed; and (ii) to the form which dictates how the material parts composing these objects must be arranged or configured in order for these material parts to compose a whole of kind, K. In Koslicki (2008a), I leave open whether, in addition to kind-level forms, neo-Aristotelian mereologists should also take on board individual forms. Below, I endorse a version of the individual forms hypothesis for reasons connected with the cross-world identity of hylomorphic compounds. Thus, the position to which I am now most sympathetic is a version of the fourth position according to which forms are individuals, but the essence of a hylomorphic compound is not exhausted by its individual form, e.g., on the grounds that the type of matter or other factors that are essential to the object under consideration also figure in its essence. 25   Proponents of this position include, for example, Frede (1985, 1987a, 2000); and Frede and Patzig (1988). We might also count Lowe (1999) as an advocate of the third position, though (as I shall bring out shortly) care is needed when we try to subsume Lowe’s position under one of the headings provided in the text. 26   The mismatch in question only arises given the anti-Leibnizian assumption that whatever qualitative content is given in a definition is not by itself sufficient to single out a unique individual form/essence and differentiate it from all the other individual forms/essences that are present in other members of the kind in question.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  85 uniform across all members of a species, even though the forms/essences themselves assumed by the third position are unique to each particular hylomorphic compound. Thus, insofar as a definition can capture what is essential to each hylomorphic compound, it will do so only at the level of kinds, but not at the level of the individual members of these kinds, despite the fact that this approach considers kind-level forms/ essences to be derivative and individual forms/essences to be primary. Individual Form ≠ Essence. The fourth combination of views holds (like the third position) that forms are particular or individual entities of some sort; but (unlike the third position) this approach takes the essence of a hylomorphic compound not to be exhausted by its form, on the grounds that the matter composing a hylomorphic compound also figures into its essence at least in a generic way.27 As in the case of the third position, this approach will likely need to recognize that the (general) content of a definition stating the essence of a hylomorphic compound does not fully capture the uniqueness of the (particular) explanatory principle that is in fact present in each particular specimen of a given kind. Thus, the pattern attributed to kind-level definitions by the fourth position need not diverge from that endorsed by the second position, though the underlying metaphysical commitments will be distinct. The four positions mentioned so far all assume, in one way or another, that essences are themselves entities of some sort, which either are or are not fully exhausted by the (universal or individual) form of a hylomorphic compound. E. J. Lowe has argued that “although all entities have essences, essences themselves should never be thought of as further entities” (Lowe (2008), p. 23), on the basis of the following regress argument: For one thing, if the essence of an entity were just some further entity, then it in turn would have to have an essence of its own and we would be faced with an infinite regress that, at worst, would be vicious and, at best, would appear to make all knowledge of essence impossible for finite minds like ours.  (Lowe (2008), p. 39)

Lowe (channeling Locke and Aristotle) describes the essence of an entity, X, as “what X is,” “what it is to be X,” or “the very identity of X” (Lowe (2008), p. 35), though with respect to the latter phrase we must bear in mind that “identity” here is not to be understood as denoting the relation of identity which everything necessarily bears to itself and nothing else.28 An essence, according to Lowe, may be either general or individual: the general 27   Jennifer Whiting’s interpretation of Aristotle’s hylomorphism can be read as subscribing to either the third or the fourth position (see, for example, Whiting (1984, 1986, 1991, 1992)). 28   I assume that Lowe cautions us against understanding “identity,” in its occurrence in the phrase “the very identity of X,” as denoting the identity relation for the following reason: relations are further entities that are numerically distinct from their relata; therefore, if we were to understand the essence of an entity, X, in terms of X’s bearing the relation of identity to itself, then we would be understanding X’s essence in terms of a further entity besides X itself, viz., the identity relation. This construal of essence would in turn open us up to the threat of the regress of essences outlined by Lowe in the cited passage. We should also keep in mind here that some entities (e.g., tropes), in Lowe’s view, are essentially identity-dependent on other entities numerically distinct from them (viz., the concrete particular objects that are the bearers of these tropes). In these cases, the essence of an entity, X (e.g., a trope), does involve X’s bearing a relation to a further numerically distinct entity, Y (viz., the concrete particular object that is the trope’s bearer), but

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

86  form essence of an entity, X, is what it is to be a K, where K is the ontological category to which X belongs; the individual essence of an entity, X, is what it is to be the individual of kind K that X is (Lowe (2008), p. 35). When Lowe speaks of essences, he has in mind a “serious” notion of essence, which can serve as the ground for metaphysical possibility and necessity, rather than an “ersatz” notion of essence, which presupposes and is defined in terms of an antecedently understood notion of metaphysical possibility and necessity (p. 34). Given Lowe’s approach, then, what is the relation between a concrete particular object, its matter, its form, and its essence?29 Forms, for Lowe, can either be construed as universals or as particulars. When forms are construed as universals, they are the substantial kinds (e.g., humanity, equinity, etc.) instantiated by concrete particular objects (viz., individual human beings, horses, etc.). When forms are construed as ­particulars, they are the particular instances of these substantial kinds; on this latter construal, a concrete particular object (e.g., an individual human being, horse, etc.), in Lowe’s view, simply is, i.e., is numerically identical to—a particularized form (i.e., the particular instance of the substantial kind to which the object belongs, e.g., humanity, equinity, etc.). Insofar as Lowe can make sense of what hylomorphists might mean by “matter,” he understands the matter of a concrete particular object to be its proximate matter or material parts; the material parts of a concrete particular object, in Lowe’s view, are either themselves concrete particular objects (e.g., the material parts of a hydrogen atom are its proton and electron) or portions of stuff (e.g., the material parts of a rubber ball are portions of rubber), which themselves may turn out to be (mereologically) composed of concrete particular objects (e.g., microscopic particles). Given Lowe’s approach, we cannot take the matter composing a concrete particular object, X, to be part of X’s essence, regardless of whether “essence” here is understood not in a way which, in Lowe’s view, gives rise to a regress of essences. For suppose we have characterized the essence of the ontologically dependent item, X (viz., the trope), in terms of the numerically distinct further entity on which it depends, Y (viz., the concrete particular object that is the trope’s bearer). Then we could find ourselves in one of two situations: either (i) Y is itself not ontologically dependent on any other entity numerically distinct from itself; or (ii) Y is itself ontologically dependent on some numerically distinct further entity, Z. In the first case, we are done appealing to numerically distinct entities in the process of specifying the essences of X and Y. In the second case, specifying Y’s essence will require a reference to some further numerically distinct entity, Z. At some point in this process, however, the need to refer to numerically distinct further entities will come to a stop, namely, when we have reached entities which, in Lowe’s view, are not themselves essentially identity dependent on anything numerically distinct from themselves. Thus, Lowe’s own ontology does not fall prey to his regress of essences, despite the fact that he assigns a place within his ontology to ontologically dependent entities which essentially bear certain relations to numerically distinct further entities. We will come back to these issues when we investigate Lowe’s views concerning ontological dependence further in Chapters 5 and 6. 29   See especially Lowe (1998), Ch. 9, “Matter and Form,” pp. 190–203; and Lowe (1999). For his more recent thoughts on the matter–form distinction, see Lowe (2012b). As noted earlier (see Chapter 1, n. 28), in the context of this discussion, we find Lowe distancing himself from hylomorphism, but only insofar as the latter is interpreted as giving rise to a particular kind of constituent ontology. Provided that the matter–form distinction is understood properly, Lowe continues to see an application for this distinction within his own approach, viz., the “four category ontology,” which he views as neither a constituent ontology nor a relational ontology.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  87 as “general essence” or “individual essence”; for, in Lowe’s view, X’s identity at a time (synchronically) or over time (diachronically) is not determined by the material parts composing X. Rather, as the following passage brings out, it is the form of a concrete particular object, in Lowe’s view, which provides both a “principle of individuation” (i.e., a criterion of synchronic identity) and a criterion of diachronic identity for the object in question: Might it then be that matter [. . .] provides a criterion of identity for [individual concrete things]? [. . .] But this idea can certainly be challenged, on the simple grounds that individual concrete things like tigers can and do change their component matter. Indeed, in principle, two tigers could, over a period of time, entirely exchange their component matter with one another. So the most that can be said is that two different tigers cannot share the same matter at the same time. But why not? Simply because that would require the two tigers to exist in exactly the same place at the same time, which we deem to be impossible. However, then it transpires that what really makes for the diversity of our two tigers is their difference in space-time location, from which their difference in component matter at any time merely follows as a consequence. Moreover, that tigers are differentiated by their space-time locations is clearly itself a consequence of their form, since it has to do with what kind of thing they are. [. . ..] This is brought out by the fact that certain (different) kinds of entity apparently can exist in the same place at the same time, such as a rubber ball and the piece of rubber composing it, for this highlights the fact that the mutual exclusion of different tigers from the same place at the same time is simply a consequence of their being things of the same substantial kind. I conclude, then, that matter provides neither a principle of individuation nor a criterion of identity for individual concrete things: their form alone provides both.  (Lowe (2008), pp. 201–2)

As Lowe argues in this passage, the material parts composing a concrete particular object, X, do not provide a diachronic criterion of identity for X, since concrete particular objects can persist through changes with respect to its material parts over time. But X’s material parts also do not provide a “principle of individuation” or synchronic criterion of identity for X, in Lowe’s view, since he allows that numerically distinct objects which belong to distinct substantial kinds (e.g., a rubber ball and the portion of rubber constituting it) can be composed of the same material parts, viz., when they occupy a single region of space at the same time. Instead, so Lowe argues, it is the form, rather than the matter, which provides both a criterion of synchronic identity and a criterion of diachronic identity for concrete particular objects. Given the intimate connection between essence and identity, Lowe’s remarks therefore lead us to believe that the essence of a concrete particular object, in his mind, is exhausted by its form. But forms, on Lowe’s view, are entities: forms, when construed as universals, are substantial kinds; when construed as particulars, they are the individual instances of substantial kinds. We might thus wonder whether the proposed identification of the essence of a concrete particular object with its form would violate Lowe’s injunction cited earlier, according to which the essence of an entity should not be taken to be a further entity, on pain of regress. This worry, however, can be discharged. For Lowe’s regress argument is meant to warn us not so much against taking essences to be

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

88  form entities as such, but rather against taking essences to be further entities, i.e., entities that are numerically distinct from the entities whose essences they are. But a concrete particular object, in Lowe’s view, just is (i.e., numerically identical to) its particularized form. Thus, we can safely interpret Lowe as holding the view that the individual essence of a concrete particular object is exhausted by its particularized form, since the entities in question are one and the same.30 Summary. My aim in this section was a relatively modest one, viz., to illustrate that the identification of essence with form as well as the rejection of this thesis can be represented in an apparently coherent way by both the universal and the individual forms theorist. Therefore, the need to develop a clarified understanding of the relationship between form and essence by itself does not seem to help us narrow down the field of options that are open to hylomorphists when it comes to the ontological category to which forms should be assigned. As it turns out, both universal and individual forms theorists have a way of representing both of the relevant possibilities, viz., the possibility that the essence of a hylomorphic compound is identical to its form, as well as the possibility that the essence of a hylomorphic compound is not exhausted by its form, e.g., on the grounds that the matter composing it also in some way figures into its essence. If forms are particular or individual entities, then the individual essences of hylomorphic compounds are either exhausted by their individual forms or other factors, e.g., the matter composing them, also figure in some way into their individual essences, if only generically. Alternatively, if forms are universal or general entities, then the essences under discussion are generic essences shared by all members of a certain kind. Still, this latter option is also compatible with both possibilities mentioned earlier: either kind essences are purely formal or other factors are also included in the generic essence of a certain kind of hylomorphic compound, e.g., that it be ­suitably enmattered. Thus, to advance further in our attempt to assign forms to their proper ontological category, we must appeal to a different set of considerations besides those already canvased concerning the causal role of forms and the relationship between form and essence. In Section 3.4.3, we will discover that the remaining set of desiderata and decision points which are immediately relevant to our discussion of the ontological status of form (see Section 1.6, “Identity and Indiscernibility”) can be used to provide motivations in favor of a particular combination of views among those ­outlined in previous sections. 30   Recall that, according to the interpretation of Aristotle developed in Peramatzis (2011), the essence of a matter–form compound, X, is its (universal) form, conceived of as a certain kind of property, “a mode or way of being,” associated with X’s membership in the kind to which it belongs: for example, Socrates’ essence or form is to be a human being, while Bucephalus’ essence or form is to be a horse. But forms themselves, in Peramatzis’ view, have essences, which are also modes or ways of being. Does Peramatzis therefore run into the regress of essences outlined by Lowe in the passage just cited? Peramatzis’ method for stopping the regress of essences is to hold that, in the case of forms, the essence of the thing and the thing whose essence it is (viz., the form) are identical (Peramatzis (2011), p. 5). Thus, at most, we might be charged with a misleading way of speaking when we say that forms have essences, since the more proper form of expression would be to describe forms as being (numerically identical to) essences.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  89

3.4.3  Identity and Indiscernibility How (if at all) are questions concerning the (synchronic, diachronic, and cross-world) numerical identity or distinctness of concrete particular objects to be settled?31 Recall, from Section 1.6, that hylomorphists and non-hylomorphists alike face various questions concerning the numerical identity and distinctness of concrete particular objects. A scenario such as that presented in Black (1952) raises the question of how to distinguish between objects that are allegedly numerically distinct but qualitatively indiscernible at a single time. In addition to these questions concerning synchronic distinctness, however, we may also ask what accounts for an object’s synchronic, diachronic, and cross-world identity: what, if anything, makes a single concrete particular object identical to itself synchronically, i.e., at a single time? What, if anything, makes a single concrete particular object that exists at one time diachronically identical to the very same concrete particular object existing at a different time, despite the fact that the object in question may have undergone qualitative change? And what, if anything, makes a single concrete particular object which exists in one world cross-world identical to the very same object existing in another world? Hylomorphists in particular will want to take a stand on whether and how the matter or the form of a matter–form compound plays into facts about its numerical identity or distinctness from other objects. In what follows, I concentrate in particular on the very challenging case of cross-world identity, since it brings into focus the strengths and weaknesses of particular proposals which have been made in this area. When I use the language of possible worlds to describe various scenarios which raise puzzling questions, I do so merely as a convenient shorthand. Those who are uncomfortable with approaching questions of possibility and necessity in terms of a possible worlds framework can reformulate the various scenarios I consider in terms of what is possible or necessary for the entities which centrally figure in these scenarios.32 31   Earlier we mentioned that Aristotle scholars disagree with one another over whether matter–form compounds should be individuated by reference to their matter or their form (see item (d), Section 3.3). These concerns are taken up again in this section in connection with the fourth set of desiderata and decision points (see Section 1.6: “Identity and Indiscernibility”), which challenges hylomorphists to say how (if at all) the form of a hylomorphic compound should figure in an account of the synchronic, diachronic, or crossworld identity and distinctness relations which obtain among these objects. 32   The remainder of this section consists of material that is extracted from Koslicki (2018a). The main topic on which I focus in this section is how to settle questions concerning the cross-world identity of concrete particular objects. One might find it odd or surprising that I now take up the topic of cross-world identity when, up to this point, I have been discussing primarily the question of whether we should conceive of forms as universal or general entities or as particular or individual entities. However, as will become apparent shortly, these two topics are intimately connected, since I will be recommending that those who are already committed to hylomorphism for other reasons are best served by settling questions concerning the cross-world identity of matter–form compounds by appeal to their forms. Proponents of the universal forms hypothesis, however, do not have the option of invoking forms as cross-world identity principles for matter–form compounds, since these theorists hold that the very same form is present in numerically distinct matter–form compounds which belong to a single kind. Thus, a universal forms theorist could not very well answer the question of whether a certain object is numerically identical to Socrates or Callias, for example, by appeal to Socrates’ or Callias’ form, since by hypothesis the universal form associated with

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

90  form Why should we even take seriously questions concerning the cross-world identity and distinctness of concrete particular objects? Some philosophers (e.g., van Inwagen (1985)) claim to have shown that there really is no genuine metaphysical problem about the cross-world identity of concrete particular objects at all (see also Kripke (1971 and 1980)). Such philosophers adopt the attitude that the demand for necessary and sufficient conditions governing the cross-world identity of concrete particular objects really poses no genuine challenge, since we should not expect questions of the form, “Why is an entity, x, in a world, w1, numerically identical to an entity, y, in a distinct world, w2?,” to have an informative answer. (One motivation for such a position might be the conviction that all there is to numerical identity is simply the identity of each entity with itself and nothing could be more unproblematic than that.) To this, I reply that even philosophers who are attracted to this line of argumentation still owe us an answer to such questions as “Could Nixon have been a poached egg?”. This question, as I have just formulated it, makes no explicit appeal to cross-world identity; but we could have asked what I take to be the very same question in a way which does explicitly mention cross-world identity, viz., “Is there a possible world, w, containing a concrete particular object which is a poached egg, such that our actual Nixon is numerically identical to this individual in w?”. If the answer to this question is “No, there is no such (metaphysically) possible world and no such (metaphysically) possible concrete particular object which is a poached egg and cross-world identical to our actual Nixon” (or, more simply put, “No, Nixon could not have been a poached egg”), then we still need to understand why a counterfactual scenario involving the supposition that Nixon might have been a poached egg does not in fact manage to describe a genuine (metaphysical) possibility involving the actual concrete particular object, Nixon. Presumably, the answer to the question of why Nixon could not have been a poached egg would in some way appeal to Nixon’s essence in order to bring out what is wrong with the supposition that Nixon could have been a poached egg. Moreover, if the appeal to Nixon’s essence to which the proposed answer refers merely states necessary conditions for the cross-world identity of concrete particular objects, then (as we will discover in what follows) certain tricky questions concerning the numerical identity and distinctness of concrete particular objects will be left unanswered. For we will then not have succeeded in ruling out counterfactual situations in which numerically distinct concrete particular objects are indiscernible with respect to their alleged essences, but the criteria in question do not settle which of these concrete particular objects is numerically identical to Nixon in the actual world. I will therefore, in what follows, proceed under the assumption that we ought to take seriously questions concerning the cross-world identity and distinctness of concrete particular objects.

Socrates is numerically identical to the universal form associated with Callias. Thus, the argument I offer here in favor of designating forms as the cross-world identity principles for matter–form compounds is in effect an argument in favor of (some version of) the individual forms hypothesis.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  91 The following are prominent candidate principles by means of which one might try to settle questions concerning the numerical identity and distinctness of concrete particular objects across worlds: (i) an object’s qualitative character;33 (ii) its matter; (iii) its origins;34 (iv) its haecceity or primitive non-qualitative thisness property (e.g., the property Socrates, and Socrates alone, has of being numerically identical to Socrates); (v) what Plantinga (1974) calls “world-indexed properties” (e.g., the property Socrates, and Socrates alone, has of being Xanthippe’s husband in the actual world); or (vi) its form. As we will see in what follows, the first five of these candidates are problematic in certain ways. In the case of (i), (ii), and (iii), it is doubtful whether the proposed principle in question actually succeeds in providing both necessary and sufficient conditions for the cross-world identity of individuals. In the case of (iv) and (v), the question is not so much whether the proposed principle manages to supply the desired necessary and sufficient conditions for cross-world identity, since haecceities and world-indexed properties are tailored precisely to fit this purpose; rather, the more pressing question in this case is whether the proposed principles accomplish this task in an explanatorily satisfying manner. For these reasons, I will recommend that (vi) is an option we should take very seriously. It is difficult to see, however, how forms could serve as the cross-world identity principles for matter–form compounds, unless these forms are themselves individual or particular entities. The option of invoking forms as the principles of cross-world identity for matter–form compounds therefore seems to carry with it a commitment to the individual forms hypothesis. Once individual forms are on the table, the strategy of settling questions concerning the numerical identity and distinctness of concrete particular objects by appeal to them, while also not entirely problem-free, nevertheless turns out to have various advantages over competing approaches, some of which I will briefly highlight. (A more extended version of the argument that is provided in what follows in an abbreviated fashion can be found in Koslicki (2018a).)35 33   As noted in Chapter 1, it is tricky to say precisely how the distinction between qualitative and nonqualitative properties is to be drawn. For present purposes, I rely on the rough characterization of this distinction given in Adams (1979): “ . . . a property is purely qualitative—a suchness—if and only if it could be expressed, in a language sufficiently rich, without the aid of such referential devices as proper names, proper adjectives and verbs (such as ‘Leibnizian’ and ‘pegasizes’), indexical expressions, and referential uses of definite descriptions” (Adams (1979), p. 7). 34   Below, we will consider a proposal which conceives of (iii), the origins of a concrete particular object, as a special case of (ii), the matter composing a concrete particular object. However, even for entities which are composed of matter, it is in principle possible to think of an entity’s origin in a different way which does not lead to the subsumption of (iii) under (ii). For example, in Chapter 8, we will encounter a type of origin essentialism about artifacts defended by Simon Evnine, according to which it is the act of creation, rather than the original matter, that is taken to be essential to the resulting artifact. It thus makes sense to distinguish (ii) and (iii), even if in some cases, and according to some proposals, instances of (iii) may turn out to be a special case of (ii). 35   See also Whiting (1986) for an argument in favor of individual forms on the basis of considerations concerning the numerical identity of matter–form compounds. Whiting outlines cases in which the same species form is realized in the same matter, even while one concrete particular object goes out of existence and another one comes into existence, thereby indicating that the traditional doctrine which views matter

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

92  form Qualitative Indiscernibility, Sameness of Matter, and Sameness of Origins. I turn first to the question of whether the first three proposed principles of cross-world identity (viz., (i) qualitative indiscernibility, (ii) sameness of matter, and (iii) sameness of origins) can, even jointly, yield sufficient conditions for cross-world identity. To see why they fail to do so, we consider the following scenario described in McKay (1986):36 It seems that two distinct individuals could originate at different times from the same matter configured in exactly the same way. (I am not suggesting that this is likely, merely that it is possible.) Individual O1 might cease to exist, its matter become disorganized, and then its originating matter m might, by chance, come together in exactly the same configuration at t2. A new object O2 would then come into existence at t2. Now find a time t halfway between t1 and t2. Object O1 could have originated at t. Object O2 could have originated at t. These might be facts about the potentialities of O1 and O2 that ought to be respected by a modal system. But the possible situations, O1’s coming into existence at t (without O2 existing at all) and O2’s coming into existence at t (without O1 existing at all) are qualitatively indistinguishable. No features of origin distinguish these two, even though they are clearly distinct objects. (McKay (1986), p. 297)

We can provide an illustration of the kind of scenario McKay has in mind by considering the following course of events. Suppose that in a world, w1, Socrates’ original matter, at the time of his birth in 469 bc, is m. Socrates goes around his merry ways in w1 until his death in 399 bc. Twenty years later, Callias is born in 379 bc in w1 and his original matter in w1 is also m. In the intervening period between 469 bc and 379 bc, the matter which originally composes Socrates at the time of his birth in 469 bc somehow miraculously reconfigures itself into a qualitatively indiscernible arrangement and comes to compose Callias at the time of his birth in 379 bc. Now consider a distinct world, w2, in which a human being comes into existence in 419 bc. This human being is in every conceivable respect qualitatively indistinguishable as the principle of individuation must be abandoned in favor of the position that this role must instead be assigned to individual forms. 36   I will in what follows construe the essentiality of origins thesis, due to Kripke (1971 and 1980), as it applies to composite material objects, as a special case of sameness of matter as a criterion for cross-world identity: instead of considering the matter of which individuals are composed at some later point during their career, the essentiality of origins thesis instead shifts the focus to the original matter of which an entity is composed when it first comes into existence. For example, in the case of artifacts, e.g., a table, the original matter might be the wood which composes the table at the time at which it is created by the craftsperson. The case of organisms is slightly more complicated. A human being, for example, originates from a particular sperm and egg, which together result in a zygote. This zygote then splits into two zygotes, and so on. When we look closely at discussions of the essentiality of origins thesis as it applies to organisms (e.g., Forbes (1985)), we see that what is supposed to be essential to the organism is the matter which is passed on from the sperm and egg to the first zygote, and not, for example, the genetic information encoded in these antecedently existing entities, since the genetic information could be reduplicated in, and hence shared by, numerically distinct organisms, e.g., identical twins. Those who want to deny that the essentiality of origins for composite material objects can be understood in terms of sameness of original matter will, I think, have to appeal instead, implicitly or explicitly, to one of the other principles of cross-world identity I go on to discuss (viz., (iv) haecceities, (v) world-indexed properties, or (vi) forms), when they attempt to settle questions concerning the numerical identity of individuals across worlds.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  93 from Socrates and Callias, as they were at the time of their respective births in w1 in 469 bc and 379 bc. The matter which originally composes this human being in w2, when he comes into existence in 419 bc, is m as well, i.e., the same as the matter which originally composes both Socrates and Callias at the time of their respective births in w1, configured in exactly the same way in which Socrates’ and Callias’ original matter was configured when Socrates and Callias came into existence. The scenario just described immediately confronts us with the question of how we managed to identify the matter, m, in question across times and across worlds. For we assumed just now, firstly, that it is somehow possible to establish a cross-temporal identity within a single world, w1, between Socrates’ original matter in 469 bc and Callias’ original matter in 379 bc. Secondly, we took for granted, in addition, that Socrates’ and Callias’ original matter in w1 can be cross-world identified with the matter which originally composes the human being in w2 when he comes into existence in 419 bc. To help us in this endeavor, I will make the simplifying assumption that we can identify the matter in question with a certain collection of particles, {p1, . . . , pn}, as follows:37 Intra-world Identities for w1: Socrates’ original matter in 469 bc = m = {p1, . . . , pn} Callias’ original matter in 379 bc = m = {p1, . . . , pn} Cross-temporal Identities for Collections of Particles for w1: {p1, …, pn} in 469 bc = {p1, …, pn} in 379 bc Cross-world Identities for Collections of Particles: {p1, …, pn} in w1 = {p1, …, pn} in w2 When we approach the scenario just described with these assumptions in mind, further questions arise. Firstly, the relevant cross-temporal and cross-world identity facts concerning collections of particles must be settled in some way; and, secondly, it must be determined how the particles themselves, which go into these collections, are to be identified across times and across worlds. Perhaps we may presuppose, for present purposes, that this latter question concerning the cross-temporal and cross-world identity of particles is resolved in some way, either because the cross-temporal and cross-world identity of particles may legitimately be taken for granted, or because the

37   According to the simplifying assumption I make here, we suppose (for the sake of the argument) that the matter, m, from which both Socrates and Callias are supposed to originate, can be identified with a collection of particles, {p1, . . . , pn}. The notation I use suggests that the collection of particles, {p1, . . . , pn}, is a set. But “m” could just as well be taken to be a term of plural reference which denotes a plurality of particles, p1, . . . , pn, instead. Nothing I say in what follows hangs on whether we prefer one of these options over the other.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

94  form demand for necessary and sufficient conditions for their cross-temporal and cross-world identity can be met in some fashion.38 Now, with these assumptions in place, we may extract the following results from the scenario outlined earlier. Unless we want to be committed to the position (which I take to be highly implausible) that the time at which an individual comes into existence is essential to that individual, we should allow that Socrates might have been born fifty years later than when he was born in w1, and similarly that Callias might have been born forty years earlier than when he was born in w1. In that case, both Socrates and Callias seem to have an equally good claim at being cross-world identified with the human being who comes into existence in 419 bc in w2 and is originally composed of {p1, . . . , pn}. Given that Socrates and Callias are obviously numerically distinct in w1, they cannot both be identified with a single individual in w2, since doing so would ­constitute a violation of the transitivity of identity. If this scenario describes a genuine possibility, however unlikely in practical terms, then we ought to conclude that, even together, the first three contenders described earlier (viz., (i) qualitative indiscernibility, (ii) sameness of matter, and (iii) sameness of origins) do not yield sufficient conditions for cross-world identity. (iii) is satisfied in the specific version of McKay’s scenario just described, since the original matter was assumed to be the same in each case, viz., a certain collection of particles {p1, . . . , pn}. Moreover, the second criterion is satisfied as well in this case, since (iii) sameness of original ­matter is just a special case of (ii) sameness of (original or non-original) matter. Lastly, since the collection of particles in question was said to be configured in exactly the same way each time it originally composes one of the individuals in question, we are thus presumably dealing with individuals who are qualitatively indiscernible at the time of their respective births. Therefore, unless our conception of qualitative indiscernibility, sameness of matter, and sameness of origins can be amended somehow to recognize a difference in these circumstances, McKay’s scenario by itself establishes that even the combined force of the three proposed criteria together will not answer all tricky questions

38   We have now assumed that questions concerning the numerical identity of the matter, m, from which both Socrates and Callias are supposed to have originated, need not concern us further here, either because m’s cross-temporal and cross-world identity can somehow be settled or because it can legitimately be taken as basic. Although, following my simplifying assumption, I speak of m as a collection of particles, {p1, . . . , pn}, we could equally well have taken m to be a portion of prime matter, following the Thomistic prime matter hypothesis discussed in Chapter 2. As we noted in Section 2.3.2, Brower (unlike Oderberg) assumes that portions of prime matter are the primitive source by appeal to which questions concerning the numerical identity and distinctness of matter–form compounds are to be ­settled derivatively. Thus, despite my “un-Thomistic” formulation of the scenario currently under discussion, it can nevertheless be rephrased in terms which would be congenial to prime matter theorists. In that case, however, we should note that McKay’s argument can also be used to show that primitively individuated portions of prime matter alone do not suffice to settle all remaining questions concerning the numerical identity and distinctness of matter–form compounds. For Brower’s most recent thoughts on these topics, see Brower (2017).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  95 which arise concerning the numerical identity and distinctness of concrete particular objects.39, 40, 41 Haecceities. Haecceities (“thisnesses”) are typically taken to be primitive nonqualitative identity properties.42 For example, on this construal, Socrates’ haecceity (if he has one) is the property of being numerically identical to Socrates. Similarly, for each individual, its haecceity (if it has one) is the property of being numerically identical to that individual. A haecceity must be distinguished from the property of being selfidentical, since this latter property is shared by all entities. Since numerical identity is a one-to-one relation, each haecceity can only be exemplified by exactly one individual at particular times, across times, and across worlds. For example, no one other than 39   In response to McKay’s objection involving the recycling of an organism’s original matter, Forbes c­ onsiders adopting a conception of cross-world identity which would count such properties as being first or being second in the order of creation as identity-relevant properties (see Forbes (1997 and 2002)). But this move faces the same difficulties as those encountered by best-candidate accounts of personal identity, according to which the question of whether an individual is identical to itself across worlds is apparently turned into an extrinsic matter, sensitive to what goes on with numerically distinct individuals. 40   Even if the criteria considered so far do not succeed in yielding sufficient conditions for cross-world identity, essentialists may nevertheless point out in their defense that some of these criteria still at least provide necessary conditions for cross-world identity, in particular indiscernibility with respect to qualitative essential properties and sameness of origins. (Sameness of matter at later points during an object’s career is in many cases implausible as a necessary condition for cross-world identity, if we want to allow that objects can persist through changes with respect to the matter which composes them at some later point during their careers.) If indiscernibility with respect to qualitative essential properties and sameness of original matter at least provide necessary conditions for cross-world identity, essentialists will have made some progress in avoiding cross-world identity judgments which are apparently completely ungrounded in anything other than facts about the numerical identity of the individuals involved in actual and possible situations. Still, a Quinean critic, who challenges essentialists to provide principles of cross-world identity that are both necessary and sufficient, will not be completely satisfied until both halves of his or her demand have been met. 41   Given that we have allowed an in-principle distinction between (ii) sameness of matter and (iii) sameness of origins, one might wonder whether a response could be formulated to McKay’s challenge which does not turn on (iii)’s being a special case of (ii). One such possibility would be to conceive of an object’s origin in terms of the event which led to the creation of the object in question. In this vein, as noted earlier, Evnine (2016a), for example, develops a version of origin essentialism for artifacts according to which an artifact is essentially the product of a certain act of creation in which the artifact’s maker successfully exercises his or her creative intention to produce a certain thing. Evnine’s account, which will be examined further in Chapter 8, turns out to give rise to serious difficulties concerning the individuation of artifacts. But even if an account of this type could be made to work for artifacts, one wonders how it could be successfully extended to members of natural kinds: for, in that case, some analogue of an “act of creation” would have to be found that does not make reference to the intentions guiding the artifact’s maker during the process of production. Since we have already assumed that Socrates and Callias could have come into existence slightly earlier or later than they in fact did, the time at which the “creation event” occurred is not plausibly taken to be part of the essence of the resulting object. I assume that the place at which the “creation event” occurred can be ruled out for similar reasons. But if neither the time nor the place at which the “creation event” occurred are part of the essence of the resulting object, and the participating original matter is not sufficiently fine-grained to yield a satisfactory principle of cross-world identity, then it is difficult to see what other options there are (besides those that will be considered in the remainder of this section) to generate a successful response to McKay’s challenge. (For further discussion concerning the individuation of events, see, for example, Diekemper (2009, 2015).) 42   Historically, Duns Scotus is credited with the introduction of haecceities into our philosophical discourse; see, for example, Cross (2003) for discussion. See also Rosenkrantz (1993).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

96  form Socrates can instantiate the property of being numerically identical to Socrates at every time and in every world in which Socrates exists; and the same holds for all ­entities whose individual essences include haecceities. There is no question, then, that haecceities provide necessary and sufficient conditions for the cross-world identity of those individuals with which haecceities are associated: an individual’s haecceity picks out precisely that individual and no other, whenever and wherever that individual exists. But the worry does arise as to whether an appeal to haecceities settles questions concerning the numerical identity and distinctness of concrete particular objects in an explanatorily satisfactory way: for one might feel that haecceities have been invented by philosophers for no other reason than precisely to resolve the sorts of puzzles we considered earlier which apparently cannot be put to rest in any other way. Ideally, questions of the form, “Why is entity, x, in a world, w1, numerically identical to an entity, y, in a distinct world, w2?,” should not be given a purely stipulative answer, “It just is, because we said so.” The trouble with haecceities is that they have just such a postulated air about them. According to the haecceitistic strategy, whether x in w1 is identical to y in w2 turns on whether the haecceity exemplified by x in w1 is numerically identical to the haecceity exemplified by y in w2. And to the question, “But why is the haecceity exemplified by x in w1 numerically identical to, or distinct from, the haecceity exemplified by y in w2?,” no further answer is forthcoming, since facts about the numerical identity or distinctness of haecceities are to be accepted as basic. Moreover, since haecceities are also non-qualitative, there is no hope of being able to point to some other feature which does not already presuppose the numerical identity or distinctness of the object involved and which accompanies the presence of a certain haecceity in the object in question. In order for haecceities to do their intended job, we must assume that they satisfy certain requirements. For example, if Socrates’ haecceity is to settle questions concerning Socrates’ numerical identity in all possible situations, it must be the case that Socrates’ haecceity is numerically distinct from the haecceity of every other object numerically distinct from Socrates. That is, in every world in which Socrates’ haecceity is exemplified at all, it must be the case that Socrates’ haecceity is only ever exemplified by Socrates. Furthermore, it cannot be the case that what explains why only Socrates can ever exemplify Socrates’ haecceity, and why it is impossible for, say, Callias to exemplify Socrates’ haecceity, is that Socrates is numerically distinct from Callias; for the explanation is supposed to proceed in the other direction, viz., the distinctness of Socrates’ and Callias’ haecceities is meant to explain the fact that Socrates is numerically distinct from Callias. Since haecceities are taken to be primitive nonqualitative identity properties, we cannot ever expect to be given a further reason of any sort, qualitative or otherwise, as to why Socrates’ haecceity is numerically distinct from Callias’ haecceity and from that of any other object numerically distinct from Socrates. One might thus come away feeling that haecceities are simply stipulated to satisfy whatever requirements they need to satisfy in order to resolve questions

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  97 concerning the numerical identity of those objects with which these haecceities are said to be affiliated. The complaints against haecceities just voiced may appear to be motivated at least in part by their role as explanatory primitives, rather than by anything that is peculiar to haecceities as primitive non-qualitative identity properties. To be sure, it is sometimes difficult, if only for psychological reasons, to resist the temptation to ask for further explanation in the face of being told that one has reached a basic level beyond which no further explanations are available. But we may still press those who favor haecceities as their preferred principles of cross-world identity to justify that their particular choice of explanatory primitive really locates the point at which all explanation is supposed to come to a halt in a good place. Here is another way of bringing out why one might think that the stipulativity worry just raised does pose a serious challenge after all for the defender of haecceities which turns specifically on the non-qualitative character of haecceities, rather than their role as explanatory primitives. Consider the following question which gives rise to what I will refer to as the “poached egg” objection: what, so we might ask the haecceitist, prevents Socrates’ haecceity, say, from being exemplified by a poached egg? Since anything which exemplifies Socrates’ haecceity is numerically identical to Socrates, a situation in which Socrates’ haecceity is exemplified by a poached egg is, eo ipso, a situation in which Socrates is (numerically identical to) a poached egg. Thus, so the objection goes, it appears as though a commitment to haecceities provides truth-makers for such apparently implausible modal judgments as “Socrates might have been a poached egg.” Haecceitists may respond to the “poached egg” objection by imposing certain constraints which concrete particular objects must satisfy in order for them to be able to exemplify particular haecceities. For example, the haecceitist might require, among other things, that only a human being can exemplify Socrates’ haecceity. Given the nonqualitative character of haecceities, however, such constraints cannot in any way be motivated by appeal to factors internal to the particular haecceities themselves. There is nothing about Socrates’ haecceity itself which helps to explain why this particular haecceity can only be exemplified by a human being, and not, as it might be, by a poached egg. Any such constraint that is imposed on a concrete particular object in order to insure that it is eligible for the exemplification of a particular haecceity must therefore be an externally imposed necessary connection which is required to hold between the haecceity and its exemplifier. We will see shortly that other cross-world identity principles, which have at least a partially qualitative character, are in this respect preferable to haecceities, since they can do a better job in motivating the necessary constraints which must obtain between the identity principle in question and the object with which it is affiliated. World-Indexed Properties. A further potential resource for essentialists in their quest to resolve difficult questions concerning the numerical identity and distinctness of concrete particular objects is to invoke what Plantinga (1974) calls “world-indexed properties.” According to Plantinga, if P is any property Socrates, and Socrates alone,

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

98  form has in the actual world, α, then having P in α is an individual essence of Socrates, e.g., being married to Xanthippe in α, being the shortest Greek philosopher in α, or being A. E. Taylor’s favorite philosopher in α (see Plantinga (1974), pp. 72ff). What is more, Plantinga’s world-indexed properties apparently succeed in providing necessary and sufficient conditions for the cross-world identity of individuals: in every world in which Socrates exists, Socrates’ world-indexed essences are exemplified by one and only one individual, viz., Socrates. At the same time, it is questionable whether Plantinga-style world-indexed properties meet the demand for a cross-world identity principle in a more explanatorily satisfactory way than our previous contender, viz., haecceities. For one thing, the strategy of settling cross-world identity questions by appeal to world-indexed properties does not account for the cross-world identity of all entities belonging to some specific ontological category (e.g., set, organism, artifact, event, etc.) in a systematic, category-wide manner, since, for each such entity, x, which world-indexed properties qualify as individual essences for x, will depend on which features are unique to x in a given world. For example, if Xanthippe had remarried after Socrates’ death, then, for some world β in which Xanthippe remarries after Socrates’ death, being married to Xanthippe in β does not qualify as a cross-world identity principle for Socrates. But Xanthippe’s and Socrates’ marital status is a contingent feature of these objects in each world in which they exist and not one which we would expect to be terribly central to their numerical identity. In other respects, the appeal to world-indexed properties behaves in ways that are quite similar to the haecceitistic strategy considered just now. In particular the troublesome stipulativity worry, which was voiced earlier in connection with the haecceitistic strategy, also affects the success of the current approach. Suppose the question arises as to whether an individual, S, in a world, w, is numerically identical to Socrates in α. One of the properties which allegedly uniquely identifies Socrates across possible worlds, according to Plantinga, is being married to Xanthippe in α. Call this uniquely identifying world-indexed property, “E.” Now we may ask: does S in w exemplify E? Given the strategy currently under consideration, this question must be answered without presupposing that S in w is identical to, or distinct from, Socrates in α. Otherwise, the appeal to world-indexed individual essences would be circular, since after all whether S in w is identical to, or distinct from, Socrates in α is supposed to be explained by appeal to whether S in w exemplifies E, and not the other way around. Unlike the haecceitistic strategy, the appeal to world-indexed individual essences does set some qualitative constraints on individuals in order for them to be eligible for the exemplification of some particular individual essence. In this case, in order for S in w to exemplify E, S in w would have to be such that in α he is Xanthippe’s only husband. This tells us at least that S in w must be the sort of thing that is capable of entering into the legal relationship of marriage.43 But even if we narrow down the range of entities 43   This restriction is perhaps sufficient to exclude certain categories of entities in the domain of w from being able to exemplify E, e.g., poached eggs as well as all those entities which do not belong to the right

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  99 under consideration from the domain of w to only those which are suited for the role of being Xanthippe’s husband in another world, there is no reason to think this condition would uniquely single out S in w from among the eligible candidates in the domain of w which might be Xanthippe’s husband in another world. Thus, the question of whether S in w exemplifies E has yet to be settled in some way. If, on the one hand, it is settled by brutely declaring that S either does or does not exemplify E, then the appeal to world-indexed individual essences is subject to the stipulativity concern voiced ­earlier in connection with the haecceitistic strategy. If, on the other hand, it is settled by appeal to the fact that S in w is either identical to, or distinct from, Socrates in α, then the strategy under consideration is circular. In this way, while world-indexed properties do provide a criterion which is strictly speaking materially adequate (i.e., necessary and sufficient) for cross-world identity, there is reason to be skeptical as to whether the appeal to world-indexed properties settles questions concerning the cross-world identity of objects in a way that is more explanatorily satisfactory than the haecceitistic approach considered earlier. Forms. In the preceding sections, we have evaluated some of the major options which are available to essentialists in their attempt to settle questions concerning the numerical identity and distinctness of concrete particular objects across possible worlds. We found that the first three proposed principles of cross-world identity (viz., an object’s qualitative character; its matter; and its origins) fail to establish criteria that are both necessary and sufficient for the cross-world identity of individuals, while the fourth and fifth candidates (viz., haecceities and world-indexed properties) are open to various charges of explanatory inadequacy. A hylomorphic analysis of concrete particular objects as compounds of matter and form opens up a further possibility: that the form of a hylomorphic compound might serve as its principle of cross-world identity. If forms are to supply necessary and sufficient conditions for the cross-world identity of concrete particular objects, then facts about the numerical identity of forms must either be taken as primitive or they must themselves be explained by reference to some further cross-world identity principle. Given the sorts of difficulties we have already encountered in our attempts to arrive at non-haecceitistic explanations of facts about the numerical identity of concrete particular objects, I will opt for the first strategy which takes facts about the numerical identity of forms as primitive; however, nothing I say in what follows is incompatible with a non-haecceitistic explanation of facts about the numerical identity of forms, should such a possibility arise. It is difficult to see how forms could serve as cross-world identity principles for hylomorphic compounds, unless these forms are themselves particular or individual entities. The option of invoking forms as principles of cross-world identity for hylomorphic compounds thus seems to carry with it a commitment to the individual forms

kind to be able to enter into the legal relationship of marriage. In this respect, the appeal to world-indexed properties, due to their partially qualitative content, does a somewhat better job in providing a non-stipulative response to the “poached egg” objection considered earlier than the haecceitistic strategy.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

100  form hypothesis. And while contemporary metaphysicians might feel some hesitation in turning to individual forms to settle questions concerning the numerical identity and distinctness of concrete particular objects, we must also bear in mind that, in light of the arguments put forward in the preceding sections, essentialists may at this point simply have run out of other options. For once we rule out an object’s qualitative profile, its matter, its origins, its haecceity, and its world-indexed properties as possible contenders for the role of supplying explanatorily adequate, necessary and sufficient conditions for its cross-world identity, what other candidates remain which could be considered for this office? It seems that, for those who already feel the pull of hylomorphism for other reasons, individual forms are the one explanatory principle left standing which could plausibly be invoked to settle questions concerning the numerical identity and distinctness of concrete particular objects across possible situations. The strategy of invoking forms as the cross-world identity principles for hylomorphic compounds offers a unique combination of explanatory advantages which places it ahead of its competitors. By appeal to individual forms, cross-world identity facts concerning concrete particular objects can be settled in a way that is (1) non-circular, (2) not susceptible to the threat of an infinite regress, (3) complete within its intended domain of matter–form compounds, (4) non-ad hoc, (5) uniform, and (6) non-stipulative. Firstly, given that facts about the numerical identity of forms are distinct from facts about the numerical identity of hylomorphic compounds whose cross-world identity and distinctness are supposed to be explained by appeal to their individual forms, the strategy under consideration does not fall prey to the threat of circularity. Secondly, since we assumed that facts about the numerical identity of forms are to be taken as primitive, the appeal to individual forms also does not give rise to an infinite regress, i.e., a chain of alleged explanations in which questions concerning the cross-world identity of entities belonging to class A are settled by appeal to facts about the cross-world identity of entities belonging to a distinct class B, and so on, ad infinitum. Thirdly, the individual forms strategy will yield answers to questions concerning the cross-world identity of individuals for all those entities with which individual forms are associated, i.e., for all hylomorphic compounds. Fourthly, hylomorphists may point to a whole arsenal of independent reasons, such as those laid out in Koslicki (2008a) and in other parts of the present study, for positing forms as explanatory principles that are associated with matter–form compounds and thus avoid the charge that they are postulating entities simply as an ad hoc measure designed specifically to meet the demand for necessary and sufficient conditions for the cross-world identity of concrete particular objects. Fifthly, the individual forms account also fares well with respect to uniformity considerations, in that it explains facts concerning the numerical identity and distinctness of concrete particular objects across possible situations in a way that is uniform across specific ontological categories. For even though the strategy at issue ascribes to each hylomorphic compound its very own individual form, numerically distinct from that of every other hylomorphic compound, nothing prevents the proponent of individual forms from also recognizing significant similarities between the individual forms that

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

in defense of individual forms  101 are present in hylomorphic compounds belonging to a single species. For example, while Socrates’ individual form is, by hypothesis, numerically distinct from Callias’, their respective individual forms are nevertheless both human forms and therefore can be expected to resemble each other in systematic and principled ways, e.g., in that they enable both Socrates and Callias to engage in growth, nourishment, locomotion, perception, and thought. Thus, hylomorphic compounds which belong to the kind, human being, will have facts concerning their cross-world identity explained in the same way, even though each such particular explanation will make reference to a numerically distinct explanatory principle. These explanations will take the following form: a human being, x, in a world, w1, is identical to a human being, y, in a world, w2, just in case x’s individual form (e.g., x’s human soul) in w1 is identical to y’s individual form (viz., y’s human soul) in w2. Finally, the individual forms account favorably compares to its two closest competitors, viz., the haecceitistic strategy and the appeal to world-indexed individual essences, with respect to the most challenging issue facing proposed cross-world identity principles, viz., the stipulativity worry. We noted earlier that the haecceitistic strategy is especially vulnerable to the stipulativity worry, not only because haecceities are postulated as explanatory primitives, but also, and especially, because their nonqualitative character precludes haecceities from putting to rest questions about any of the other features of the entities with which they are associated besides questions concerning their bare numerical identity. World-indexed properties were seen to fare somewhat better in this respect, since they at least impose some qualitative constraints on the entities which exemplify them. However, in the case of the particular example we considered above (viz., being Xanthippe’s husband in α), the qualitative condition at issue concerned a contingent state of affairs that obtains in another world, α, which turns out to be not particularly helpful in adjudicating, in a non-stipulative and non-circular fashion, the question of whether a given entity in the world under consideration, w, does or does not exemplify the world-indexed property in question. Individual forms constitute an improvement with respect to these stipulativity concerns over both the haecceitistic strategy and the appeal to world-indexed individual essences in at least the following two ways. Firstly, individual forms impose some qualitative constraints on the hylomorphic compounds with which they are associated. Secondly, these qualitative conditions do not merely concern contingent states of affairs which obtain in particular worlds; rather, they apply to the relevant hylomorphic compounds across worlds, since they impact not only their contingent qualitative makeup, but also the features they exhibit as a matter of de re necessity. Consider again the world-indexed property which Socrates, and Socrates alone, has of being Xanthippe’s husband in α. The fact that Socrates has this property does not contribute in any meaningful way to an explanation of some of the other striking features Socrates shares with other typical members of his biological kind, e.g., the ability to think, act, deliberate, perceive, laugh, speak, move, grow, ingest and digest food, procreate, breathe, sleep, dream, or engage in any of the myriad of other activities

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

102  form human beings are typically able to manifest in the course of a lifetime. Socrates’ form, in contrast, as the explanatory principle in virtue of which he is a living human being, is much more suitable to carrying significant weight when faced with the question of why the range of characteristic human behaviors is as much open to Socrates as it is to other members of his species.44 Relatedly, a commitment to individual forms may be further justified on the grounds of theoretical utility, since it can be used to accomplish three important tasks simultaneously: the commitment to individual forms opens up the possibility of settling questions concerning the cross-world identity of hylomorphic compounds, their intra-world synchronic identity, and their intra-world diachronic identity by appeal to a single explanatory principle. The first and second explanatory role for individual forms is endorsed, for example, in Lowe (1999), who takes concrete particular objects (or “individual substances,” as he calls them) to be identical to their individual forms. Thus, questions concerning, for example, Socrates’ synchronic intra-world identity, for Lowe, are answered by appeal to Socrates’ individual form, as are questions concerning his cross-world identity. Michael Frede, e.g., in Frede (1985 and 1987a), as noted earlier, leans on individual forms in his account of the intra-world diachronic identity of hylomorphic compounds. In Frede’s view, when a hylomorphic compound persists over time, despite the fact that its matter or qualitative character may change, what stays numerically the same from one time to another is the hylomorphic compound’s individual form: it is what accounts for the continuity of organization as well as the organism’s disposition to function or behave in certain characteristic ways from one time to another. All of these additional explanatory benefits which come with a commitment to individual forms work together in protecting this account from stipulativity concerns in a way that is not also available to its two closest competitors, viz., the haecceitistic strategy and the appeal to world-indexed individual essences. Summary. My primary focus in this section has been on how questions concerning the cross-world identity of concrete particular objects should be answered. I examined six contenders for the role of cross-world identity principles: (i) an object’s qualitative character; (ii) matter; (iii) origins; (iv) haecceities; (v) world-indexed properties; and (vi) forms. The first three fail to provide conditions that are necessary and sufficient for the cross-world identity of individuals; the fourth and fifth criteria are open to the charge that they do not succeed in settling the relevant questions concerning crossworld identity in an explanatorily adequate fashion. And while the sixth strategy also involves an appeal to some presupposed cross-world identity facts (viz., those concerning individual forms), this option, on the whole, can nevertheless take on a much greater explanatory burden with respect to the hylomorphic compounds in which they are 44   I take it that one of Aristotle’s main projects in the biological treatises is to show how the forms of living organisms are causally, and hence explanatorily, relevant in precisely this way to an organism’s ability to engage in the range of activities that are characteristic of members of its biological kind. However, as I argue in Koslicki (2012b), an appeal to facts about the essences of other related entities will also be required in order to arrive at a plausible derivation of an entity’s de re modal profile (see also Section 4.5).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

conclusion  103 present than the two closest competing cross-world identity principles, viz., haecceities and world-indexed properties. On balance, then, the sixth option deserves to be taken very seriously as a possible response to the demand for necessary and sufficient conditions for the cross-world identity of concrete particular objects, especially by neoAristotelians who are already committed to a hylomorphic analysis of these objects for other reasons. Many of these hylomorphists also accept a non-modal conception of essence and thus face the further difficult task, over and above what is required to meet the challenge concerning cross-world identity, of having to explain an object’s de re modal profile in terms of facts about its essence which are accepted as basic. Haecceities and world-indexed properties are unlikely to be of much help with respect to this second challenge, while the forms of hylomorphic compounds are, in fact, well suited for this purpose.

3.5 Conclusion In this chapter, I considered various proposals which have been outlined in the literature concerning the ontological category to which forms should be assigned. My own view, as I argued in Section 3.4.3, is that individual forms are better able to take on ­certain important explanatory tasks than their most salient competitors. In particular, individual forms are well suited to settle questions concerning the numerical identity and distinctness of matter–form compounds as well as their modal profile, which will be discussed further in Chapter 4. Other issues, e.g., the causal roles ascribed to form (Section  3.4.1) or the relation between form and essence (Section  3.4.2), perhaps ­surprisingly, turned out to be neutral between the individual forms hypothesis and the universal forms hypothesis. In Chapter 4, I now turn to the defense of my preferred conception of forms as “robust” particulars which do not simultaneously bear the same relation to the matter–form compound (essentially) and to the matter composing it (accidentally).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

4 Hylomorphic Relations 4.1  Introductory Remarks Hylomorphists regard concrete particular objects not only as compounds of matter and form but also as the bearers of properties. As noted earlier in connection with the first set of desiderata and decision points (see Section 1.6: “Property Possession”), one therefore wonders whether the relation of property possession, which holds between a concrete particular object and its properties, should be taken to be the same as or different from the hylomorphic relations which hold between a matter–form compound, its form, and its matter. Given the structure that is attributed to matter–form compounds by hylomorphists, there are at least three such hylomorphic relations: (i) the “compound–form relation”—the relation between a matter–form compound and its form; (ii) the “matter–form relation”—the relation between the matter and the form composing a matter–form compound; and (iii) the “matter–compound” relation (i.e., “material constitution”)—the relation between the matter composing a matter– form compound and the compound it composes. In this chapter, I take up the question of how hylomorphists should conceive of these hylomorphic relations which hold between a matter–form compound, its matter, and its form. I do so by considering how hylomorphists should respond to the challenge issued by the Grounding Problem (see the eighth desideratum and decision point ­discussed in Section 1.6). The Grounding Problem asks why an object and its matter (e.g., a statue and the clay that constitutes it) can apparently differ with respect to ­certain of their properties (e.g., the clay’s ability to survive being squashed, as compared to the statue’s inability to do so), even though they are otherwise so much alike. I argue that, in response to the Grounding Problem, hylomorphists should opt for a “robust” construal of form, according to which forms do not simultaneously bear the same relation both to the matter–form compound (essentially) and to the matter composing it (accidentally). Armed with a robust conception of forms, hylomorphists can solve the Grounding Problem by invoking independently motivated differences between numerically distinct spatiotemporally coincident objects. According to my own preferred version of robust hylomorphic pluralism, the differences between numerically distinct spatiotemporally coincident objects are to be explained by appeal to a non-modal conception of essence (Koslicki (2012b)), together with the mereological approach to

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

the grounding problem  105 the relation between a matter–form compound, its matter, and its form, developed in Koslicki (2008a).1

4.2  The Grounding Problem Coincidence theorists, also known as “pluralists” or “multi-thingers,” hold that numerically distinct objects can occupy the very same region of space-time, while “monists” or “one-thingers” deny this claim. Why should we believe that numerically distinct objects can occupy the very same region of space-time? Consider, for example, a statue and the lump of clay that constitutes it. In a typical case, these objects differ, among other things, with respect to their temporal properties: one comes into or goes out of existence before or after the other. But we can certainly imagine a case in which the statue and the lump of clay that constitutes it come into and go out of existence at exactly the same time, as is the case, for example, in Allan Gibbard’s well-known scenario involving Lumpl and Goliath: I make a clay statue of the infant Goliath in two pieces, one the part above the waist and the other the part below the waist. Once I finish the two halves, I stick them together, thereby bringing into existence simultaneously a new piece of clay [“Lumpl”] and a new statue [“Goliath”]. A day later I smash the statue, thereby bringing to an end both statue and piece of clay. The statue and the piece of clay persisted during exactly the same period of time. (Gibbard (1975), p. 191)

Even if the statue and the piece of clay imagined by Gibbard are indiscernible with respect to their temporal properties, they nevertheless appear to differ with respect to other properties. If indeed there are such properties with respect to which coincident objects differ, then Leibniz’s Law dictates that these objects are numerically distinct, even though they occupy the very same region of space-time.2 What are these properties with respect to which coincident objects appear to differ? Prominent examples of such properties which are apparently not shared by coincident objects include the following: de re modal properties (e.g., being essentially or accidentally statue-shaped); sortal or kind properties (e.g., being a statue or being a lump of clay); properties pertaining to their persistence conditions (e.g., being capable of being squashed or not being capable of being squashed); as well as evaluative or other types of properties (e.g., being innovative, valuable, well made, expressing a certain content,   The contents of this chapter largely coincide with Koslicki (2018b).   According to Leibniz’s Law, necessarily for all objects, x and y, if x and y are numerically identical (i.e., x = y), then x and y are qualitatively indiscernible (i.e., they have all the same properties and stand in all the same relations). Leibniz’s Law should be distinguished from the much more controversial converse principle, the Identity of Indiscernibles, according to which necessarily for all objects, x and y, if x and y are qualitatively indiscernible (i.e., have all the same properties and stand in all the same relations), then x and y are numerically identical (i.e., x = y). See Koslicki (2005) for further discussion of the cogency of Leibniz’s Law-style arguments in favor of the numerical distinctness of coincident objects. 1 2

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

106  hylomorphic relations etc.)3. Perhaps some of these apparent differences can be explained in terms of others, so that some of these properties may turn out to be derivative (or more derivative than others), while others may turn out to be basic (or less derivative than others). Karen Bennett calls those properties with respect to which coincident objects can apparently differ “sortalish” properties; and she calls the remaining properties which are apparently shared by coincident objects “non-sortalish” properties (see Bennett (2004), p. 341). Using Bennett’s distinction between sortalish and non-sortalish properties, we can thus state the Grounding Problem as follows: given that coincident objects appear to share all of their non-sortalish properties, what grounds the apparent differences between them with respect to their sortalish properties (Bennett (2004), pp. 341–2)? Although the Grounding Problem has been raised as an objection primarily against pluralists, monists as well must formulate some kind of response to the Grounding Problem. For if, as monists hold, each region of space-time is occupied by only a single object, then we wonder why it is the case that we nevertheless find ourselves attributing apparently contradictory properties to this single object.4 As Bennett points out (pp. 343–4), the distinction between sortalish and nonsortalish properties should not be confused with the distinction between intrinsic and extrinsic properties, since coincident objects share both intrinsic and extrinsic properties; and the sortalish properties with respect to which they appear to differ may also include both intrinsic and extrinsic properties, depending on one’s view concerning these properties. Thus, the Grounding Problem does not ask how coincident objects which are intrinsically alike can apparently differ from each other in other non-intrinsic respects. Bennett considers three types of responses to the Grounding Problem: (i) nonsortalish grounds—attempts to ground the sortalish differences between coincident objects somehow in their non-sortalish properties (e.g., by way of a supervenience principle of some sort); (ii) conceptualism/conventionalism—attempts to attribute the apparent sortalish differences between coincident objects in some way to our concepts or conventions; and (iii) primitivism—the position that objects have whatever sortalish properties they have primitively, and therefore nothing grounds the apparent sortalish differences between coincident objects. (Primitivism, in Bennett’s view, is not so much a solution to the Grounding Problem as it is a denial that there is a genuine question in the vicinity which requires an answer.) Bennett takes the first style of response to be hopeless: given that coincident objects share all the same non-sortalish properties, it is difficult to see how the apparent sortalish differences between them could be grounded in their non-sortalish properties, regardless of the various attempts to the contrary to concoct suitable supervenience principles.5 The second style of response is also unattractive, in Bennett’s view, due to   For illustrations along these lines, see, for example, Fine (2000).   For additional discussion of the Grounding Problem, see, for example, deRosset (2011); Divers (2008); Fine (2008); Jago (2016b); Olson (2001); Sutton (2012); Zimmerman (1995). 5   See, for example, Baker (2000); Rea (1997); Sider (1999b); Zimmerman (1995). 3 4

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

sidelle’s challenge to hylomorphists  107 the questionable commitments it incurs: proponents of the second style of response must either hold that our concepts or conventions possess near-magical creative powers which they do not, in fact, seem to possess; or they must allow that objects can exist with no sortalish properties whatsoever prior to the imposition of such properties by means of our conceptual or conventional activities.6 Since Bennett takes both of these positions to be highly implausible, she urges pluralists not to opt for the second style of response to the Grounding Problem. The third option, viz., primitivism, is the best hope for pluralists, according to Bennett, though this response also comes with substantial costs. The apparent unattractiveness of primitivism can be alleviated, so Bennett argues, if pluralists take on board the following plenitude principle: “ . . . every region of space-time that contains an object at all contains a distinct object for every possible way of distributing ‘essential’ and ‘accidental’ over the non-sortalish properties actually instantiated there” (Bennett (2004), p. 354).7 Bennett argues that, given this plenitude principle, the question of why any particular sortalish profile is instantiated in a given region of space-time can be answered by pointing to the fact that every possible such sortalish profile is instantiated in the same region of space-time.

4.3  Sidelle’s Challenge to Hylomorphists Alan Sidelle has recently posed an interesting challenge to hylomorphists (Sidelle (2014)): depending on how these theorists conceive of the form present in a matter– form compound, hylomorphism either offers no solution to the Grounding Problem; or, if it does respond to the Grounding Problem at all, it does so in a way that is not proprietary to hylomorphism, and hence can be adopted by non-hylomorphic approaches to the metaphysics of concrete particular objects as well. Either way, so Sidelle reasons, hylomorphism fails to add anything distinctive to the dialectical situation pluralists face when attempting to explain what grounds the apparent differences between numerically distinct spatiotemporally coincident objects. To arrive at the first horn of the dilemma (viz., that according to certain conceptions of form, hylomorphism provides no solution to the Grounding Problem), Sidelle ­supposes, for the sake of the argument, that form is construed by hylomorphists as something along the lines of the organization or arrangement that is exhibited by an object’s material parts. Given this supposition, so Sidelle reasons, it is difficult to see why the matter composing a hylomorphic compound does not also exhibit the very same form as the hylomorphic compound itself. After all, while the matter in question composes the hylomorphic compound, both are organized and arranged in the very same manner, for as long as the entities in question are spatiotemporally coincident. At the same time, it is thought to be essential to the hylomorphic compound that its   See, for example, Sidelle (1989, 1992).   See, for example, Yablo (1987, 1992), though Yablo’s motivation for endorsing the plenitude principle is not driven by the desire to provide a response to the Grounding Problem. 6 7

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

108  hylomorphic relations material parts are organized or arranged in the way that is dictated by the form in ­question, but only accidental to the matter composing the hylomorphic compound that these material parts are organized or arranged in this way. But now, so Sidelle argues, we can see that no progress at all has been made toward a distinctively hylomorphic solution to the Grounding Problem, given the conception of form currently under consideration. For the natural next question we want to pose is how it can be that the very same form belongs to a hylomorphic compound essentially, while it belongs to the matter composing it accidentally; and this question presents us with just another instance of the Grounding Problem. We have therefore apparently landed right back where we started and the appeal to form seems not to have contributed ­anything at all to providing a solution to the Grounding Problem. The second horn of the dilemma (viz., that under a different hypothesis, hylomorphism fails to provide a distinctive solution to the Grounding Problem) requires a conception of form that is robust enough to block the problematic move just outlined. According to this more robust conception, it is taken as a brute fact that the form which is essentially present in a hylomorphic compound does not also accidentally belong in the same way to the matter composing it. (This second response, thus, is a version of what we referred to earlier as “primitivism.”) To illustrate, this approach assumes that no further explanation can be given as to why a statue, say, has a statue form present in it, while the clay constituting the statue does not: statues, according to this conception, by their very nature have statue forms present in them; and, since lumps of clay are not statues, lumps of clay by their very nature do not have statue forms present in them. While the account currently under consideration does not fall prey to the previously raised difficulty, it does, in Sidelle’s view, suffer from a different sort of weakness. For since the central thesis of hylomorphism (viz., that concrete particular objects are, in some sense, compounds of matter and form) seems to play no role in generating this second response to the Grounding Problem, the same style of response is also available to theorists who do not share the hylomorphic outlook. Instead of appealing to a brute formal difference between numerically distinct spatiotemporally coincident objects, these theorists can point to a brute difference of some other sort (e.g., a brute sortal difference, a brute difference with respect to persistence conditions, or what have you). Insofar as the second response provides a solution to the Grounding Problem, then, Sidelle considers it not to be proprietary to hylomorphism. In addition, however, since the second response appeals to a brute difference between numerically distinct spatiotemporally coincident objects, Sidelle is skeptical as to whether it even yields a genuine solution to the Grounding Problem at all. For the question at issue was, after all, to say what grounds the differences between numerically distinct spatiotemporally coincident objects, and merely to point to the brute fact that there are such differences without explaining how these differences arise strikes Sidelle as not so much a solution to the Grounding Problem as “a denial that the basic differences between the objects need to be grounded” (Sidelle (2014), p. 402). In any

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

towards a hylomorphic solution to the grounding problem   109 case, so Sidelle concludes, hylomorphists are guilty of false advertising when they claim that their approach to the metaphysics of concrete particular objects is superior to ­others in that it yields a distinctive solution to the Grounding Problem which is not also available to those who do not endorse the basic hylomorphic premise.

4.4  Towards a Hylomorphic Solution to the Grounding Problem In what follows, I argue for a robust hylomorphic pluralist response to Sidelle’s challenge. This approach rejects the non-robust construal of form problematized in the first horn of the dilemma Sidelle poses for hylomorphists, according to which forms lack the robustness necessary to prevent them from simultaneously bearing the same relation both to the matter–form compound (essentially) and the matter composing it (accidentally). Instead, I argue, hylomorphists should embrace the robust construal of form problematized in the second horn of his dilemma. Hylomorphists who adopt a robust construal of form can respond to the remainder of Sidelle’s challenge as follows. First, hylomorphists need not accept the responsibility of having to provide a distinctively hylomorphic solution to the Grounding Problem; rather, their response to the Grounding Problem may turn on features which can be utilized by hylomorphists and non-hylomorphists alike. (In that case, as Sidelle correctly notes, hylomorphists should of course not falsely advertise that their solution to the Grounding Problem is  distinctively hylomorphic, when in fact it turns on features which generalize to other non-hylomorphic accounts as well.) In addition, I argue for a version of robust hylomorphic pluralism which invokes independently motivated essential differences between coincident objects. When conjoined with a non-modal conception of essence, this response to the Grounding Problem outlines a strategy for explaining the derivative modal differences between coincident objects in terms of essential differences between them that are taken as basic by the account in question.8

4.4.1  The Scenario Let’s return once more to Gibbard’s Lumpl and Goliath scenario cited earlier. Gibbard asks us to imagine a case in which a clay statue (“Goliath”) is fashioned by sticking 8   I, myself, am guilty of Sidelle’s charge that hylomorphists engage in false advertising when they claim that their account generates a distinctively hylomorphic solution to the Grounding Problem. I argued in Koslicki (2008a) that a mere difference in formal parts between numerically distinct spatiotemporally coincident objects is sufficient to yield a distinctively hylomorphic solution to the Grounding Problem (Koslicki (2008a), pp. 181ff). I no longer hold this position. Instead, as I propose here, in addition to the independently motivated essential differences between coincident objects, I now believe that a non-modal conception of essence is also needed to formulate a credible response to the Grounding Problem. But hylomorphists and non-hylomorphists alike can help themselves to this strategy; hence, the resulting response to the Grounding Problem is not distinctively hylomorphic. As long as hylomorphists are careful not to engage in false advertising, however, they should not be concerned if the main features of their response to the Grounding Problem can be replicated by non-hylomorphists as well.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

110  hylomorphic relations together two pieces of clay, the part above the waist and the part below the waist; once the two pieces of clay are stuck together, a new piece of clay (“Lumpl”) comes into existence and the statue, Goliath, thereby simultaneously comes into existence as well. A day later, Goliath is smashed to pieces, thereby destroying Lumpl as well.9 Following a hylomorphic analysis of concrete particular objects, Goliath is a matter– form compound consisting of some matter, m, and a form, f. As noted in Chapter 3 (see Section 3.2), Goliath’s form, f, could be construed either as a universal or general entity or as a particular or individual entity, depending on whether we characterize forms in accordance with the universal forms hypothesis or the individual forms hypothesis. I will, in what follows, refer to Goliath as “O” (short for “object”) and use “+” to stand for the relation of hylomorphic composition by means of which m and f together give rise to O. We can then state the hylomorphic claim that Goliath is a ­compound of some matter, m, and a form, f, as follows:10 (1) O = m + f Presumably, hylomorphists will find it attractive to say that Lumpl, in Gibbard’s story, is just (numerically identical to) m, i.e., Goliath’s matter. Further, this scenario involves the following three hylomorphic relations: the ­compound–form relation Goliath bears to its form, f; the matter–form relation Lumpl bears to Goliath’s form, f; and the matter–compound relation Lumpl bears to Goliath. Given the close connection between form and essence (see Section 3.4.2), hylomorphists will want to say that Goliath’s form in some way figures in its essence:11 (2)  f figures in O’s essence. In contrast, it is plausible to hold that Goliath’s form, f, does not figure in Lumpl’s essence, since Lumpl appears to be related to f only accidentally: (3)  f does not figure in m’s essence. 9   Although this particular case happens to involve artifacts, I assume that a case in which a concrete particular object and its matter are spatiotemporally coincident could, in principle, be constructed for ­living or non-living members of natural kinds as well. I thus intend my reasoning in what follows to be completely general and not to turn on any characteristics that might be specific to the case of artifacts. 10   The representation in (1) is an oversimplification, since it ignores the possibility that O can persist through changes with respect to its matter over time. Thus, a more realistic representation of the claim that Goliath is a compound of some matter, m, and a form, f, would have to accommodate such possibilities as the following: Goliath exists at a time, t1, and at a distinct time, t2, and the matter, m1, which composes Goliath at t1 is distinct from the matter, m2, which composes Goliath at t2. For the sake of simplicity, I will in what follows continue to suppress the necessary relativization to time. 11   The term, “figure,” is intentionally vague to allow for different ways of filling in the relation between form and essence. As noted in Section 3.4.2, proponents of the individual and universal forms hypothesis alike have the option of holding either that the essence of a matter–form compound is exhausted by (i.e., identical to) its form or that the essence of a matter–form compound includes but is not exhausted by its form, e.g., on the grounds that the matter composing the compound also in some way figures in its essence, if only generically. Either way a statement of the essence of a matter–form compound will include a reference to its form.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

towards a hylomorphic solution to the grounding problem   111 After all, even though Lumpl and Goliath, in Gibbard’s scenario, come into and go out of existence at the same time, Lumpl nevertheless could have existed in a scenario in which Goliath does not exist and in which Lumpl is not coincident with any other statue. In such a scenario, Lumpl would have existed without being related to Goliath’s form, f, or to Goliath. Lumpl, therefore, is only accidentally related to Goliath’s form, f, and f does not figure in Lumpl’s essence. At this point, an apparent sortalish difference between Lumpl and Goliath has emerged, as is brought out in (2) and (3): while both Lumpl and Goliath stand in some relation to the very same entity, viz., Goliath’s form, f, this form figures only in Goliath’s essence but not in Lumpl’s. The Grounding Problem, in its current incarnation, now asks what (if anything) explains this apparent sortalish difference between Lumpl and Goliath: (4)  Sortalish Difference: In virtue of what is it the case that f figures in O’s essence, but not in m’s essence? Different hylomorphists will produce different answers to (4), depending in part on how they approach the following questions: (5)  Compound–Form Relation: How is a matter–form compound related to its form? (6)  Matter–Form Relation: How is the matter composing a matter–form compound related to the compound’s form? (7)  Matter–Compound Relation: How is the matter composing a matter–form compound related to the compound? (8)  Form: What sort of entity is the form composing a matter–form compound? (9)  Matter: What sort of entity is the matter composing a matter–form compound?

4.4.2  Hylomorphic Monism Hylomorphic monists hold that matter–form compounds and their matter are numerically identical or numerically the same concrete particular object: that is, in their view, the matter–compound relation, at issue in (7), is numerical identity or numerical sameness.12 From the perspective of the monist, therefore, the region of space-time occupied by Lumpl and Goliath, despite appearances to the contrary, is in fact occupied by only a single concrete particular object. As noted earlier, even though monists deny that Gibbard’s case involving Lumpl and Goliath presents us with a genuine case of coincidence, these theorists nevertheless still face an explanatory task of their own: for monists as well must in some way account for our practice, as illustrated in (2) and (3), of attributing apparently contradictory properties to Lumpl and Goliath, despite 12   See Rea (1998), for example, for an approach to the relation between a matter–form compound and its matter which distinguishes between numerical sameness and numerical identity.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

112  hylomorphic relations the fact that, in their view, the region of space-time in question is occupied by only a single concrete particular object. The universalist powers approach defended in Marmodoro (2013), for example, presents us with a version of hylomorphic monism. According to this approach, the pre-existing material ingredients (e.g., the bricks and mortar) from which a hylomorphic compound (e.g., a house) is made cease to exist as a result of the process which produces a new hylomorphic compound. Thus, there is no separate entity or collection of entities (viz., the matter or plurality of material parts composing a newly created hylomorphic compound) which is numerically distinct but spatiotemporally coincident with the newly created hylomorphic compound. Marmodoro’s approach has the benefit of escaping the ontological commitments incurred by pluralists, but only at the cost of diverging dramatically from the persistence conditions we ordinarily attribute to the pre-existing entities which go into the making of a newly created hylomorphic compound. Thus, no brick, for example, can, strictly speaking, survive the building of a house, even when (as we would ordinarily put it) all that apparently happens to the brick, say, is that it comes to be surrounded by other bricks and connected to them with mortar. In such a case, the brick ceases to exist once it becomes “absorbed” into the newly created house, though many of the characteristics which previously inhered in the brick prior to its destruction continue to “live on” in the newly created house.13

4.4.3  Non-Robust Hylomorphic Pluralism Hylomorphic pluralists deny that a matter–form compound and its matter are numerically identical or numerically the same concrete particular object: that is, in their view, numerical identity or numerical sameness are not live options for the matter– compound relation at issue in (7). The Grounding Problem challenges these theorists to explain how Lumpl and Goliath, despite the fact that they are otherwise so much alike, can nevertheless differ with respect to their sortalish profile, as illustrated in (2) and (3). Non-robust hylomorphists construe forms in the non-robust way problematized by the first horn of Sidelle’s dilemma. Given this combination of views, as will become clear shortly, non-robust hylomorphic pluralists make it unnecessarily difficult for themselves to formulate a response to the Grounding Problem. Suppose, for example, that forms are construed in the following non-robust manner: for Goliath to “have” a certain form is simply for Goliath’s material parts to be arranged in a certain way. Of course, if all it takes for an entity to “have” a certain form is for its material parts to be arranged in a certain way, then nothing will prevent Lumpl from 13   Versions of hylomorphic monism can also be found in Brower (2014); Jaworski (2016b); Oderberg (2007); Rea (1998); and Toner (2008). Hylomorphic monists often conceive of the relation between an object and its matter in terms of the distinction between what is actually present and what is merely potentially present in a given region of space-time (see also the distinction between actual and “virtual” parts discussed in connection with Oderberg’s and Brower’s approaches in Section 2.2.3 and Section 2.3.2, respectively).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

towards a hylomorphic solution to the grounding problem   113 also “having” Goliath’s form, f, simply by virtue of being spatiotemporally coincident with Goliath. After all, spatiotemporally coincident objects are composed of the same material parts and their material parts are arranged in the same way. Thus, given this non-robust construal of form, both Lumpl and Goliath count as standing in the very same relation to Goliath’s form, merely by virtue of being spatiotemporally coincident. But now we may ask the question posed in (4): why is it that Goliath’s form, f, figures in Goliath’s essence, but not in Lumpl’s? In general, hylomorphic pluralists would be expected to appeal to a formal difference between Lumpl and Goliath in order to explain their apparent sortalish differences. In this case, however, the relevant apparent sortalish difference between Lumpl and Goliath precisely concerns their relation to Goliath’s form, f. Since non-robust hylomorphic pluralists hold that Lumpl and Goliath both “have” Goliath’s form, f, in virtue of having the same material parts arranged in the same way, it becomes utterly mysterious how an appeal to Goliath’s form, f, could possibly help to explain why Goliath and Lumpl are nevertheless apparently differently related to Goliath’s form, f, as brought out in (2) and (3). The lesson to draw from this characterization of Gibbard’s scenario is not that ­hylomorphic pluralists in general are unable to formulate a solution to the Grounding Problem; but rather that hylomorphic pluralists should reject the assumptions on which the non-robust construal of form and the compound–form relation are based. Where non-robust hylomorphic pluralists go wrong is that their construal suggests a reductive analysis of what it means for a compound to be related to its form in terms of conditions that are so minimal that the matter composing the compound meets these conditions as well, merely by virtue of being spatiotemporally coincident with the compound. With these impoverished resources at their disposal, it is no wonder, then, that non-robust hylomorphic pluralists have trouble explaining the apparent sortalish differences between coincident objects by appeal to their formal differences. As we will see in what follows, robust hylomorphic pluralists not only reject the minimal conditions proposed by their non-robust counterparts; they also reject the very idea that it is necessary at all for hylomorphists to give a reductive analysis of a compound’s relation to its form.14

4.4.4  Robust Hylomorphic Pluralism: A Compositional Approach I endorse the following robust pluralist hylomorphic response to the Grounding Problem which incorporates views I have developed elsewhere (see especially Koslicki (2008a,  2012b)): (i) numerically distinct spatiotemporally coincident objects are 14   Sidelle quite clearly construes non-robust hylomorphic pluralism as a position which, in the course of attempting to generate a response to the Grounding Problem, also goes in for a reductive analysis of a compound’s relation to its form. For example, the title of Section 2 of Sidelle (2014) is “What is it for an object to have a form?” and the question reoccurs numerous times throughout Sidelle’s discussion. As Sidelle recognizes, robust hylomorphic pluralists, by contrast, reject the idea that they are obliged to give a reductive account of a compound’s relation to its form, as part of their response to the Grounding Problem or for any other reason. In their view, after all, no further analysis in more basic terms can or needs to be given of a compound’s relation to its form.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

114  hylomorphic relations f­ ormally distinct, i.e., they have distinct essences; (ii) these basic essential differences between numerically distinct spatiotemporally coincident objects, when conjoined with a non-modal conception of essence, can be used to explain their derivative modal differences; and (iii) the considerations on which (i) and (ii) are based are independently motivated, i.e., based on reasons not directly related to our desire to formulate a hylomorphic response to the Grounding Problem. A response to the Grounding Problem which exhibits these three features still counts as a version of primitivism, since it takes as basic essential differences between numerically distinct spatiotemporally coincident objects. However, since the main features of this account are independently motivated and earn their keep by doing explanatory work, the robust hylomorphic pluralism I propose does not simply deny that the apparent sortalish differences between numerically distinct spatiotemporally coincident objects need to be grounded; rather, it outlines a strategy for explaining the derivative modal differences between numerically distinct spatiotemporally coincident objects in terms of their basic essential differences. In the remainder of Section 4.4.4, I focus on the first component of my response to the Grounding Problem and argue that we have independently motivated reasons for thinking that coincident objects have distinct essences. In Section 4.5, I turn to the second component of my strategy and discuss how an independently motivated nonmodal conception of essence can be used to explain the derivative modal differences between coincident objects. Although other theorists may disagree with the details of  my account, (i)–(iii) are in principle transferable to other hylomorphic or nonhylomorphic versions of robust pluralism as well. Hence, the overall strategy I am proposing does not, and is not intended to, yield a distinctively hylomorphic solution to the Grounding Problem. 4.4.4.1  Robust Mereological Hylomorphic Pluralism: the Basics To begin with, let’s review the basic hylomorphic characterization of Gibbard’s Lumpl and Goliath scenario (see Section 4.4.1: (1)–(3)). Goliath is a compound of some matter, m, and some form, f; Lumpl is Goliath’s matter; Goliath’s form, f, figures in Goliath’s essence, but does not figure in Lumpl’s essence. Next, given pluralism, Lumpl and Goliath, by Leibniz’s Law, are numerically distinct, since they are not indiscernible with respect to all of their properties. Given robust hylomorphic pluralism, Goliath’s relation to its form, f, is not analyzed reductively in terms of conditions that are so ­minimal that Lumpl, merely by virtue of being spatiotemporally coincident with Goliath, meets these conditions as well and therefore counts as standing in the very same relation to f as Goliath. More generally, robust hylomorphic pluralists reject the very idea that a reductive analysis of a matter–form compound’s relation to its form can be given, or needs to be given. In addition, I supplement this basic characterization of robust hylomorphic pluralism with the following further positions. Firstly, following the hylomorphic conception of matter defended in Chapter 2, I take the material parts of matter–form compounds

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

towards a hylomorphic solution to the grounding problem   115 themselves to be matter–form compounds, as long as these material parts are also structured wholes. Secondly, given the individual forms hypothesis I endorsed in Chapter 3, I construe forms as particular or individual entities that are by their very nature not repeatable or sharable among multiple distinct entities by being wholly present in each of them at a single time. Thirdly, according to the thoroughly mereological conception of composition I developed in Koslicki (2008a), both the matter and the form composing a matter–form compound are literally and strictly speaking proper parts of the compound, according to a single relation of proper parthood.15, 16 In response to the questions posed in (5)–(9), this combination of views gives rise to the following position, to which I will refer in what follows as “(RMHP)” (short for “Robust Mereological Hylomorphic Pluralism”): (5’)  Compound–Form Relation: f is a proper part of O; and f figures in O’s essence. (6’)  Matter–Form Relation: m satisfies constraints dictated by f. (7’)  Matter–Compound Relation: m is a proper part of O. (8’)  Form: f is a robust particular which does not simultaneously bear the same relation to O (essentially) and to m (accidentally). (9’)  Matter: m is a matter–form compound (assuming m is a structured whole); m and O are both concrete particular objects and hence belong to the same ontological category.17 15   Very briefly, my argument in Koslicki (2008a), Chapter 7, goes as follows. In order to avoid a proliferation of primitive sui generis notions of parthood and composition, whose characteristics must be stipulatively imposed on them by means of distinct bodies of postulates (Armstrong (1978, 1986, 1989, 1991, 1997); Fine (1999)), I assume a single notion of parthood, at least for the domain of concrete particular objects, which satisfies at least the following minimal formal characteristics: Asymmetry, Transitivity, and Weak Supplementation. (The Weak Supplementation Principle states that an object which has a proper part must have at least another proper part disjoint from the first.) By Leibniz’s Law, a whole is not identical to its proper parts, individually or collectively, since a whole and its proper parts are not indiscernible with respect to all of their properties. Next, it is plausible to think that a matter–form compound has its matter or material components as proper parts. Suppose now that it is possible for a whole (e.g., a statue) to be composed of just a single material component (e.g., a piece of clay). Then, by Weak Supplementation, the whole must have one or more additional proper parts besides its single material component. The best candidates for these additional proper parts within the whole are of course its form or formal components. I conclude that a whole therefore has both its matter or material components and its form or formal components as proper parts, strictly and literally speaking and according to a single relation of proper parthood. 16   One might think that my view is susceptible to a regress argument of the sort familiar from Aristotle’s discussion in Met. Z.17. I disagree and defend my reading of the relevant passage as well as my response to the regress argument in Koslicki (2006). The unity of matter–form compounds is taken up in more detail in Chapter 7. 17   I am focused here on scenarios in which a whole is composed of a single material component. In a scenario in which a whole is composed of more than one material component, its matter in my view is a plurality of material components. In that case, each of these material components, assuming that it is a structured whole, is itself a concrete particular object that is analyzed hylomorphically as a compound of matter and form. As noted earlier, the more precise formulation of the hylomorphic conception of matter includes the following important proviso: the material parts of matter–form compounds are themselves matter–form compounds, unless or until we reach an empirically confirmed level in the compositional hierarchy at which the material parts of matter–form compounds are not themselves structured wholes. The possibility envisaged in the proviso is discussed in more detail in Chapter 2.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

116  hylomorphic relations 4.4.4.2  Essential Differences between Coincident Objects As stated in (5’) and (7’), the compound–form relation and the matter–compound relation overlap, in that both the matter and the form, in my view, are proper parts of the matter–form compound they compose. However, as brought out in (5’), the compound– form relation also includes a further dimension, in addition to the mereological relationship between a matter–form compound and its form. This further dimension concerns the close connection between form and essence: the form, according to my thoroughly mereological conception of composition, is a proper part of the matter–form compound; but, in addition to that, the form also figures in the compound’s essence, either by exhausting (i.e., being identical to) its essence or by being included in its essence. (More on the connection between form and essence shortly.) Contrary to what is assumed by non-robust hylomorphic pluralists, the matter– form relation, at issue in (6’), is not the same relation as the compound–form relation. Rather, the relation borne by the material components composing a whole to the form associated with the matter–form compound in question is that of satisfying the constraints specified by the form. As I have emphasized in previous work, among the primary responsibilities of the form within a matter–form compound is to specify a range of structural requirements which must be satisfied by the material parts composing a matter–form compound. We may think of the form associated with a matter–form compound, on the one hand, as an entity which provides “slots” to be filled by objects of a certain kind, i.e., a sort of “recipe” for how to build wholes of that kind. An object’s material components or matter, on the other hand, may be visualized as what fills the “slots” specified by the object’s form: the compound’s material components are thus, so to speak, the “ingredients” that are called for in the “recipe”; they are the objects which, in a successful case of composition, satisfy the conditions dictated by the object’s form. These structural constraints specified by the object’s form usually result in requirements concerning the type, configuration, and sometimes even the number of material components from which wholes of the kind in question must be composed. In addition to these structural constraints, however, the form also sets other constraints governing the object’s material components: in general, the form is that explanatory principle within a matter–form compound which accounts for the compound’s structure, identity, and unity.18 18   I discuss the structural constraints set by the form of a matter–form compound in Koslicki (2008a). For the form’s role in determining the numerical identity of matter–form compounds, see Section 3.4.3 in this work; and Koslicki (2018a). The unity of matter–form compounds will be taken up in Chapter 7. The three types of constraints I have just identified (i.e., constraints concerning the structure, identity, and unity of a matter–form compound) are not intended to constitute an exhaustive list of all possible types of constraints which can be placed on an object’s material components by the object’s form. One would expect, for example, that specific kinds of objects are governed by further formal constraints which do not generalize across the board. To illustrate, in the case of chemical kinds, for example, the material components composing a molecule, say, must be chemically bonded to one another by sharing electrons; but in the case of artifact kinds, the material components composing an ax, say, must be properly fastened to one another in a different way.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

towards a hylomorphic solution to the grounding problem   117 Given that, as stated in (5’), forms are proper parts of the matter–form compounds with which they are associated, the logical properties of parthood and composition, in my view, legislate that matter–form compounds cannot be taken to be numerically identical to their individual forms: contrary to the composition-as-identity model, I hold that wholes are numerically distinct from their proper parts, individually or collectively.19 The present approach, therefore, allows us to take a stand against identity theorists who hold that the compound–form relation at issue in (5) is numerical identity.20 We noted in Chapter 3 (see Section 3.4.2) that various options are open to both universal and individual forms theorists when it comes to the development of a clarified understanding concerning the relation between form and essence. Both approaches are able to represent the possibility that the essence of a hylomorphic compound is identical to its form as well as the possibility that the essence of a hylomorphic compound is not exhausted by its form, e.g., on the grounds that the matter composing it also in some way figures into its essence, if only generically. If forms are particular or individual entities, then a compound’s individual essence is either exhausted by (i.e., is identical to) its individual form or it includes but is not exhausted by the compound’s individual form, e.g., on the grounds that the matter also in some way figures in its individual essence, if only generically. Alternatively, if forms are universal or general entities, then the essences under discussion are generic essences shared by all members of a certain kind. Still, this latter option is also compatible with both of the possibilities just mentioned: either kind essences are purely formal or, e.g., it is also part of the generic essence of a certain kind of hylomorphic compound, that it is suitably enmattered. Either way, forms and essences are intimately linked, since the essence of a matter– form compound is either exhausted by or includes its form.21 According to RMHP, Goliath’s form, f, is a proper part of Goliath. Even though Lumpl is coincident with an object, viz., Goliath, which has f as a proper part, f is not therefore also a proper part of Lumpl. Recall that robust hylomorphic pluralists reject the idea that they are under any obligation to give a reductive analysis of a compound’s 19   For diverging perspectives concerning the composition-as-identity model, see, for example, Lewis (1991); Cotnoir and Baxter (2014). 20   See, for example, Frede (1985, 1987a); Frede and Patzig (1988); Lowe (1999); Whiting (1984, 1986, 1991, 1992). 21   Beyond that, additional commitments concerning the precise specification of the essence of a matter– form compound require delving more deeply into what it means to be an object of a specific kind. To illustrate, for the specific case of artifacts, say, to state exactly what goes into the specification of an artifact’s essence requires a commitment to a particular essentialist account of artifacts. But those who have proposed such accounts disagree with one another as to what exactly belongs to a proper specification of the essence of an artifact (e.g., its function, the particular creative intention guiding the artifact’s maker, the artifact’s original matter, the particular creative act in which the artifact’s maker successfully exercised his or her creative intention, etc.). In addition, in the case of artifacts, we must confront the more general question as to whether a coherent and plausible essentialist account is even possible or appropriate, given that artifacts seem to be in certain ways dependent on human minds, interests, practices, and so forth. For further discussion of the particular case of artifacts, see, for example, Baker (2007); Evnine (2016b); Thomasson (2003b); and Chapter 8 in this work.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

118  hylomorphic relations relation to its form. Thus, the fact that Lumpl occupies the very same region of space-time as an object, viz., Goliath, that has Goliath’s form, f, as a proper part, is not a reason for thinking that Lumpl must therefore also have f as a proper part. Although I have adduced general philosophical reasons in my previous work for thinking that a matter– form compound has its form as a proper part, these arguments are not offered in the spirit of giving a reductive analysis of the relation between a matter–form compound and its form. Rather, as a proponent of RMHP, I maintain that it is a basic fact about Lumpl and Goliath that they differ with respect to their formal parts. Turning now to (8’), we can use the basic outlines of RMHP to bring us one step closer to the larger goal of arriving at a clarified understanding of what sorts of entities forms are. Given the robust construal of forms, we know that forms cannot be reductively analyzed in terms of a set of minimal conditions which are simultaneously met by the compound’s matter, merely in virtue of the fact that it materially composes the compound. These further constraints that are placed on our conception of forms by the robust construal are not specific enough to single out a particular choice among the range of ontological categories available to individual forms theorists. But they do narrow down our options to some extent by removing from our consideration approaches, according to which, forms are construed non-robustly.22 If we want to cut down the range of viable options available to individual forms theorists even further beyond their status as robust particulars, this choice must be made on independent grounds which go beyond the considerations raised here. Next, as stated in (9’), Lumpl, i.e., Goliath’s matter, given the hylomorphic conception of matter advocated in Chapter 2, is itself to be construed as a matter–form compound, O*, consisting of some matter, m*, and a form, f*, assuming that Lumpl is also a structured whole: (10) O* = m* + f* Therefore, Lumpl, on this view, belongs to the very same ontological category as Goliath, i.e., both are concrete particular objects and both are analyzed as matter–form 22   As noted earlier (Section 3.2.3), individual forms theorists have a range of options available to them, depending on whether they take individual forms to be objects, properties, relations, powers, states, functions, activities, processes, facts, actions, sui generis entities, or what have you. (In each case, the ontological category in question is, in this case, construed as containing particulars as its members.) The robust construal of individual forms does not in itself favor one of these ontological categories over another; rather, it rules out certain conceptions of what the occupants of the relevant ontological category are like. The robust construal of forms requires that—regardless of what sorts of entities individual forms are—they cannot bear the very same relation both to the compound (essentially) and to the matter (accidentally). Thus, suppose, for example, that Socrates’ individual form is construed as a particularized property (e.g., a humanity trope) which essentially characterizes Socrates. Then, the robust construal requires that Socrates’ humanity trope does not also characterize Socrates’ matter, viz., his body, even accidentally, despite the fact that Socrates and Socrates’ body are spatiotemporally coincident for as long as Socrates is alive. Nevertheless, so the robust hylomorphic pluralist reasons, nothing that is human at all is human only accidentally. Therefore, even though Socrates’ body is spatiotemporally coincident with an object, viz., Socrates, which is characterized by Socrates’ humanity trope and is so essentially, Socrates’ body is not therefore also ­characterized by Socrates’ humanity trope, even accidentally.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

towards a hylomorphic solution to the grounding problem   119 compounds, assuming that Lumpl and Goliath are both structured wholes. The same reasons which motivate us to hold that Goliath has its matter, i.e., Lumpl, and its form, f, as a proper part, apply to Lumpl as well; therefore, we construe Lumpl similarly as a hylomorphic compound which has both its matter, m*, and its form, f*, as proper parts. Notice that it now follows, perhaps surprisingly, that Lumpl’s form, f*, is also a proper part of Goliath: for f* is a proper part of Lumpl; Lumpl is a proper part of Goliath; the proper parthood relation which relates Lumpl’s form, f*, to Lumpl is the very same proper parthood relation as that which relates Goliath’s matter, viz., Lumpl, to Goliath; and, finally, this single relation of proper parthood is transitive. Since Lumpl’s form, f*, is a proper part of a proper part of Goliath (viz., Lumpl), f* is therefore also a proper part of Goliath. Nevertheless, we should not infer from the fact that Lumpl’s form, f*, is a proper part of Goliath that Lumpl’s form, f*, is therefore also Goliath’s form or that it must therefore also figure in Goliath’s essence. Rather, I take it to be a basic fact about Goliath that f, and not f*, is Goliath’s form and that f, and not f*, figures in Goliath’s essence; and correspondingly for Lumpl and its relations to f and f*, respectively. Just as robust hylomorphic pluralists reject the idea that a reductive an­alysis of a matter–form compound’s relation to its form can be given or needs to be given, so proponents of a non-modal conception of essence reject the idea that a reductive analysis can be given or needs to be given of facts about the essences of objects. We can now return to the question, posed earlier in (4), concerning the particular sortalish difference between Lumpl’s and Goliath’s respective relationship to Goliath’s form, f: “In virtue of what is it the case that f figures in O’s essence, but not in m’s essence?”. Insofar as (4) issues a request for a reductive analysis of why Goliath and Lumpl have distinct essences, this question must be rejected as misguided. The best we can do is to respond to (4) by pointing out that it is a basic fact about Goliath that f (but not f*) figures in its essence, and a basic fact about Lumpl that f* (but not f) ­figures in its essence. In addition, general philosophical reasons can be adduced for preferring a thoroughly mereological conception of composition and a non-modal conception of essence over rival approaches. Beyond that, however, we must simply accept that all explanation comes to an end once we have reached facts about the essences of matter–form compounds that are taken as basic by the approach in question. While the basic essential differences between Lumpl and Goliath therefore ­cannot be further explained in terms of anything more basic, the fact that coincident objects have distinct essences, as we will see in the Section 4.5, can be used in conjunction with a non-modal conception of essence to explain why they differ with respect to their derivative modal profile.23 23   I said earlier that I will not lean on any special features concerning the specific example under discussion which arise from the fact that Lumpl and Goliath are artifacts, since I want to leave open the possibility that a similar case can, in principle, be constructed for living or non-living members of non-artifactual kinds as well. Given that Lumpl and Goliath are assumed to be artifacts, however, various options are available in this particular case which may not generalize to a case involving the members of natural kinds. For example, suppose that the expression, “statue,” singles out a genuine kind, but the expression, “piece of

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

120  hylomorphic relations

4.5  Explanatory Work for Non-Modal Essences As noted earlier (see Section 1.6: “Essence and Accident”; “The Grounding Problem”), pluralists can hope to make some headway toward formulating a response to the Grounding Problem if they combine their analysis of concrete particular objects with a non-modal conception of essence.24 According to such approaches to the essence/ accident distinction, certain facts about the essences of concrete particular objects are taken as basic and other facts concerning their modal profile (viz., facts about what is necessary but non-essential to these objects) are regarded as, at least in part, derivable from these primitive facts about their essences. As a result, a hylomorphic analysis of concrete particular objects, when combined with a suitable approach to the essence/ accident distinction, can hope to explain certain of the differences which obtain between numerically distinct spatiotemporally coincident objects, viz., differences with respect to their modal profile, in terms of differences between their respective essences, together with whatever further considerations turn out to be relevant to the derivation of the non-essential but necessary features of hylomorphic compounds from facts about their essences. As I argued in Koslicki (2012b), to make headway in explaining how the necessary accidents inhering in a hylomorphic compound follow from its essence, it is likely that we find ourselves having to appeal not just to the essence of the hylomorphic compound in question, but also to facts about the essences of distinct but related entities. To illustrate, as Aristotle brings out in his biological treatises, there is no reason to expect that the necessary (but non-essential) features of camels, say, can be derived solely from facts about the essences of camels. Rather, when we inspect Aristotle’s reasoning in these texts, it becomes evident that he takes facts about the essences of various other types of entities (e.g., animals and living organisms in general, stomachs, deserts, or earthy material) also to be relevant to a proper explanation of why camels have the particular necessary (but non-essential) features they do. To illustrate, the fact that the palates of camels are hard and durable, so Aristotle reasons, is at least in part to be clay,” does not single out a genuine kind. Then, the plurality of material parts composing Goliath compose a matter–form compound which belongs to the kind, statue, but they fail to compose an additional matter– form, viz., the alleged piece of clay, Lumpl, which is spatiotemporally coincident but numerically distinct from the statue Goliath. The strategy just outlined allows for a monist response to the particular case at hand. I consider the special features of artifacts in greater detail in Chapter 8. 24   As noted in Ch. 1 (see n. 25), I am thinking here of a conception of essence of the type encountered, for example, in the works of Aristotle (e.g., Posterior Analytics) and Kit Fine (see especially Fine (1994a, 1995a, 1995b, 1995c)). Other contemporary Aristotelians have gravitated toward a non-modal conception of essence as well; see, for example, Gorman (2005); Koslicki (2012a, 2012b, 2013a, 2013b); Lowe (1994a,  1998, 2006, 2007, 2008, 2012a, 2013); Oderberg (2007, 2011); Tahko and Lowe (2015). The conception of essence and necessity advanced in Hale (2015) is similar to Fine’s, though strictly speaking the label, “nonmodal,” does not apply to it, since Hale takes as basic certain modal truths, viz., the essential truths, from which other necessary truths are supposed to be derivable. The contrasting modal conception of essence, which takes essence to be reducible to modality, dominated the metaphysical landscape for many decades starting in the 1970s; see, for example, Forbes (1985); Lewis (1986); Mackie (2009); and Plantinga (1974), for some representative examples.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

explanatory work for non-modal essences  121 explained by reference to the fact that the palates of camels are made of earthy material. But the fact that things made of predominantly earthy material are hard and durable, for Aristotle, is itself directly traceable to facts about the essence of earth, one of the four elements which, in Aristotle’s view, materially compose the inhabitants of the sublunary sphere in different ratios. Thus, an account of why camels have the particular necessary (but non-essential) features they do (e.g., hard and durable palates) may also terminate in facts about the essences of other types of entities which are in some way implicated, in particular, in the activities, physiology, or habitat of camels. Such an approach is compatible with both the individual forms hypothesis and the universal forms hypothesis, as long as the modal properties which are supposed to be explained by reference to basic facts about essences are general modal properties, i.e., modal properties which are shared by all (or at least typical) members of a kind. If, however, some of the modal properties at issue are peculiar to the matter–forms compound at hand, then kind essences, needless to say, will not do the trick. I argued in Section 3.4.3 that some such modal properties, viz., those concerning the crossworld identity and distinctness of concrete particular objects, are in fact peculiar to specific matter–form compounds and should be explained by appeal to facts about the numerical identity and distinctness of their individual forms. Thus, proponents of the individual forms hypothesis can help themselves to the following consideration in favor of their view: assuming that specific matter–form compounds have peculiar modal properties (i.e., modal properties that are unique to these matter–form compounds and not shared by other members of the same kind), individual forms offer a better prospect for allowing us to ground these particular modal properties in the essences of matter–form compounds than do universal forms. A response to the Grounding Problem which utilizes a non-modal conception of essence will not serve to explain all the differences which obtain between numerically distinct spatiotemporally coincident objects in terms of other more basic features; rather, it assumes that all such explanations must eventually bottom out once we have reached certain basic non-modal facts about essences. To illustrate, suppose, for example, that a specification of Socrates’ essence makes reference to the fact that Socrates is a living organism that is composed of a human soul and a suitable body, while no reference to a human soul is included in a specification of the essence of Socrates’ body. Then, the question of why a specification of Socrates’ essence includes a reference to a human soul, while a specification of the essence of Socrates’ body does not include a reference to a human soul, cannot be expected to receive an answer in more basic terms, since non-modal facts about the essences of matter–form compounds are assumed to be explanatorily basic according to the approach at issue. Given this difference between Socrates’ essence and the essence of Socrates’ body, however, certain of the other derivative modal differences between a hylomorphic compound and its matter can now be explained by appeal to their basic essential differences, e.g., why upon Socrates’ death his body slowly starts to deteriorate, whereas during Socrates’ lifetime his material parts interact and hang together in a sufficiently unified manner to allow

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

122  hylomorphic relations Socrates to engage in various activities that are characteristic of human and non-human animals, e.g., perceiving, sleeping, dreaming, procreating, moving, etc. The crucial piece of apparatus which is still missing from the robust pluralist response to the Grounding Problem just outlined is a proper treatment of the relation of “following from” which, according to a non-modal conception of essence, connects basic non-modal facts about the essences of hylomorphic compounds to their derived modal profile.25 As I discuss in Koslicki (2012b), both Aristotle and Kit Fine, in their conception of the relation between essence and modality, rely on such a distinction between what belongs to the essence proper of an object and what merely follows from the essence proper of an object. On both Fine’s and Aristotle’s conception, the de re necessary truths state the necessary (but non-essential) features which merely follow from, but are not included in, the essence proper of an object. In order for this type of approach to the relation between essence and necessity to be successful, we must be able to identify an appropriate consequence relation which generates the result that the de re necessary truths about objects follow from basic truths about their essences. For example, supposing that it is part of the essence of triangles that they have three angles and merely necessary but non-essential to triangles that they have three sides, then we must be given some indication of how the second feature derives from the first. The approach Fine defends in his pioneering work on essence and modality in the 1990s certainly has an advantage over the traditional modal conception of essence, in that it is set up to reflect the sensitivity of essential truths toward their grounds, viz., facts concerning the identity of those objects in virtue of which these claims are true. But in this work Fine assumes that the relevant consequence relation which connects propositions stating what is necessary but non-essential to objects to propositions ­stating what is essential to them is that of logical entailment.26 A triangle’s being threesided, however, is not logically entailed by its being three-angled, unless additional premises are added which take the relationship in question for granted and hence make the derivation in question philosophically uninteresting (e.g., that every closed geometrical figure with three angles also has three sides and that triangles are closed geometrical figures). Aristotle’s central idea, in contrast, is that the explanatory power of definitions (statements of the essence) derives from the causal power of essences.27 This approach has the potential to open the door to a philosophically satisfying account of the   See also Hashemi (2017) for further discussion.   See also Hale’s distinction between “direct” and “indirect” essential truths (see Hale (2015), p. 154). The direct essential truths are those propositions which are immediately, or directly, true in virtue of the nature of the entities under consideration, i.e., whose truth is guaranteed by the entity’s definition. The indirect essential truths, in Hale’s view, are those which logically follow from an entity’s definition. 27   Our earlier discussion in Section 3.4.1 brought out how Aristotle divides up the causal responsibilities within a matter–form compound between its matter, which serves as the material cause of the hylomorphic compound, and its form, which serves as its formal, final, and efficient cause. See also Koslicki (2014) for further discussion of the causal priority Aristotle assigns to the form of a matter–form compound over the matter–form compound itself and its matter. 25 26

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

conclusion  123 distinction between what belongs to an entity’s essence proper and what merely follows from its essence. The relevant consequence relation which characterizes this contrast, according to Aristotle, is that supplied by his technical concept of demonstration [apodeixis], as developed in the Posterior Analytics. Demonstration encompasses more than deductive entailment, in that the explanatory order of priority represented in a successful demonstration must mirror precisely the causal order of priority present in the phenomena in question. In particular, as essences are the causal bedrock of Aristotle’s metaphysics, so definitions, the linguistic counterparts of essences, are the explanatory bedrock of Aristotle’s theory of demonstration. Armed with a suitable analysis of causation, Aristotle’s central idea, viz., to trace the explanatory power of definitions to the causal power of essences, thus points the way toward a promising direction to pursue for those who are in search of an appropriate non-logical asymmetric explanatory connection between basic non-modal facts about the essences of objects and derivative facts about their modal profile.28

4.6 Conclusion This chapter focused on the nature of the hylomorphic relations which obtain between a matter–form compound, its matter, and its form. In order to clarify the nature of these relations, I examined the question of how hylomorphists should respond to the challenge issued by the Grounding Problem (see the eighth desideratum and decision point discussed in Section 1.6). The Grounding Problem asks why an object and its matter (e.g., a statue and the clay that constitutes it) can apparently differ with respect to certain of their properties (e.g., the clay’s ability to survive being squashed, as compared to the statue’s inability to do so), even though they are otherwise so much alike. I argued for the following robust pluralist response to the Grounding Problem. According to a robust construal of forms, a matter–form compound’s relation to its form cannot be reductively analyzed in terms of conditions that are so minimal that they are met simultaneously by the compound itself (essentially) and by its matter (accidentally) merely by virtue of the fact that they are spatiotemporally coincident. In fact, robust hylomorphic pluralists reject the very idea that a reductive analysis of a compound’s relation to its form can be given or needs to be given. In combination with my independently motivated endorsement of the individual forms hypothesis (see Section 3.4.3), we thus arrive at a construal of forms as robust particulars. A form’s status as a robust particular does not single out a specific choice from among the range of ontological categories that are available to individual forms theorists (see Section 3.2.3); but the arguments I have presented here do serve to narrow down our options further to some extent, by removing from our consideration approaches which take forms to be non-robust particulars.   This causal approach to essences is, for example, developed further in Peramatzis (2018).

28

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

124  hylomorphic relations In addition, my response to the Grounding Problem follows the thoroughly mereological conception of composition I developed in Koslicki (2008a), which holds that a matter–form compound has both its matter and its form literally and strictly speaking as proper parts, according to a single relation of proper parthood. Given the close connection between form and essence (see Section 3.4.2), a compound’s relation to its form includes a further dimension, which goes beyond their mereological connection: for the compound’s form also figures in the compound’s essence either by exhausting (i.e., being identical to) the compound’s essence or by being included in the compound’s essence, along with whatever else turns out to be essential to the compound in question. Given robust mereological hylomorphic pluralism, it is a basic fact about the essences of coincident objects that they differ with respect to their formal parts. When combined with a non-modal conception of essence, this approach leads to a promising strategy for how the derivative modal differences between coincident objects may be grounded in their basic essential differences. One outstanding piece of the puzzle, which is still needed in order to complete the response I outlined in the foregoing remarks, is a proper treatment of the relevant asymmetric consequence relation which connects what is, strictly speaking, part of the essence proper of an entity and what is merely necessary (but not essential) to it. While this “following from” relation must be distinguished from logical entailment, a positive treatment of it (e.g., in terms of a suitable notion of causation) poses, as of yet, a still open research project for robust hylomorphic and non-hylomorphic pluralists who are sympathetic to a non-modal conception of essence.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

PA RT I I

Substance Introduction to Part II In Part I of this inquiry, I argued that a hylomorphic analysis of concrete particular objects is well equipped to compete with alternative approaches when evaluated against a range of desiderata. A successful application of the doctrine of hylomorphism to the specific case of concrete particular objects, among other things, hinges on how hylomorphists construe the matter or material parts composing a concrete particular object; its form; and the hylomorphic relations holding between a concrete particular object, its matter, and its form. In my view, hylomorphists should, firstly, conceive of the matter composing a concrete particular object not as prime matter or stuff, but as belonging to the same ontological type as the concrete particular object itself: according to the hylomorphic conception of matter I favor, the material parts composing a concrete particular object are to be construed as themselves matter–form compounds, as long as these material parts are also structured wholes. Secondly, hylomorphists, I argued, should reject the universal forms hypothesis (according to which forms are universal or general entities of some sort) and instead endorse a version of the individual forms hypothesis (according to which forms are particular or individual entities of some sort). Thirdly, I recommended that hylomorphists should opt for a robust compositional version of hylomorphic pluralism over its rivals, viz., nonrobust hylomorphic pluralism and hylomorphic monism. As a result, according to my preferred analysis, matter and form are to be counted literally and strictly speaking as proper parts of the concrete particular object they compose, according to a single notion of parthood; and forms, given their robust construal, do not simultaneously bear the same relation to the matter–form compound (essentially) and to the matter (accidentally). In Part II of this inquiry, I turn to the question of whether concrete particular objects, when construed as matter–form compounds, deserve a privileged position of some kind within our ontologies and, if so, according to what criterion or criteria of “ontological privilege.” This question is particularly pressing for those who find themselves drawn to the following three tenets: (i) a hylomorphic analysis of concrete

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

126  Substance ­ articular objects; (ii) an independence criterion of substancehood; and (iii) the desire p to assign substance status to at least some hylomorphic compounds. Firstly, as we know from Part I, the doctrine of hylomorphism, in its application to the specific case of ­concrete particular objects, holds that these entities are in some sense compounds of matter (hulē) and form (morphē). Secondly, proponents of an independence criterion of substancehood classify those entities as deserving of substance status which are ontologically independent of entities numerically distinct from themselves, while other entities, in turn, ontologically depend on them, according to some preferred notion(s) of ontological dependence and independence. Finally, philosophers who endorse the third tenet opt for the inclusion of certain composite entities (e.g., living organisms) in the ontologically privileged category of substances, despite their hylomorphic structure. These three commitments, when taken together, appear to lead to the following conflict. Suppose, for example, that living organisms are analyzed in the hylomorphic fashion, as compounds of matter and form. Given their metaphysical complexity, one wonders whether these entities will not then turn out to be ontologically dependent on other entities numerically distinct from them, viz., their form and possibly their matter as well. Such an outcome, however, would seem to jeopardize the inclusion of these hylomorphic compounds in the category of substances.1 The question of how best to resolve the apparent conflict which arises from these commitments should certainly be of central interest to those who find themselves attracted to the three tenets identified here. My main aim in the second part of this inquiry will be to present my own proposal as to how Aristotelians should proceed if they wish to resolve the apparent tension just noted in a way that is maximally consistent with their other central commitments. I begin Part II by considering various relations which have been defined in the ­literature under the heading of “ontological dependence” (Chapter 5). I examine first existential construals of ontological dependence and turn next to construals of ontological dependence which are formulated in terms of a non-modal conception of essence. I argue that even the most promising ones among these candidate relations are nevertheless open to objections when evaluated against various plausible measures of success. Chapter 6 examines some initially attractive attempts at formulating an independence criterion of substancehood in terms of a particular essentialist construal of ontological dependence. I argue that, in order for a criterion of substancehood to yield the desired results when applied to hylomorphic compounds, a unity criterion of substancehood for composite entities is more suitable to the task at hand than an independence criterion, despite a general preference among Aristotelians for the latter. Unity, however, has been a notoriously underdeveloped notion and my next job, in Chapter 7, therefore consists in providing a serviceable account of unity which 1   The apparent tension just identified will sound very familiar to those who have tried to wrestle with the relationship between Aristotle’s Categories and Metaphysics Z, H, Θ.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Introduction to Part II  127 captures the idea that hylomorphic compounds exhibit a higher degree of unity than other composite entities (e.g., heaps, collections, or mereological sums). Chapter 8 investigates how artifacts fare with respect to the hylomorphic analysis of concrete particular objects developed in the foregoing chapters. What hangs on the question of whether matter–form compounds can or cannot be included in the category of substance and, if so, relative to what criterion or criteria of substancehood? Although I will have much more to say on these matters in what follows, I take it that what is at issue here is a notion of substancehood that is construed as in some way indicative of ontological fundamentality. For the purposes of this introduction, we may consider Plato’s well-known Euthyphro dilemma as an illustration of what I have in mind when I speak of ontological fundamentality. In his dialogue, Euthyphro, Plato famously has Socrates object to the proposed definition of piety as what is loved by the gods on the following grounds: I’m afraid, Euthyphro, that when you were asked what piety is, you did not wish to make its nature [ousia] clear to me, but you told me an affect or quality [pathos] of it, that the pious has the quality of being loved by all the gods, but you have not yet told me what the pious is. Now, if you will, do not hide things from me but tell me again from the beginning what piety is, whether loved by the gods or having some other quality—we shall not quarrel about that—but be keen to tell me what the pious and the impious are.2

Socrates and Euthyphro agree on the existential question of whether the world contains things (e.g., acts, persons, etc.) that are loved by the gods as well as pious. Moreover, Socrates is also willing to admit that Euthyphro has managed to single out an “affect” or “quality” of the pious, namely that it is loved by the gods. In fact, Socrates would presumably even grant that being pious and being loved by the gods are necessarily correlated, i.e., that it is necessarily the case that whatever is loved by the gods is pious and vice versa. And yet, Socrates nevertheless wants to deny that Euthyphro’s proposed account has succeeded in specifying a proper definition of piety, i.e., a statement of what piety really is, one that singles out its ousia.3 Even if necessarily something is pious if and only if it is loved by the gods, it is still the case, as Plato has Socrates argue in the Euthyphro, that something’s being pious explains why it is loved by the gods, and not the other way around. It would be wrong to say, so Socrates reasons, that something is pious because it is loved by the gods: for such a conception of piety would give the gods an oddly stipulative and capricious air, as if they just decide by fiat what to love, with no basis in the thing itself that is loved. Rather, a more plausible approach to how the gods operate is that, when they love something, there is some further basis in the thing itself which explains why it is worthy of the gods’ love, namely, its being pious. Plato’s Euthyphro dilemma thus provides us with a case with the following features. Firstly, in it, we encounter two phenomena (viz., the pious and what is loved by the gods)  Plato, Euthyphro, 11b–c. Translation by G. M. A. Grube.   The term, ousia, which derives from the verb, einai (“to be”), is translated by Grube in the Euthyphro passage cited above as “nature.” But the same expression is also Aristotle’s technical term for substance. 2 3

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

128  Substance whose existence, or occurrence in the world, is not called into question by either of the disputants involved in the exchange at issue. The disagreement between Socrates and Euthyphro can therefore not be correctly characterized as an existential one, concerning such questions as “Does the pious exist?” or “Are there things which are loved by the gods?”. Secondly, the properties at issue (viz., being pious and being loved by the gods) are instantiated by the very same things (viz., acts, persons, etc.) at the very same time. And, thirdly, even though (as I am assuming) both disputants take these two phenomena to be necessarily correlated, they nevertheless disagree over what explanatory relationship holds between them. After all, in Socrates’ view, something is loved by the gods because of, or in virtue of, its being pious (and not the other way around), whereas for Euthyphro the reverse is the case: something is pious because of, or in v­ irtue of, its being loved by the gods (and not the other way around). There surely are different sorts of relations which give rise to an explanatory asymmetry and many of these are not relevant to the phenomena with which we will be concerned in Part II of this study. For example, one such relation is that of causation, at least when causation is understood as a contingent and typically diachronic relation between events (e.g., the relation between Caesar’s death and Brutus’ stabbing). According to such a conception of causation, the cause and the effect are both events; the cause usually precedes the effect in time (exempting special cases, such as those allegedly involving time travel, causal loops, and the like); and the occurrence of the earlier event causally explains the occurrence of the later event. Moreover, the relation between these events, in such cases, is considered to be contingent, in the sense that, for example, it is true to say that Brutus might not have killed Caesar. Clearly, if this is what is meant by “causation,” then the relation between the pious and what is loved by the gods would not be appropriately described as causal. A second sort of relation which gives rise to an explanatory asymmetry and which seems not to capture what is at issue in Plato’s Euthyphro dilemma is the logical relationship which obtains, for example, between the truth of the premises of a valid argument and the truth of its conclusion. In such cases, the premises entail the conclusion and, we might say, the truth of the premises logically explains the truth of the conclusion. In Plato’s Euthyphro dilemma, however, the proposition that a particular act or person is pious does not by itself entail that this act or person is also loved by the gods, or vice versa. We can therefore assume that the explanatory asymmetry in question is not due to a purely logical relation between the phenomena in question. A third category of relations which applies in probabilistic cases and also gives rise to explanatory asymmetries may safely be bracketed for present purposes as well. For example, the decay of a particle can be probabilistically explained by appeal to its half life or a patient’s recovery from strep throat can be probabilistically explained by that patient’s having taken antibiotics. But evidently the relationship between something’s being pious and something’s being loved by the gods is also not of this sort. Given its long and illustrious philosophical pedigree, the idea that there can be explanatory asymmetries even between necessarily correlated phenomena of the sort

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Introduction to Part II  129 Plato singles out in the Euthyphro thus is, or at least should be, quite familiar to us. It is fair to say, however, that no consensus has as of yet emerged among philosophers as to how best to capture and illuminate such connections. Whatever exactly is the nature of the connection between the pious and what is loved by the gods, it seems quite clear that it is not a contingent and diachronic relation between events; it is not a purely logical relation; and it is also not a probabilistic relation. These characterizations are of course all negative. It is difficult to add much by way of a positive characterization of our subject matter to what has already been said without immediately becoming embroiled in controversy. For now, I will therefore limit myself only to the following preliminary specification: when I speak below of substantive non-existential disagreements in ontology over questions of fundamentality, I have in mind phenomena which are connected by means of a relation that is synchronic, necessary, non-logical, non-probabilistic, and gives rise to an explanatory asymmetry such as that noted by Plato in the Euthyphro. In addition, relations which satisfy these characteristics are also sometimes described as “non-causal.” As I remarked earlier, I am willing to go along with this characterization, as long as we are absolutely explicit that what is meant by “causation” here is a contingent and typically diachronic relation between events, such as that exhibited in the case of Brutus’ killing of Caesar. However, there are also good reasons to avoid the characterization, “non-causal,” just to minimize possible sources of confusion. Plato, for one, would want to describe the relationship between the pious and what is loved by the gods as causal, just not in the sense of “causal” identified earlier. For the position he endorses in the Phaedo and other middle period dialogues is precisely one according to which forms are causes. In these texts, Plato argues that the cause of a particular statue’s being beautiful, i.e., that which makes the statue beautiful and or that in virtue of which the statue is beautiful, is its participation in the form, beauty. I am sympathetic to Plato’s assessment that the asymmetric explanatory relations in question is causal (see Section 4.5), though I would take the causal principles in question to be Aristotelian, and not Platonic, forms. Given this perspective, then, it seems wise not to adopt the label, “non-causal,” to delineate the asymmetric explanatory relations that are relevant to my current project. As I have argued in Koslicki (2016a), some of the most interesting and important debates which properly belong to the study of being, whether we call it “metaphysics” or “ontology,” do not concern existential questions at all; rather, such disputes, as I have claimed, focus in some cases on non-existential disagreements over questions of fundamentality. To illustrate, consider a specific example of such a dispute in ontology, namely, that between two different kinds of trope theorists: E. J. Lowe, who takes tropes to be ontologically dependent on their bearers (see Lowe (2006)), and Keith Campbell, who accepts a reductive analysis of the bearers of tropes as trope bundles (see Campbell (1990)). If my diagnosis of this dispute is correct, then these two trope theorists precisely find themselves in a substantive non-existential disagreement in ontology over questions of fundamentality. The dispute at issue should not strike us as

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

130  Substance one in which the engaged parties are simply talking past each other; rather, by reading the dispute in question as a substantive non-existential disagreement over a question of fundamentality, we can see that the reductive trope theorist precisely affirms what the non-reductive trope theorist denies, and vice versa. The idea that such non-existential disagreements over questions of fundamentality can be substantive also plays a significant role in the defense of a realist outlook in metaphysics. For metaphysical realists have an important stake in maintaining, against their anti-realist opponents, that some disputes in ontology are non-verbal, i.e., cannot be resolved merely by pointing to a linguistic confusion or equivocation. If some of these substantive ontological disputes furthermore turn out to be non-existential, then the anti-realist’s job will have been made even more difficult, since in that case no single piece of apparatus that is designed specifically to deal with existential disputes (e.g., a certain treatment of the existential quantifier) can be effective in showing that all apparently substantive disputes in ontology are really based on verbal disagreements.4 Thus, the proper classification of such substantive non-existential disagreements over questions of fundamentality has far-reaching consequences for how we should conceive of the nature and business of the study of being as a discipline. Since it is not at all obvious how Quinean or Carnapian conceptions of our discipline could accommodate the possibility and intelligibility of substantive non-existential disagreements in ontology over questions of fundamentality, the recognition of such disputes therefore bolsters the overall case for an alternative conception of the study of being, different from those put forth by Quine and Carnap or their more recent followers.5 The Aristotelian outlook defended here aims to develop such an alternative proposal, in continuation with the project I began in Koslicki (2008a). When I spoke of “fundamentality” earlier, I had in mind relative fundamentality, rather than absolute fundamentality. Surely, in the case of Plato’s Euthyphro dilemma, most philosophers, including Plato, would not be particularly tempted to say that some­ thing’s being pious is an absolutely fundamental matter, even if they agree that something’s being pious is more fundamental than its being loved by the gods.6 In what follows, although I remain neutral in my commitments concerning the absolutely   See, for example, Hirsch (2002).   See, for example, Quine (1948); Carnap (1950). For a more recent expression of the Quinean approach to ontology, see also “Thesis 2” of van Inwagen (2009): “Being is the same as existence” (p. 480). The neoCarnapian approach to ontology and metaphysics is particularly well represented in Chalmers, Manley, and Wasserman (2009); see, for example, the contributions by Chalmers, Hirsch, Hofweber, and Thomasson; also in this vein is Chalmers (2012). Quine’s conception of ontology as concerning questions of existence is so mainstream that it is usually just taken for granted as a presupposition which does not stand in need of justification. Quine’s take on metaphysics at large, however, is considerably more controversial and assumes additional machinery such as his rejection of the analytic/synthetic distinction, holism, pragmatism, ontological relativity, and so forth, on which there is much less consensus among contemporary philosophers. 6   I take it that, for the Plato of the middle period, only the form of the good would count as an absolutely fundamental entity. All the other entities in Plato’s middle period ontology (e.g., the other forms, the sensible particulars, or the mathematical objects) would receive only a comparative ranking depending on their relation to the form of the good. 4 5

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Introduction to Part II  131 fundamental, I do recommend that we maintain a cautious attitude towards attempts to define absolute fundamentality in terms of relative fundamentality. This line of reasoning would proceed by taking an entity, x, to be absolutely fundamental just in case there is no entity, y, numerically distinct from x which is more fundamental than x in some specific way, or in any way at all. In what follows, we will consider an attempt to define a type of absolute fundamentality, viz., substancehood simpliciter, in terms of a relation which is often understood as indicative of relative fundamentality, viz., ontological dependence. As we will discover, this attempted definition, even if it might seem initially appealing, has unattractive consequences which need to be addressed by those who are tempted to endorse it.7 The notion of grounding has been hailed by philosophers in recent years as the right tool by means of which to approach relative fundamentality, with the potential to lead to a definition of absolute fundamentality in terms of ungroundedness.8 Grounding theorists generally assume that this notion is not reducible to, or analyzable in terms of, other familiar idioms such as identity, necessity, parthood, supervenience, realization, causation, or counterfactual dependence.9 At least with respect to its formal properties, grounding does appear to hold more promise than earlier modal–existential notions, such as supervenience, for the purposes of developing an approach to relative fundamentality, if only because grounding is commonly stipulated to be asymmetric and not definable in modal terms. However, as I have argued elsewhere (Koslicki (2015a, 2016b)), grounding suffers from some of the same deficiencies which were noted in connection with the concept of supervenience: most prominently, grounding, like supervenience, also fails to be sufficiently fine-grained to do its intended explanatory work. In addition, there is doubt as to whether the phenomena collected together under the rubric of grounding are really unified by the presence of a single relation. And, finally, grounding turns out not to be particularly helpful in capturing and illuminating what is philosophically important about the traditional substance/non-substance distinction. In the end, although grounding performs better than supervenience in some respects, it does not solve all of the problems to which supervenience-based approaches were found to fall prey. If the absolutely fundamental turns out not to be straightforwardly definable in terms of available notions of relative fundamentality, one might be led to consider one of the following possible attitudes toward the absolutely fundamental. First, one might try to do away with the category of the absolutely fundamental altogether. A second strategy would be to aim at treating absolute fundamentality as a defined notion, 7   See also Zylstra (2014) for discussion concerning the relationship between ontological dependence and fundamentality. 8   See, for example, Audi (2012a, 2012b); Bennett (2011a, 2011b, 2017); Correia and Schnieder (2012); Fine (2001, 2012); Hoeltje, Schnieder, and Steinberg (2013); Jenkins (2011); Raven (2012); Rosen (2010); Schaffer (2009, 2010); Trogdon (2013). 9   See, for example, Kim (1984, 1990); McLaughlin (1995); and McLaughlin and Bennett (2005), for useful resources concerning the state of the literature on supervenience.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

132  Substance but  one whose definition can be stated in terms of something other than relative ­fundamentality. Yet a third strategy would be to accept absolute fundamentality as a primitive undefined notion. Since my primary interest in what follows lies with relative fundamentality, I will not, within the current context, pursue the question of which approach to the absolutely fundamental yields the best results. Overall, the conclusions reached in Parts I and II together allow us to classify hylomorphic compounds in a comparative way, as more deserving of substance status than other composite entities, due at least in part to their high degree of unity. When ­substancehood is understood in this way, a comparative notion of substancehood ­corresponds to a similarly comparative ranking of entities by their degree of relative fundamentality: such a notion of fundamentality allows us to assess an entity or type of entity as more or less fundamental (or derivative) than some other entity or type of entity. A comparative classification of entities, however, should not be confused with an absolute classification of entities or types of entities as substances simpliciter. When we encounter cases in which one type of phenomenon is alleged to be more or less fundamental (or derivative) than another, we should not immediately assume that we are dealing with a case in which the purportedly more fundamental phenomenon in question can also be correctly described as absolutely fundamental. Thus, I will take care in what follows to separate questions concerning relative (or comparative) fundamentality from those concerning absolute fundamentality. A comparative classification of entities as more or less deserving of substance status suggests that a correct characterization of the world we live in requires the recognition of complex and widespread networks of relative fundamentality. As noted just now, such a relative fundamentality ranking is not be confused with a classification of hylomorphic compounds as belonging to the category of the absolutely fundamental or the substances simpliciter (if there are any such entities). At the same time, however, the importance of assigning hylomorphic compounds to their proper position within the hierarchy of the relatively fundamental should not be underestimated. As I will go on to illustrate, many of the most interesting metaphysical projects, especially those dear to Aristotelians, take place within the sphere of what, in absolute terms, would count as ontologically derivative. In particular, in subsequent chapters, we will encounter reasons for thinking that multiple explanatory factors are relevant to the classification of an entity as more or less fundamental (or derivative) than another. To illustrate, tropes, holes, boundaries, or Aristotelian universals are sometimes taken to be abstracted from, and hence derivative in one particular way of, the concrete particular objects that are their bearers. And yet, despite their abstracted status, these entities may nevertheless be regarded as highly unified, e.g., perhaps because they are taken to be simple in the sense of not being constructed out of constituents, and hence they would count as unified by default. Moreover, their abstracted status also leaves open whether the entities in question are artificial (e.g., a boundary around a particular piece of land) or natural (e.g., an Aristotelian universal or a naturally formed hole in a rock formation). Heaps, mereological sums, collections, artifacts, artworks, and

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Introduction to Part II  133 natural unified wholes, in contrast, may be taken to be in some sense constructed out of their parts or constituents; but whether they are unified or disunified, artificial or natural, is not immediately settled by their status as constructed entities. And while artifacts, artworks, and intentional objects are arguably the results of certain kinds of creative acts involving intentional agents, some of these entities may be classified as abstracted (e.g., an after-image); others as constructed (e.g., a sculpture). In addition, some of these entities may have a relatively high degree of unity (e.g., certain artifacts whose constituents work together to fulfil a certain function), whereas others may exhibit a relatively low degree of unity (e.g., an artwork consisting of components that are scattered across multiple geographical locations). Finally, heaps, mereological sums, and collections, which may be taken to be constructed entities, also appear to exhibit a relatively low degree of unity compared to other more unified entities (e.g., natural unified wholes, tropes, simples, Platonic universals, etc.). Given this multidimensional approach to relative fundamentality, one and the same entity can be classified, without giving rise to contradictions, as more fundamental than another in certain respects (e.g., its degree of unity) but not in others (e.g., its artificiality). According to this approach, when we ask in a particular case whether some entity, or type of entity, is more or less fundamental than another, we must always specify the particular respect in which something is to be categorized as more or less fundamental than something else to which it is being compared. Currently available approaches to relative fundamentality do not reflect the various dimensions of relative fundamentality and non-fundamentality I have distinguished earlier: (i) whether an entity is essentially abstracted from something more complex; (ii) whether an entity is essentially constructed out of other entities; (iii) whether an entity is essentially artificial; or (iv) whether an entity is highly unified. Just as one and the same thing can be both good in some respects (e.g., dancing) and bad in others (e.g., playing basketball), one and the same entity, according to this multidimensional approach, can be both more fundamental than another in certain respects (e.g., its degree of unity) and less fundamental than another in others (e.g., its naturalness). In its broadest statement, then, the second part of this book aims to bring out that the recognition of multiple dimensions of relative fundamentality and derivativeness is required in order to do justice to the data that presents itself to those engaged in the study of being.10

10   This multidimensional approach to questions of relative fundamentality is discussed in more detail elsewhere; see especially Koslicki (2012a, 2013a, 2013b, 2015a, 2015b, 2016a, 2016b). In Chapter 7, I go on to develop a detailed account of unity as involving a particular kind of mutual interactional dependence among the parts of a whole and the whole to which they belong. The second dimension of fundamentality, viz., construction, has already been extensively studied by metaphysicians, including myself, as can be gleaned from the voluminous literature on various specific construction operations, such as mereological composition and constitution. The third dimension of non-fundamentality, viz., artificiality, will be addressed in Chapter 8. A detailed examination of the first dimension of non-fundamentality, viz., abstraction, will have to await another occasion.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

5 Ontological Dependence 5.1 Preliminaries Ontological dependence is typically taken to be a relation whose relata are entities.1 The following cases are often cited in the literature as putative candidates of pairs of entities which exhibit a relation of ontological dependence of some sort: (a) Smiles ontologically depend on mouths that are smiling. (b) (Non-empty) sets ontologically depend on their members. (c) Events or states of affairs (e.g., lightning or heat) ontologically depend on their participants (e.g., electrons, molecules, and the like). (d) Chemical substances (e.g., water) ontologically depend on their molecular/ atomic constituents (e.g., H2O molecules, hydrogen and oxygen atoms, and the like). (e) Tropes (e.g., the redness of a particular tomato) ontologically depend on their bearers (e.g., the tomato). (f) Aristotelian universals (e.g., redness) ontologically depend on their bearers (e.g., objects that are red). (g) Holes (e.g., the holes in a piece of Emmentaler cheese) ontologically depend on their “hosts” (e.g., the piece of Emmentaler cheese). (h) Boundaries (e.g., the boundary around a football field) ontologically depend on their “hosts” (e.g., the football field). In all of these cases, if in fact they do constitute examples of pairs of entities related by an ontological dependence relation of some sort, the dependence relation in question may plausibly be taken to be asymmetric. Any number of relations have been defined in the literature under the heading of “ontological dependence.” As we will see in what follows, some of the most popular definitions are formulated in modal terms; others in non-modal (e.g., explanatory or essentialist) terms; some (viz., the existential construals of ontological dependence) emphasize requirements which must be met in order for an entity to exist; others (viz., the non-existential essentialist construals of ontological dependence) focus on requirements which must be met in order for an entity to be the very entity that it is at   Sections 5.1–5.3 are drawn from Koslicki (2012a) and Koslicki (2013a), with minor revisions.

1

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

136  Ontological Dependence each time at which it exists; some are rigid, in the sense that they involve a relation between particular entities; others are generic, in the sense that they involve only a relation between an entity and some entities or other which bear certain characteristics. In Section 5.2, I discuss existential construals of ontological dependence; non-existential essentialist construals will be examined in Section 5.3 and in Chapter 6.2 With this plethora of defined relations, it is particularly important to set out clear and principled standards by which to evaluate the explanatory usefulness of one such technical notion as compared to another. One possible way to measure success in this area is by considering how well a particular notion does in classifying putative cases of ontological dependence, such as those mentioned earlier. This explanatory goal of classifying particular cases correctly, as is to be expected, does not yield an uncontroversial measure of success. For different philosophers, depending on their particular views concerning specific cases, will disagree over which putative cases constitute good examples of pairs of entities exhibiting a relation of ontological dependence of some sort and in which direction the dependence relation in question runs. For example, some trope theorists (namely, those who view concrete particular objects as bundles of tropes) will disagree with other trope theorists (namely, those who do not view concrete particular objects as bundles of tropes) over the status of (e). Trope theorists of the first kind may take the bearers of tropes to depend ontologically on the tropes that compose the bundle in a manner analogous to (b), e.g., the way in which non-empty sets ontologically depend on their members, while trope theorists of the second kind may agree with the placement of tropes in (e) among the ontologically dependent items. Moreover, trope theorists of the second kind may or may not find it necessary to be committed to Aristotelian universals in addition to their commitment to tropes, which will affect the question of whether the entry (f) presents us with another genuine example of an ontologically dependent type of entity. Thus, whatever putative classifications we take as our starting point, they cannot be viewed as representing judgments that are universally agreed upon from the outset. Nevertheless, specific proposals can often be evaluated in an ad hominem way, ­relative to the commitments of a particular philosopher who is putting forward the definition under consideration; or, in at least some cases, they can be evaluated in a way that will, or at least ought to be, found compelling by reasonable interlocutors in general. Secondly, as noted earlier, it is quite common for those who define relations of ontological dependence to employ these notions as markers of ontological fundamentality, especially in the context of formulating an independence criterion of substancehood. Such a criterion is intended to classify those entities as substances on which other types of entities ontologically depend and which do not themselves ontologically depend on other entities numerically distinct from themselves, according to some preferred 2   For other useful discussion and surveys of the literature on ontological dependence, see also Correia (2008), Mulligan (1998), Smith (1982), Smith and Mulligan (1982), and Tahko and Lowe (2015).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

preliminaries  137 notion of ontological dependence. Those who find this connection between ontological independence and substancehood congenial can thereby avail themselves of a second possible measure of success by which to evaluate a particular proposed definition of ontological dependence, viz., to what extent the relation in question contributes to the formulation of a plausible criterion of substancehood. Unsurprisingly, this second possible measure of success also does not yield an uncontroversial measure by which to evaluate the explanatory usefulness of a particular definition of ontological dependence, since different philosophers will take different positions on the question of which entities (if any) should be classified as more deserving of substance status than entities belonging to some other relevant comparison class, or as substances simpliciter. Moreover, the question of whether and how ontological fundamentality or substancehood are connected with ontological dependence (or the lack thereof) is itself a matter of dispute.3 A third potential measure of success is how well the account in question accommodates the possibility that reasonable philosophers who agree on questions of existence might nevertheless carry on a substantive disagreement in ontology over questions of fundamentality. Thus, two philosophers might, for example, agree that both particulars and universals exist, but disagree over whether one category of entities should be classified as absolutely fundamental (e.g., as substances simpliciter) or as more fundamental (e.g., as more deserving of substance status) than the other. One of these philosophers (e.g., Aristotle in the Categories) might hold the position that particulars are more deserving of substance status than universals, perhaps precisely because, in this philosopher’s view, universals in some way ontologically depend on particulars. In contrast, the other philosopher (e.g., Plato during the middle period) might hold the position that universals should be classified as more deserving of substance status than particulars, perhaps precisely because, according to this second philosopher, particulars in some way ontologically depend on universals. But if a restriction to particulars were simply built into the definition of fundamentality or substancehood, then the thesis that particulars are more fundamental or more deserving of substance status than universals would simply follow from the definition in question and the opposing thesis, that universals are more fundamental or more deserving of substance status than particulars, would be blocked by the definition in question. Ideally, however, an account of ontological dependence should be sufficiently neutral on this question, so that it can make room for the possibility that disagreements in ontology over questions of fundamentality, such as the dispute just outlined, might be substantive, i.e., neither trivially answerable nor resulting in an inconsistency or incoherence of some sort. Since it is a contested matter whether (and, if so, which) particular disputes in ontology are of the kind just described, this third potential measure of success once 3   In addition to the remarks concerning ontological dependence and substancehood in this chapter, I will also take up this topic again in Chapter 6, where a number of proposed independence criteria of substancehood will be examined in more detail.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

138  Ontological Dependence again cannot be expected to yield uncontroversial and universally agreed upon results when it is used to evaluate specific accounts of ontological dependence.4 Fourthly, suppose a particular definition of ontological dependence performs quite well with respect to the three goals just stated: it classifies some collection of phenomena in some particular desired way; it helpfully contributes to the formulation of a criterion of substancehood; and it makes intelligible how non-existential ontological disputes over questions of fundamentality can be substantive. It would be an added benefit if, in addition to satisfying these three explanatory goals, a definition of ontological dependence could also provide some degree of illumination concerning the nature of the connections at issue. To illustrate, an entity can be classified as more ontologically derivative than some other relevant comparison class for a variety of different reasons: e.g., because the entity in question is abstracted from something more complex (e.g., as a trope might be thought to be abstracted from its bearer); or because it is constructed from other entities (e.g., as a non-empty set might be thought to be constructed from its members); or because it is artificial (e.g., an artifact or artwork) as opposed to natural; or because it exhibits a low degree of unity compared to other more highly unified entities (e.g., as heaps might be thought to exhibit a lower degree of unity than other types of composite entities, such as structured wholes). The fourth criterion considers it to be a mark of explanatory success if a candidate definition of ontological dependence gives us some insight into the particular explanatory factors that are at play in classifying an entity as ontologically dependent on, or independent of, some other entity. As in the previous three cases, however, it is again a matter of dispute whether the fourth criterion yields a plausible measure of explanatory success, since some may deny that candidate definitions of ontological dependence should be expected to illuminate the nature of the connections at issue. One might be tempted to add to these four criteria by which to measure the explanatory success of a particular definition of ontological dependence a fifth one, viz., how well the definition in question manages to capture our so-called “intuitions” concerning the phenomena in question. It is a contested question among philosophers what exactly might be meant by “intuition,” when some role is assigned to these epistemic states as yielding a potential data set with respect to which a given theory may allegedly be evaluated. Often, so-called “intuitions” are divided into those that are “pre-theoretic” or “pre-philosophical” and those that already carry with them some degree of theoretical commitment. Consider, for example, case (g), e.g., the relation between the holes in a particular piece of Emmentaler cheese and the piece of Emmentaler cheese in which they reside. If by “intuition” we have in mind the first kind (i.e., so-called “pre-theoretic” or “pre-philosophical” intuitions), then I am not sure that it is possible to have an “intuition” of this kind to the effect that holes are, in fact, related in this way to their “hosts.” For how could one be in a position to judge that the case at issue follows the 4   For a more detailed discussion of non-existential disputes in ontology over questions of fundamentality, see also Koslicki (2016a). I examine ground-theoretic approaches to fundamentality in Koslicki (2015a, 2016b).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

varieties of existential dependence  139 pattern identified earlier, unless one is already in possession of some conception of what sorts of entities holes are and what sort of asymmetric explanatory relation holds between holes and their “hosts.” But such a conception, if we have one at all, cannot very well be characterized as “pre-philosophical” or “pre-theoretic.” If, on the other hand, by “intuition,” we have in mind an epistemic state that is already informed by philosophical theorizing, then it is quite likely that the “intuition” in question is somehow wrapped up with the four explanatory goals I have identified earlier as potential measures of success against which a particular definition of ontological dependence may be evaluated. Or, if in addition to these four explanatory goals I have identified, there are additional ones that I have missed, then these should be made explicit as well. Once this has been done, their plausibility can be subjected to scrutiny; but, at that point, it seems that we have left appeals to “intuition” behind and have entered into the business of explicit theorizing. For these reasons, I see no need to add “Agreement with Intuitions” as a fifth criterion over and above the four already cited.

5.2  Varieties of Existential Dependence 5.2.1  Ontological Dependence in Aristotle’s Categories Ontological dependence is often construed existentially.5 For example, Aristotle is standardly read as putting forward an existential claim when he famously states in the Categories: Thus all the other things are either said of the primary substances as subjects or in them as ­subjects. So if the primary substances did not exist it would be impossible for any of the other things to exist. (Categories, Ch. 5, 2b3–6; my emphasis)6

But where the English translation has “exist,” the Greek simply has the verb “einai” (“to be”), which can sometimes be rendered in an existential sense, but need not be so rendered. Thus, using the more neutral terminology of “being” rather than “existence,” we may read Aristotle as putting forward the following counterfactual dependence claim: “If the primary substances were not [or: did not have being], it would be impossible for any of the other things to be [or: to have being].” Notice that Aristotle in the passage quoted just now speaks of two different ways in which the primary substances can be marked off from everything else: (i) all the other entities are either said of the primary substances as subjects; or (ii) they are (present) in the primary substances as subjects. Aristotle understands the phrase, “being in a subject,” in the following way: what is in something (a) not as a part, and (b) cannot be separately from what it is in (Categories, Ch. 2, 1a24–5). (Again, the occurrence of “to be” [“einai”] here is standardly translated as “exist” as well: “ . . . and cannot exist separately from what it is in”; but the point raised above applies here just the same.)   For more discussion on ontological dependence in Aristotle, see also Corkum (2013).   Translation by J. L. Ackrill (see Barnes (1984)).

5

6

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

140  Ontological Dependence I interpret Aristotle’s two relations, being said of a subject and being in a subject, as indicating two ways in which all the other entities depend ontologically on the primary substances, while the primary substances do not in turn ontologically depend on anything else in either of these two ways. These two varieties of ontological dependence, for Aristotle, correspond to two different forms of predication: essential predication, as when we say of something (e.g., Socrates) that it is a member of a certain taxonomic category, i.e., a species (e.g., human being) or a genus (e.g., animal); and accidental predication, as when we say of something (e.g., Socrates) that it bears a certain accidental feature (e.g., paleness).7 Neither the variety of ontological dependence which corresponds to the relation, being in a subject (as indicated by accidental predication), nor the variety of ontological dependence which corresponds to the relation, being said of a subject (as indicated by essential predication), should be read in an exclusively existential way.8 If the particular instance of paleness that inheres in Socrates at a particular time is construed as a non-repeatable and non-transferable entity (e.g., a trope, mode, or moment), then it is certainly true that it would be impossible for this individual instance of paleness to exist unless Socrates exists as well. And, given an Aristotelian conception of universals, it would similarly be impossible for universals in any category to exist unless concrete particular objects exist as well. For in order for the universal, color, to exist for example, individual instances of color must exist as well; and in order for individual instances of color to exist, concrete particular objects must exist in which these individual color instances can inhere. At the same time, as has often been observed, if we construe ontological dependence in Aristotle’s Categories in a purely existential fashion, then the entities he identifies there as most deserving of substance status compared to everything else (e.g., concrete particular objects, such as human beings or horses) would lack the distinctive asymmetric ontological independence from other entities which Aristotle seems to 7   Although Aristotle does not overtly speak of dependence in the passage just cited, I nevertheless interpret his two relations, being said of a subject and being in a subject, as indicating two ways in which all the other entities depend ontologically on the primary substances, while the primary substances do not in turn ontologically depend on anything else in either of these two ways. The following passage from Met. Γ.2 does contain a Greek term which is rendered by W. D. Ross into English as “depend” (“ἤρτηται” from “ἀρτάω,” “to fasten to,” “to hang one thing upon another;” hence, in the passive voice, “to be fastened to,” “to be hung upon another thing,” i.e., “to depend”):

But everywhere science deals chiefly with that which is primary, and on which the other things depend [ērtētai], and in virtue of which they get their names. If, then, this is substance, it will be of substances that the philosophers must grasp the principles and the causes. (Met. Γ.2, 1003b16–19; trans. W. D. Ross) According to Met. Z.1, substance is primary in three senses, viz., in formula [logō(i)], in order of knowledge [gnōsei], and in time [chronō(i)]. (See also Met. Θ.8, where the actual is said to be prior to the potential in three senses of “priority,” viz., in definition [logō(i)]; in time [chronō(i)]; and in substance [ousia(i)].) For further discussion, see Koslicki (2014, 2015b). 8   Aristotle’s dependence claim in the Categories, when read in an exclusively existential way, gives rise to the following definition: “x existentially depends on y ≡def If y did not exist, then it would be impossible for x to exist.” A non-existential reading of ontological dependence in Aristotle is also defended in Corkum (2008, 2013) and Peramatzis (2008, 2011).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

varieties of existential dependence  141 want to attribute to them. For in order for a concrete particular object, such as Socrates, to exist, some other individuals and universals must exist as well which can be either said of Socrates or present in him as a subject. Even though Socrates can exist without the particular color instances that are predicated of him accidentally at any particular time, some color instances or other must be present in him at any time at which he exists. And the existence of some color instances or other in turn necessitates the existence of some more general qualitative universals as well, such as paleness and color, to which these individual color instances essentially belong. Finally, if the more general taxonomic categories of which Socrates is essentially a member (e.g., the species, human being, or the genus, animal) did not exist, it would be impossible for Socrates to exist as well. Therefore, if the notion of ontological dependence at work in Aristotle’s Categories were interpreted in a purely existential manner, then concrete particular objects, such as Socrates, would come out as roughly on a par with respect to their degree of ontological independence with other entities, e.g., qualitative universals such as color. But in the Categories Aristotle clearly thinks that concrete particular objects, such as Socrates, are ontologically independent of other entities in a way in which other types of entities are not and that entities of other types are ontologically dependent on concrete particular objects in a way in which they are not in turn also ontologically dependent on other types of entities. We should thus conclude that an exclusively existential construal of Aristotle’s independence thesis in the Categories does not provide the most charitable reading of what Aristotle has in mind there when he claims that all the other entities in some way asymmetrically depend on those entities which are most deserving of substance status, viz., the primary substances.9, 10

9   Corkum (2008, 2013), for example, interprets Aristotle’s conception of ontological dependence in the following way: other entities are ontologically dependent on the primary substances because they inherit their ontological status from the primary substances to which they are either accidentally or essentially related, while the primary substances do not in turn inherit their ontological status from other entities. How plausible this proposal is depends on how we understand the key notion of inheriting one’s ontological status from something; as it stands, this notion is not elaborated in Corkum (2008) to a sufficient degree to get a good handle on it. For further discussion, see Corkum (2013). Michail Peramatzis interprets Aristotle’s conception of ontological priority (the flip side of ontological dependence) as the ontological correlate of definitional priority (see Peramatzis (2008, 2011)). The notion on which Peramatzis focuses is thus perhaps more tailored to the views Aristotle espouses in his later work (e.g., the Metaphysics), where form seems to take on the status of primary substance, than to those we find in the Categories, where certain kinds of concrete particular objects (e.g., organisms) occupy the role of primary substances: for particulars, in Aristotle’s view, cannot be defined, but only perceived. 10   In contrast, Husserl, in the Logical Investigations, really does seem to have in mind existential dependence when he speaks for example of a color moment that is part of a particular more inclusive whole as being founded upon an extension moment that is part of the same whole, and vice versa (see Husserl (1900–1)). A moment, he says, is a non-independent object in the sense that it requires something in addition to itself to exist (a more inclusive whole of which the moment is a part) in order for it to exist. And a moment is founded upon another moment, if the first moment cannot exist unless it is part of a more comprehensive unity which connects it with the second moment. Both Husserl’s conception of non-independent object and his conception of foundation thus seem to appeal to a notion of existential dependence. For discussion, see, for example, Mulligan, Simons, and Smith (1984), Correia (2004) and the references found therein.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

142  Ontological Dependence

5.2.2  Modal Existential Dependence 5.2.2.1  Rigid Existential Necessary Dependence A straightforward modal/existential notion of ontological dependence is, for example, that defined by E. J. Lowe under the heading, “Rigid Existential Necessary Dependence”:11 (ND1)  Rigid Existential Necessary Dependence: x is rigidly existentially necessarily dependent on y ≡def Necessarily, x exists only if y exists. Earlier (see Section 5.1), I identified the following explanatory goals against which a particular definition of ontological dependence may be evaluated. Firstly, does it achieve a particular desired classification of paradigmatic cases of ontological dependence? Secondly, is it useful from the point of view of formulating a criterion of substancehood? Thirdly, does it allow for substantive non-existential disagreements in ontology over questions of fundamentality? Fourthly, is it sufficiently fine-grained to capture and illuminate the nature of the connections at issue? When evaluated against these explanatory goals, a straightforwardly modal/existential construal of ontological dependence along the lines of (ND1) turns out not to be satisfactory. Such a construal of ontological dependence is susceptible to objections coming from all four criteria; for reasons of brevity, I confine myself here to some well-known difficulties connected with just the first two explanatory goals. Suppose that Aristotle’s claim in the Categories is correct and everything in some way ontologically depends on concrete particular objects, while they themselves do not in the same way ontologically depend on the other entities. As pointed out, for example, in Lowe (1994a), among the apparent trouble cases which arise for (ND1) are the following. Suppose that a certain concrete particular object can exist only if certain of its constituents exist; then, by (ND1), such an entity will turn out to depend ontologically on these constituents.12, 13 The same will be true if, in some cases, 11   See also Simons’ notion of “Weak Dependence” (Simons (1998), p. 236). For Lowe’s most up-to-date views concerning ontological dependence, see Lowe (2006, 2008, 2012a, 2013); as well as Tahko and Lowe (2015). For discussions of ontological dependence in Lowe’s earlier work, see Lowe (1994a, 1994b, 1998). Also relevant are his views concerning criteria of identity; see, for example, Lowe (1989, 1997, 2009). 12   Simons’ notion of “Strong Dependence” or “Strong Rigid Dependence” simply rules out this particular group of apparent counterexamples by adding a clause which requires y not to be a proper part of x (see Simons (1987), p. 303; Simons (1998), p. 236). This exclusion of proper parts from a definition of ontological dependence raises tricky methodological issues, especially when the notion of ontological dependence in question is used in the formulation of a criterion of substancehood. For discussion, see, e.g., Toner (2010) and Koslicki (2013b); these issues will also concern us again in more detail in Chapter 6. 13   It is difficult to find uncontroversial examples which would illustrate why it might be plausible to think that a certain concrete particular object can exist only if certain of its constituents exist. I, for one, would argue that organisms, for example, provide us with a case in point. For, according to the mereological version of hylomorphism I favor, a particular organism, such as Socrates, is analyzed as a compound of matter and form; moreover, Socrates’ form and matter, on this view, are regarded as being literally and strictly speaking proper parts of Socrates. Given the strong correlation between form and essence, such a

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

varieties of existential dependence  143 a concrete particular object can exist only if it originated from certain antecedently existing entities. For example, a human being might be thought to be related in this way to the particular sperm and egg or to the particular zygote from which he or she developed. Moreover, suppose there are objects which exist necessarily (e.g., the number 8); then again, by (ND1), everything will ontologically depend on them. Finally, suppose there are particularized properties which are necessarily had by their bearers (e.g., Socrates’ humanity) or four-dimensional entities which are necessarily coexistent with the concrete particular objects with which they are affiliated (e.g., Socrates’ life); then again, by (ND1), concrete particular objects will turn out to be ontologically dependent on these entities. Cases such as these indicate that modal/ existential dependence is too coarse-grained to yield an explanatorily useful notion of ontological dependence. 5.2.2.2  Generic Existential Necessary Dependence Not all cases of existential dependence are cases in which an entity can be said to be rigidly existentially dependent on another entity. In some cases, an entity may only require for its existence that some entities or other, which bear certain characteristics, exist as well. To capture these sorts of cases, Lowe defines the notion of “Generic Existential Necessary Dependence”:14 (ND2)  Generic Existential Necessary Dependence: x is generically existentially necessarily dependent on Fs ≡def Necessarily, x exists only if some Fs exist. To illustrate, those who subscribe to an Aristotelian (as opposed to a Platonist) conception of universals will presumably take (ND2), but not (ND1), to be appropriate for a characterization of the relation between universals and the particulars that instantiate them. For it is part of the Aristotelian conception that universals can only exist if particulars instantiating them exist as well. But the existential dependence in question would have to be generic and not rigid, since the Aristotelian conception certainly does not commit one to thinking that a universal, e.g., redness, can exist only if some specific concrete particular red object, e.g., a particular tomato, exists as well, only that the existence of some red objects or other is required for the existence of the universal, redness.

view quite naturally gives rise to the consequence that Socrates can exist only if his form exists. But nearly all the assumptions used in generating this example are highly controversial and can be (and are) rejected by other philosophers who do not subscribe to this particular version of hylomorphism. Perhaps, it is sufficient for our purposes to keep in mind that even a philosopher like Simons, who is sympathetic to a modal/existential construal of ontological dependence, feels the need to add an exclusion clause for proper parts in his formulation of “Strong Dependence,” since he allows for the possibility that entities which are allegedly highly deserving of substance status can ontologically depend on their own proper parts in the sense of (ND1).   See also Simons’ notion of “Generic Dependence” (Simons (1987), p. 297).

14

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

144  Ontological Dependence (ND2), however, is overly coarse-grained for much the same reasons as (ND1) is: existentially generalized versions of all of the same counterexamples that were seen to arise for (ND1) can be generated for (ND2) as well. Suppose, for instance, that a certain concrete particular object, such as Socrates, turns out to be rigidly existentially necessarily dependent, in the sense of (ND1), on certain of his constituents, his origins, the number 8, certain necessary properties, or necessarily coexistent four-dimensional entities. Then, by (ND2), Socrates will also turn out to be generically existentially necessarily dependent on some constituents or other, some entities from which he originated or other, some necessary existents (e.g., numbers) or other, some necessary properties or other, as well as some necessarily coextensive four-dimensional entities or other. When (ND2) is interpreted as yielding a criterion of substancehood, these entities on which Socrates generically necessarily depends will then in turn be ranked as more deserving of substance status than Socrates, which, I take it, would be generally agreed to be an unwelcome result. And even if we leave aside the second explanatory goal (viz., the formulation of a plausible criterion of substancehood), (ND2) also leads to undesirable consequences with respect to the first criterion (viz., the adequate classification of paradigmatic cases of ontological dependence). For, to illustrate, according to (ND2), a universal (e.g., redness), for example, turns out to be generically existentially necessarily dependent not only on its instances (viz., the particular red objects), but also on numbers, triangles, sets, and everything else that can plausibly be taken to exist necessarily. Thus, while the notion of dependence defined in (ND2) may achieve some explanatory goals (e.g., it may be of help in characterizing the ­relationship between an Aristotelian universal and the particulars that instantiate it), it certainly cannot single-handedly satisfy all that we expect from a notion of ontological dependence.

5.2.3  Other Forms of Existential Dependence 5.2.3.1  Necessary vs. Essential Existential Dependence A persistent problem which plagues all modal/existential construals of ontological dependence, as noted earlier, is that they appear to be too coarse-grained to yield the correct results in cases involving necessarily existing entities. For example, as it stands, all of the modal/existential construals of ontological dependence considered so far seem to misclassify the relationship between concrete particular objects, such as Socrates, and the number 8 or numbers in general. For example, according to (ND1), Socrates would be classified as rigidly existentially necessarily dependent on the number 8, since necessarily Socrates exists only if the number 8 exists, given our assumption that the number 8 exists necessarily. Similarly, (ND2) classifies Socrates as generically existentially necessarily dependent on numbers in general, since necessarily Socrates exists only if some numbers or other exist, given our assumption that numbers in general exist necessarily. Moreover, we also noted that, independently of the question of whether a particular criterion succeeds in yielding a plausible criterion of substancehood, (ND2)

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

varieties of existential dependence  145 also leads to the unattractive result that the universal, redness, for example, turns out to be dependent on necessarily existing entities, such as numbers, triangles, and sets. And while one might have the reaction that these apparent difficulties could be avoided by simply excluding necessary existents from the scope of the relations in question, such a restriction is unattractive in view of the fact that some of the items to which ontological dependence relations are intended to apply are themselves presumed to be necessary existents (e.g., God, Platonic universals, sets, etc.).15 One possible solution to these problems (endorsed by E. J. Lowe and Kit Fine) is to adopt a non-modal conception of essence (see Section 1.6). Those who are attracted to such a conception of essence now have the option of formulating stronger essentialist versions of (ND1) and (ND2), as illustrated by Lowe’s (ED1) and (ED2): (ED1)  Rigid Existential Essential Dependence: x is rigidly existentially essentially dependent on y ≡def It is part of the essence of x that x exists only if y exists. (ED2)  Generic Existential Essential Dependence: x is generically existentially essentially dependent on Fs ≡def It is part of the essence of x that x exists only if some Fs exist. Presumably, given an appropriately constrained non-modal conception of essence, (ED1) and (ED2) are not open to the same range of apparent counterexamples as (ND1) and (ND2). To illustrate, even though it is a de re necessary truth about Socrates that he exists only if the number 8 or numbers in general exist, it is not similarly plausible to think that what it is to be Socrates has anything to do with numbers. Hence, with a suitably narrow conception of essence in hand, we can resist classifying the proposition that Socrates exists only if the number 8 exists or the proposition that Socrates exists only if numbers in general exist as an essential truth about Socrates. In what follows, when I speak of “essence,” “essential truths,” etc., I have in mind such a suitably constrained non-modal conception of essence. Unfortunately, even if we move from modal to non-modal formulations of existential dependence, there are still reasons to be dissatisfied with purely existential construals of ontological dependence from the point of view of meeting the explanatory goals set out earlier. Potential trouble cases for non-modal existential construals of ontological dependence include the following: Socrates’ humanity (which may be taken to be, depending on one’s outlook, for example, a trope that essentially belongs to Socrates or 15   Whether Aristotle’s counterfactual dependence claim in the Categories, when interpreted existentially (“If the primary substances did not exist, it would be impossible for the other entities to exist as well”), also misclassifies these cases depends on how one deals with counterfactuals with impossible antecedents, such as “If the number 8 did not exist, it would be impossible for Socrates to exist as well” or “If numbers in general did not exist, it would be impossible for Socrates to exist as well.” Certainly, if counterfactuals with impossible antecedents are treated as trivially true, then this counterfactual/existential construal of ontological dependence will yield the same counterintuitive results as (ND1) and (ND2) with respect to cases involving necessarily existing entities.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

146  Ontological Dependence a universal that is essentially instantiated by Socrates); Socrates’ form (according to a hylomorphic analysis of concrete particular objects); Socrates’ essential proper parts (if he has any); or Socrates’ origin (assuming Kripke’s Essentiality of Origins thesis). If in at least some of these cases it is plausible to think that it is part of the essence of Socrates, say, that he exists only if the entity in question exists, then such concrete particular objects as Socrates will again be classified as existentially dependent on other entities numerically distinct from themselves even according to (ED1) and (ED2). In this way, they will be ranked as less deserving of substance status than these other entities on which they are existentially dependent, provided that (ED1) or (ED2) are used for the purposes of formulating an independence criterion of substancehood. There is no doubt much that could be said by the defender of a non-modal existential conception of ontological dependence about each of the items on this list of apparent trouble cases. For example, one could deny that concrete particular objects are hylomorphic compounds; or that they are numerically distinct from their forms; or that they have essential proper parts; or that their origins are essential to them; and so on. One rather popular strategy which, as I argue in Koslicki (2013b) and in Chapter 6, is to be avoided is simply to exclude by stipulation some group of apparent counterexamples (e.g., universals) from one’s favorite definition of ontological dependence or from one’s preferred criterion of substancehood. Those who endorse this strategy violate the third of the four explanatory goals identified earlier (viz., to allow for meaningful non-existential disagreements in ontology over questions of fundamentality), since questions which should be considered to be substantive are then classified as either trivially answerable (because their answers follow straightforwardly from a definition) or as resulting in an inconsistency or incoherence (if it is assumed, for example, that by definition particulars are more deserving of substance status than anything else). For example, such apparently substantive disputes may concern the question of whether universals or particulars deserve to be included among the absolutely fundamental entities (e.g., among the substances simpliciter); or whether universals should be designated as more fundamental (e.g., as more deserving of substance status) than particulars or vice versa. 5.2.3.2  Rigid and Permanent Existential Dependence Fabrice Correia and Benjamin Schnieder have also put forward an existential construal of ontological dependence which is not purely modal, since it makes use of the connective, “because,” on the right-hand side of the definition. This connective, due to its explanatory nature, is not assumed to be open to an analysis in exclusively modal terms. The following definition of “Rigid Permanent Existential Dependence” (RPED) comes from Schnieder (2006):16 16   Schnieder (2006), p. 412. (My statement of (RPED) is slightly different from Schnieder’s “(Dep-7),” but only in stylistic ways.) Schnieder states that he prefers an “innocuous” interpretation of the quantifier in “There is an F . . .,” which apparently ranges over properties, along the lines of Rayo and Yablo (2001) (Schnieder (2006), p. 416, n. 26). Correia’s notion of “basing” (Correia (2005), pp. 66ff.) and his definition of “simple dependence” in terms of “basing” are similar to Schnieder’s “(Dep-7)”; see also Correia (2008).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

varieties of existential dependence  147 (RPED)  Rigid and Permanent Existential Dependence: x rigidly and permanently existentially depends upon y ≡def There is an F, such that necessarily for any time, t, at which x exists, x exists at t because y is F at t. Schnieder considers the following to be a paradigmatic case of (RPED): a particular redness trope, in his view, is rigidly and permanently existentially dependent on a particular rose, say, in which it inheres, since there is an F, viz., redness, such that necessarily for any time, t, at which the rose’s redness trope exists, it exists at that time because the rose is red at that time. (RPED) strikes me as problematic for several reasons. To avoid the apparent counterexamples considered earlier to which modal/existential construals of ontological dependence fall prey, we have to assume that the explanatory connective, “because,” sets some constraints on how F may be picked relative to the objects, x and y, under consideration.17 Consider, for example, the following blatantly unhelpful attempt at instantiating (RPED): Socrates rigidly and permanently existentially depends on the number 8 just in case there is an F, e.g., actually numbering the planets, such that necessarily for any time, t, at which Socrates exists, Socrates exists at t because the number 8 actually numbers the planets at that time. Clearly, it should turn out not to be the case that Socrates depends on the number 8 in this way. And, needless to say, it does sound extremely odd to say that Socrates exists at any time, t, because the number 8 actually numbers the planets at that time: for the fact that the number 8 actually numbers the planets at any particular time strikes us as explanatorily completely irrelevant to the question of why Socrates exists at that time. But unless more is said about how to construe explanatory relevance and irrelevance in this context, the oddness of “Socrates exists at t because the number 8 actually numbers the planets at t” does not strike me as in any way illuminating of the oddness we already recognize in “Socrates’ existence at t depends on the number 8 actually numbering the planets at t.” If anything, it seems that the direction of illumination would have to go the opposite way: the fact that the number 8 actually numbers the planets at a time t is explanatorily irrelevant to Socrates’ existence at t because Socrates’ existence at t does not depend on the number 8 actually numbering the planets at t; and not conversely. In other words, it is not that two entities, facts, states of affairs, or what have you, stand in a dependence relation because an explanatory link obtains between them; rather, a good explanation should reflect an underlying dependence relation between the relata in question, as, for example, the job of a good causal explanation is to capture an underlying causal dependence relation.18 17   The explanatory connective, “because,” might be thought to indicate the presence of a grounding connection of some sort. Koslicki (2015a) lays out my reasons for thinking that a ground-theoretic approach to relative fundamentality is unsatisfactory when evaluated against the four measures of explanatory adequacy cited earlier. 18   Schnieder does provide some further elucidation concerning the kind of explanatory connection that he takes to be relevant to (RPED). In his view, the explanatory “because” in statements of ontological dependence is to be construed in an objective conceptual way: “In general, statements involving complex or elaborated concepts are explained with recourse to more primitive concepts (which may or may not

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

148  Ontological Dependence Suppose, on the other hand, that implicit guidelines are in place for how F in (RPED) is to be chosen relative to the entities, x and y. When the entities in question are, for example, the rose’s redness trope and the rose in which it inheres, then we may assume that it is somehow determined that the explanatorily relevant F in question that must be exhibited by the rose at each time at which the rose’s redness trope exists is redness (rather than, say, color), in which case the right-hand side of (RPED) reads as follows: there is an F, redness, such that necessarily for any time, t, at which the rose’s redness trope exists, it exists at t because the rose is red at that time. But now we ought to wonder whether (RPED) has really characterized the direction exhibited by the alleged explanatory connection in question correctly for the following reasons. Only certain kinds of trope theorists would find this particular instance of (RPED) congenial. As noted in Section 5.1, some trope theorists (who take the bearers of tropes to be mere bundles of tropes) would presumably think that the ontological dependence between tropes and their bearers runs in the opposite direction, i.e., that the bearers of tropes are ontologically dependent on the tropes that make up the particular bundle in question, analogously to the way in which non-empty sets are ontologically dependent on their members. So we should ask ourselves whether (RPED) in fact correctly represents the commitments of those trope theorists who are sympathetic to the idea that tropes ontologically depend on their bearers, and not the other way around. Under certain construals of the proposal under consideration, (RPED) fails to capture the explanatory link in question correctly. For one thing, the kind of trope theorist we have in mind might not take the existence of the rose’s redness trope at t to be really a distinct fact or state of affairs from the rose’s being red at t. If, on the one hand, we are dealing here with just a single fact, then the irreflexivity of the explanatory “because” has been violated by this putative instance of (RPED), since now a single fact or state of affairs is said to be explained in a circular fashion in terms of itself.19 If, on the other hand, the existence of the rose’s redness trope at t does, for the trope theorist at issue, constitute a distinct fact from the rose’s being red at t, then such a philosopher may want to explain the rose’s being red at t in terms of the existence of a redness trope that inheres in the rose at t, and not the other way around. After all, the trope theorist’s commitment to the existence of tropes has to do some philosophical work; otherwise, enter into an analysis of the complex concepts)” (Schnieder (2006), p. 405). For example, for Schnieder, the concept denoted by the phrase, “the rose’s redness,” is a complex or elaborated concept which is to be explained with recourse to more primitive concepts, e.g., those at play in the statement “The rose is red.” But it does not seem plausible to think that facts about ontological dependence in general can be explained by what is or is not classified as primitive or complex relative to a particular conceptual system. In some cases, nothing important may hang on whether one notion or another is taken as primitive by a particular conceptual system (e.g., point versus line in some systems of geometry). In other cases, one concept can be more complex than another, even though what the first stands for is ontologically more fundamental than what the second stands for (e.g., arguably, the concepts, water and H2O molecule, illustrate this possibility). 19   But see Schnieder (2010) for a response to this kind of worry. Schnieder argues there that “because” is sensitive not only to the identity of the facts introduced by its clauses, but also to the way in which these facts are presented.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

varieties of existential dependence  149 it seems that we could get by just as well without the commitment to tropes. According to either of these two construals, then, (RPED) does not correctly capture the relevant trope theorist’s conception of the relation between the rose’s redness trope and the rose in which it inheres.20

5.2.4  Being vs. Existence In addition to the more detailed objections to various specific existential construals of ontological dependence we have considered in the foregoing sections, there are also more general reasons for wanting to divorce ontological dependence from existential dependence (whether modally analyzed or not). Even though the putative cases of ontological dependence we have considered all seem to involve existential dependence, the existential construals of ontological dependence nevertheless do not really get us to the heart of what is at issue in the distinction between the fundamental and the derivative. For this reason, the existential definitions of ontological dependence considered in this chapter are all susceptible to worries coming from the fourth criterion for explanatory success by means of which approaches to fundamentality may be evaluated (Explanatory Fine-Grainedness). In a similar vein, Fine (1995a) comments on the coarse-grainedness of existential accounts of ontological dependence as follows: The present examples [viz., impossible objects and identity properties] highlight a problem that besets any existential account of dependence, whether it be modal or essentialist in form. For it does not seem right to identify the “being” of an object, its being what it is, with its existence. In one respect, existence is too weak; for there is more to what an object is than its mere existence. In another respect, existence is too strong; for what an object is, its nature, need not include existence as a part.  (Fine (1995a), p. 274; my emphasis)21

20   In addition, (RPED)’s alleged explanatory claims (e.g., “Necessarily, at any time t at which the rose’s redness trope exists, the redness trope exists at t because the rose is red at t”) also fail to illuminate the nature of the connections at issue and therefore lack the explanatory fine-grainedness required by the fourth criterion of adequacy. The arguments given in Koslicki (2015a)) (see Section 3.4, pp. 334–5) for ground-theoretic approaches can be adapted to establish this point for (RPED) as well. 21   The two examples Fine has in mind here (impossible objects and identity-properties) are as follows. Consider round squares: if it makes sense to say that round squares have natures, i.e., that there is such a thing as what it is to be a round square, then the use of the verb, “to be,” here cannot be construed in an existential way, since it is impossible for round squares to exist. Secondly, consider the property of being identical to Socrates. Suppose (as Fine does for the purposes of this example) that this property exists necessarily and suppose further that the property of being identical to Socrates ontologically depends on Socrates (i.e., it is that very property because it is the identity-property associated with Socrates). It would again be wrong to construe the relationship in question as existential dependence, since (by hypothesis) the identity-property exists necessarily while Socrates exists only contingently. Thus, there are worlds in which the identity-property exists while Socrates does not: in order for the identity-property in question to exist, it therefore cannot be required that Socrates exists as well; but the identity-property may nevertheless depend on Socrates with respect to its being what it is, its nature. I am not endorsing the details of either of Fine’s two examples, which are quite controversial; my purpose here is only to explain what he has in mind in the passage in question. We can agree with Fine’s general point (as I wholeheartedly do), regardless of how we feel about the particular examples he chooses to illustrate this point.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

150  Ontological Dependence It is tempting to think that the relationship between smiles and mouths, for example, is not exhausted by noting that while the existence of smiles requires the existence of mouths (pace the Cheshire Cat), the reverse is not the case. Rather, as we will explore in Section 5.3, what seems to move us closer to the core of the notion of ontological dependence is an idea central to non-modal essentialist approaches: that in saying what it is to be a smile we must appeal to mouths and what these mouths can do; but in saying what it is to be a mouth, we need not in turn appeal to smiles and what smiles can do. We would not have exhausted the full force of the relation of ontological dependence if we were to construe the use of the verb, “to be,” in the context of this “what it is to be” construction, in an exclusively existential way. Nevertheless, it does seem plausible to think (in a wide range of cases at least) that ontological dependence entails modal/existential dependence, i.e., that when an entity, x, ontologically depends on an entity, y, it is also true that necessarily x exists only if y exists (see also Lowe (1998)). The converse, on the other hand, does not always hold, as the trouble cases for the modal/existential approach rehearsed earlier illustrate: it is not true across the board that if necessarily x exists only if y exists, then x ontologically depends on y. I will now turn to the difficult question of what sort of alternative reading we might substitute for an existential construal of ontological dependence.22

5.3  Varieties of Essential Dependence 5.3.1  Essential Identity Dependence Among the various existential construals of ontological dependence we encountered in the previous section, several were formulated in terms of a non-modal conception of essence, viz., in particular, E. J. Lowe’s notion (ED1), “Rigid Existential Essential Dependence,” and (ED2), “Generic Existential Essential Dependence.” Lowe agrees that, among his defined notions, even (ED1) and (ED2) do not yet suffice to establish a genuine contrast between those entities which, in his view, are highly deserving of substance status or substances simpliciter and other entities which ontologically depend on them. To establish such a contrast, Lowe introduces a further species of essential dependence, viz., “Essential Identity Dependence”: (ED3)  Essential Identity Dependence: x is essentially identity dependent on y ≡def There is some function φ such that it is part of the essence of x that x = φ(y).23 22   In order to preserve this entailment from ontological dependence to modal/existential dependence, Fine’s example involving the property of being identical to Socrates would have to be addressed in some fashion. For the entailment from ontological dependence to modal/existential dependence seems to fail in this case. 23   (ED3) also has a modal counterpart (ND3), according to which x is necessarily identity dependent on y iff there is some function φ such that necessarily x = φ(y). But since (ND3) is still susceptible to the same sorts of apparent counterexamples to modal formulations of ontological dependence as the relations

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

varieties of essential dependence  151 As a paradigmatic instance of (ED3) Lowe cites, for example, the relation between a married couple and the individual members of the married couple, or the relation between a set and its members:24 in the first of these cases, according to Lowe, there is some function, φ, viz., the “marriage function,” such that it is part of the essence of the particular married couple in question that it is the result of applying the “marriage function” to the two individual members of the married couple; and, in the second, there is some function, ψ, viz., the “set-formation” function, such that it is part of the essence of Socrates’ singleton set that it is the result of applying the “set-formation function” to Socrates. Thus, married couples and sets, according to Lowe, are essentially identity dependent on their individual members. Since (ED3) entails (ED1) and (ED2), sets and married couples are also rigidly existentially essentially dependent on their members and, a fortiori, generically existentially essentially dependent on having some members or other. The first thing to note about (ED3) is that it does in fact take us beyond a purely existential construal of ontological dependence. (ED3) does not set a condition involving an entity, y, which must be met in order for an entity, x, to exist; rather, (ED3) sets a condition involving an entity, y, that is required for x’s identity, which presumably is not to be conflated with x’s existence. The substantive content of (ED3) concerns whether or not it is part of the essence of an entity that it is the result of applying a certain function (whose existence we can presumably uncontroversially presuppose) to another entity or entities. With (ED3), we have therefore moved beyond a purely existential construal of ontological dependence and thus squarely into the realm of essential dependence. To have a good grasp of what it takes for the condition set by (ED3) to be met or to fail to be met by a given entity would require that we know what Lowe means by “function” in this context and under what circumstances it is or is not part of the essence of an entity to be the result of applying a certain function to another entity or entities. Although Lowe is less explicit on these issues than we might have wanted him to be, he appears to be construing (ED3) with the notion of a criterion of identity in mind.25 In Lowe (1989), we are offered the following general schema for a criterion of identity, where “Φ” stands for a sortal term of some kind (e.g., “set”) and “R” stands for a considered in the previous section, I will in what follows leave (ND3) aside and focus instead on the essentialist version of identity dependence, (ED3). 24   Strictly speaking, in order for (ED3) to apply to these two examples, the definition of essential iden­tity dependence would have to be modified in something like the following way: “x is essentially identity depend­ent on a plurality of entities, y1 . . . yn ≡def There is a function φ such that it is part of the essence of x that x = φ(y1 . . . yn),” In what follows, I will simply assume that this modified definition is substituted for (ED3), wherever appropriate. 25   The model advanced in Lowe (1989) and (1997) for how to think of criteria of identity takes its inspiration from Frege’s remarks concerning lines and directions in the Foundations of Arithmetic: “The judgement ‘line a is parallel to line b’ . . . can be taken as an identity. If we do this, we obtain the concept of a direction, and say: ‘the direction of line a is identical with the direction of line b’” (Frege (1953), §64). Frege here seems to be offering the following criterion of identity for the directions of lines: if x is a line and y is a line, then the direction of x = the direction of y just in case line x and line y are parallel.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

152  Ontological Dependence relation in terms of which the criterion of identity in question is formulated (e.g., the relation of having the same members): (CI)

(∀x) (∀y ) ((Φx & Φy ) → (x = y ↔ Rxy ))

Lowe takes an instance of (CI) to be given, for example, by the Axiom of Extensionality for sets: if x and y are sets, then x and y are the same set just in case x and y have the same members; or, as Lowe would put it, which set is a certain set is fixed by which members the set in question has. For entities that exist in time, we are to construe (CI) for present purposes as yielding a synchronic criterion of identity or what may also be called a “principle of individuation,” i.e., a criterion that specifies what it takes for an entity to be the very entity that it is at a time, rather than a diachronic criterion of identity, i.e., a criterion that specifies what it takes for an entity to persist over time. With this in mind, we may now approach (ED3) as follows: an entity, x, is essentially identity dependent (or, for short, “(ED3)-dependent”) on an entity, y, when x’s relationship to y at a time, t, fixes which entity x is at t.26 In contrast, an entity would count as essentially identity independent of other entities according to (ED3) (or, for short, “(ED3)-independent”) when its synchronic identity at each time at which it exists is not fixed by its relationship to any other entities numerically distinct from itself. In this way, no synchronic criterion of identity or principle of individuation which appeals to numerically distinct entities will be available at all for (ED3)-independent entities: that they are the very entities they are at each time at which they exist is simply to be taken as a non-derivative fact about these entities. Thus, if Socrates is in fact to qualify as an (ED3)-independent entity, then it must be the case that he does not owe his individuation or synchronic identity, i.e., his being the very entity that he is at each time at which he exists, to his relationship to any other entity numerically distinct from himself.27 26   One complication in applying (CI) to (ED3) is that (CI) is formulated in terms of sortal concepts, whereas (ED3) concerns individual entities. I will assume, however, that we can apply (CI) to (ED3) in the following way: since, for example, the identity of sets in general is fixed by appeal to their members, the identity of every individual set is therefore fixed by appeal to its particular members. For more discussion of Lowe’s views concerning the relation between criteria of identity and principles of individuation, see Lowe (2012a). 27   In Lowe’s view, the unavailability of a criterion of individuation or synchronic identity for (ED3)independent entities is compatible with the availability of a diachronic criterion of identity for such entities. Thus, for Lowe, composite concrete particular objects persist over time just in case a certain equivalence relation holds that is defined over their actual or possible components (see Lowe (1989), p. 168). But why the asymmetry between identity at a time and identity over time? Why is Socrates’ status as an allegedly (ED3)-independent entity not threatened by the fact that he persists over time due to a certain relationship that his proper parts at one time bear to his proper parts at another, while his status as an allegedly (ED3)-independent entity apparently would be jeopardized if something similar were true of his identity at a time? Presumably, the answer to this question, for Lowe, is that diachronic identity presupposes synchronic identity. For suppose that it is possible for Socrates (say) to have existed for only one instant: in that case, his existence at this instant would nevertheless have required that he is the very entity that he is at that instant; but his synchronic identity at that instant does not require that he persists over time. If, on the other hand, Socrates persists over time, his diachronic identity does require that, at each time at which he exists, he is at that time (i.e., synchronically) identical to himself.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

varieties of essential dependence  153 A brief preliminary evaluation of (ED3) with respect to the first and second criteria for explanatory success cited in Section 5.1 (“Classification of Phenomena”; “Criterion of Substancehood”) yields the following results. Firstly, the classification of entities to which (ED3) gives rise seems to run into trouble with the essentiality of origins. For if it were part of Socrates’ essence, for example, to have originated from a particular zygote, then it might seem that a criterion of individuation or synchronic identity could be found for an allegedly (ED3)-independent entity such as Socrates, viz., one which appeals to Socrates’ origins.28 In this case, however, Lowe can avail himself of the option of rejecting the essentiality of origins, which he finds implausible in any case. Secondly, the classification of entities resulting from (ED3) also appears to conflict with a hylomorphic analysis of concrete particular objects as compounds of matter and form. For if it were part of the essence of an allegedly (ED3)-independent concrete particular object, such as Socrates, to be a compound of matter and form, then it might appear again that Socrates can be individuated by appeal to his form or matter. In this case, Lowe avoids the apparent conflict in question by arguing that hylomorphic “compounds” should be identified with their form and therefore are not strictly speaking compounds of matter and form at all (see Lowe (1999)). Thirdly, if an allegedly (ED3)-independent concrete particular object, such as Socrates, can have essential tropes (e.g., Socrates’ particular instance of humanity), then it will have to turn out that even though it is part of the essence of the tropes in question that they are the result of applying some function to their bearer (e.g., the “abstraction” function), it is not similarly part of the essence of the concrete particular object which is the bearer of these essential tropes that it is the result of applying some function (e.g., the “construction” function) to its essential tropes, presumably because such allegedly (ED3)-independent concrete particular objects are not taken to be bundles of tropes, on Lowe’s conception.29

28   However, it should be noted that it is a controversial question even among those who subscribe to a modal conception of essence whether concrete particular objects can in fact be individuated across worlds by means of their origins. See, for example, Forbes (1985, 1986, 1997, 2002); Mackie (2002, 2009); and the references to be found therein. The question of how hylomorphic compounds are to be individuated across worlds is discussed in more detail in Section 3.4.3. I argue there that sameness of origins is not sufficient for the cross-world identity of hylomorphic compounds. 29   Earlier (see Section  5.2.3.1), I brought up the following apparent counterexamples to (ED1) and (ED2): Socrates’ humanity, conceived of either as (1) an essential trope or as (2) a universal that is essentially instantiated by Socrates; (3) Socrates’ form (according to a hylomorphic conception of concrete particular objects); (4) Socrates’ essential proper parts (if he has any); or (5) Socrates’ origin (assuming the essentiality of origins). As noted just now, Lowe has a way of defending (ED3) against (1), (3), (4), and (5). (2) is not a problem for (ED3), since the identity of a particular cannot be parasitic on its being an instance of a universal, even where the universal in question is essential to it, since other particulars may instantiate the same universal. With respect to (4), Lowe would presumably similarly deny that a concrete particular object, such as Socrates, which is (ED3)-independent by Lowe’s lights, has any essential proper parts which determine his synchronic identity, since in Lowe’s view no numerically distinct entity (including proper parts) determines that an (ED3)-independent entity is the very thing that it is at any time at which it exists.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

154  Ontological Dependence

5.3.2  Constitutive Essential Dependence Fine (1995a) defines the following essentialist notion of ontological dependence, which I will call “Constitutive Essential Dependence”: (ED4)  Constitutive Essential Dependence: x is constitutively essentially dependent on y ≡def y is a constituent of x’s essence (narrowly construed). Smiles, for example, on this account, ontologically depend on mouths in the sense that mouths are constituents of the essences of smiles; but the reverse is not the case, i.e., smiles are not also constituents of the essences of mouths.30, 31 Essence, on this construal, cannot be understood in the traditional modal way, for reasons laid out in Fine (1994a). Such a construal of essence would not capture the asymmetric manner in which, for example, singleton sets ontologically depend on their sole members, while their sole members do not in turn ontologically depend on the singleton sets of which they are the sole members. Rather, in Fine’s view, the being, nature, or essence of an object, x, is the collection of propositions that are true in virtue of x’s identity, where “it is true in virtue of the identity of x that …,” for Fine, denotes an unanalyzed relation between an object and a proposition. The collection of propositions which are true in virtue of x’s identity can also be considered a real definition of x. (Real definitions contrast with nominal definitions and concern objects themselves, rather than the linguistic expressions we use to refer to objects or the concepts we use to conceive of them; see Koslicki (2012a), Sections 7.4–7.5, for further discussion.) Fine’s approach to ontological dependence crucially relies on a distinction between essence, narrowly construed (“constitutive essence”), and essence, more widely construed (“consequential essence”). Unless some such “narrow/wide” distinction for 30   Fine’s account assumes that we may think of essences as propositions or collections of propositions and that these propositions have constituents. 31   How does Fine’s account fare in the face of the apparent counterexamples considered in Section 5.2.3.1 in connection with (ED1) and (ED2)? Whether Fine’s account runs into trouble with respect to (5), the essentiality of origins, or (3), the status of hylomorphic compounds as entities which are allegedly (ED4)-independent from all other numerically distinct entities, depends crucially on whether we find the stipulative exclusion of proper parts in the formulation of an independence criterion of substancehood to be admissible. This strategy would also take care of (4), essential proper parts, more generally. If the ontologically independent entities are just those which are (ED4)-independent from all other numerically distinct entities except for their own proper parts, then an entity would be able to qualify as ontologically independent even if its proper parts figure as constituents in its constitutive essence or real definition, so long as no other numerically distinct entities besides its proper parts figure in its constitutive essence or real definition. If, for example, the zygote from which a human being originated at one point was a proper part of it and if the form and matter of which a hylomorphic compound consists are proper parts of it (as Fine holds), then (given the exclusion of proper parts) concrete particular objects could nevertheless qualify as (ED4)-independent even if it is part of their essence to have originated from whatever entity they actually originated from and even if it is part of their essence to be a compound of some particular matter and form. I investigate these questions further in Koslicki (2013b) and in Chapter 6. As far as cases (1) and (2), essential tropes or universals, are concerned, Fine can avail himself of the same strategy that is open to Lowe as well.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

varieties of essential dependence  155 essences can be drawn, Fine’s account of ontological dependence threatens to become vacuous, since everything will turn out to depend ontologically on everything else. This result follows because, in whatever way exactly we draw the constitutive/consequential distinction, the consequential essence of any entity, on Fine’s conception, will be closed under logical consequence and all the logical truths will therefore end up in the consequential essence of everything whatsoever. Since, for example, the proposition that the number 2 is self-identical is a logical truth, the number 2 will turn up as a constituent in the consequential essence of every object whatsoever. And because the number 2 here was picked arbitrarily and every object is self-identical, every object will by the same reasoning turn up as a constituent in the consequential essence of every other object. Thus, if an entity were to count as being ontologically dependent on all those objects which figure as constituents in propositions that belong to its consequential essence, then the notion of ontological dependence would have been trivialized and every object would turn out to be ontologically dependent on every other object. It is thus important, at least for the purposes of providing an informative account of ontological dependence in terms of essence, that the approach in question can avail itself of some more restrictive conception of essence than what is given by the notion of consequential essence. To this end, Fine considers two distinct methods by which to draw the constitu­ tive/consequential distinction for essences.32 The first proposal for how the constitutive/ consequential distinction might be drawn is outlined in the following passage: A property belongs to the constitutive essence of an object if it is not had in virtue of being a logical consequence of some more basic essential properties; and a property might be said to belong to the consequential essence of an object if it is a logical consequence of properties belonging to the constitutive essence (a similar account could be given for the case in which the essence is conceived in terms of propositions rather than properties). (Fine (1995a), p. 276; his emphasis)

Since the “more basic essential properties” in question presumably just are the ones that figure in the constitutive essence of an object, the method outlined here amounts to taking as basic the notion of constitutive essence and defining that of consequential essence in terms of it by way of logical closure. A proposition then belongs to the consequential essence of an object, according to this first approach, if it is a logical consequence of a proposition that belongs to the object’s constitutive essence (which is not itself to be taken as closed under logical consequence). For example, if the proposition that Socrates’ singleton set contains Socrates as its sole member belongs to the constitutive essence of Socrates’ singleton set, then the proposition that Socrates’ singleton set contains some member or other is admitted into the consequential essence of Socrates’ singleton set by logical closure. Since the logical truths are logically entailed 32   I am here relying primarily on Fine (1995a), pp. 276–80; but similar thoughts (though presented in a more condensed fashion) are also found in Fine (1995b), Sections 3–4.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

156  Ontological Dependence by any proposition whatsoever, these propositions will end up in the consequential essence of any object whatsoever by this procedure. In case the idea of taking the notion of constitutive essence as basic is found to be objectionable in a context in which our aim is to give an account of ontological dependence, Fine also considers a second approach to the constitutive/consequential distinction: It is therefore preferable, in the interest of conceptual economy, to see if the notion of dependence can be explained in consequential terms, without appeal to an underlying constitutive conception. To this end, we need an independent way of distinguishing between those objects that enter into the consequential essence as a result of the logical closure and those that enter in “their own right”, i.e., by way of the constitutive essence. But this is readily done. For when an object enters through logical closure, it can be “generalized away”.  (Fine (1995a), p. 277)

According to Fine’s second proposed method of drawing the constitutive/consequential distinction, we are to take as basic the notion of consequential essence and define that of constitutive essence in terms of it. The central idea underlying this second procedure is this: if an object enters as a constituent into a proposition belonging to the consequential essence of another object only through logical closure, then such an object can be “generalized away.” For example, the proposition that the number 2 is self-identical belongs to the consequential essence of Socrates’ singleton set; but so does, for every object whatsoever, the proposition that that object is self-identical. In this way, the number 2 can be “generalized out” of the proposition that the number 2 is self-identical, which belongs to the consequential essence of Socrates’ singleton set. Following this second method of drawing the constitutive/consequential distinction, then, only those objects which function as constituents of propositions belonging to the consequential essence of a given entity and which cannot be “generalized out” of these propositions make it into the constitutive essence of the entity in question as constituents of propositions belonging to its constitutive essence. In this way, it seems that Fine can avoid the result that Socrates’ singleton set ontologically depends on the number 2 (and, more generally, on any object whatsoever).33

33   The following is a more precise characterization of the notion of “generalizing out” (see Fine (1995a). pp. 277–8). Consider a proposition P(y), which has an object, y, as a constituent. For example, P(y) might be the proposition that Socrates is identical to Socrates for y = Socrates. Fine’s first step is to define the notion of a “generalization” for propositions, rather than objects (i.e., constituents of propositions): the generalization of a proposition, P(y), is the proposition that P(v) holds for all objects, v. Thus, the generalization of the proposition that Socrates is identical to Socrates is the proposition that for all objects, v, that v is identical to v. (To obtain the generalization, P(v), of a proposition, P(y), all occurrences of the constituent, y, must be replaced by occurrences of v.) Given the notion of a generalization, defined for propositions, we can now make sense of the idea that an object can be “generalized out” of a collection, C, of propositions in the following way: an object, y, can be generalized out of a collection, C, of propositions if C contains the generalization of a proposition P(y), whenever it contains the proposition P(y) itself. Finally, these defined notions are now applied to the analysis of ontological dependence in the following way: an object, x, depends ontologically on an object, y, according to this method of drawing the constitutive/consequential distinction, just in case y cannot be generalized out of the consequential essence of x, i.e., just in case some

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

varieties of essential dependence  157 The trouble is that Fine’s second method of drawing the constitutive/consequential distinction for essences (viz., taking for granted consequential essence and defining constitutive essence in terms of it), as it stands, cannot be assumed to take us all the way to an object’s constitutive essence. The second proposed method of drawing the constitutive/consequential distinction is based on the idea that those constituents that are intuitively irrelevant to the essential nature of the entity under consideration can be “generalized out” of propositions belonging to an object’s consequential essence. But Fine’s “generalizing out” procedure is really effective only in removing the logical truths from the consequential essence of an object, since this procedure takes advantage of a special feature of logical truths, viz., that they remain true under all re-interpretations of the non-logical vocabulary. The “generalizing out” procedure thereby leaves us with a restricted conception of consequential essence, viz., the collection of propositions consisting of an object’s unrestricted consequential essence minus the logical truths. But this restricted notion of consequential essence cannot in general be expected to deliver the suitably narrow conception of essence that is needed for an account of ontological dependence along the lines of (ED4). For, unless the narrow notion of constitutive essence is already implicitly presupposed (as is done by the first method of drawing the constitutive/consequential contrast), the entities on which an object depends ontologically need not match exactly those which figure as constituents in propositions that belong to an object’s restricted consequential essence. Consider, for example, the proposition that the number 2 is not a member of Socrates’ singleton set. If this proposition belongs to the unrestricted consequential essence of Socrates’ singleton set, then, in accordance with the second method of drawing the constitutive/consequential distinction just considered, the “generalizing out” procedure will pass it on into the constitutive essence of Socrates’ singleton set as well. Since this proposition is not a logically necessary truth, it does not remain true under all reinterpretations of the non-logical vocabulary. For it is not true in general for every object, v, that v is not a member of Socrates’ singleton set, since Socrates, after all, is a member of Socrates’ singleton set. The number 2 therefore will not “generalize out” of the proposition that 2 is not a member of Socrates’ singleton set and it seems the proposition in question, by the second method of drawing the constitutive/consequential distinction, will therefore wrongly end up in the constitutive essence of Socrates’ singleton set with the result that Socrates’ singleton set will again turn out to depend ontologically on the number 2 (and, by the same reasoning, on every object whatsoever), on the assumption that an entity ontologically depends on just those objects that appear as constituents in propositions belonging to its constitutive essence. Suppose, on the other hand, that the proposition that the number 2 is not a member of Socrates’ singleton set is excluded from the unrestricted consequential essence of Socrates’ singleton set. Now, it seems, the second method of drawing the constitutive/ proposition P(y) belongs to the consequential essence of x without it being the case that the generalization of P(y) also belongs to the consequential essence of x.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

158  Ontological Dependence consequential distinction has really collapsed into the first. For it is difficult to see on what grounds the proposition that the number 2 is not a member of Socrates’ singleton set could be excluded from the unrestricted consequential essence of Socrates’ singleton set, unless it is excluded on the grounds that this proposition does not pertain to the essential nature of Socrates’ singleton set either directly or indirectly, by being logically entailed solely by those propositions that directly pertain to the essential nature of Socrates’ singleton set. But if these are, in fact, at least implicitly the grounds on the basis of which the proposition in question is to be excluded from the unrestricted consequential essence of Socrates’ singleton set, then the procedure for determining whether a proposition belongs in an object’s unrestricted consequential essence is tacitly defined by reference to an object’s constitutive essence. And this is just the way in which the first method of drawing the constitutive/consequential distinction proceeds, viz., by taking as basic constitutive essence and defining unrestricted consequential essence in terms of it by way of logical closure. In light of these considerations, it thus seems that there is really only one method by which to approach the constitutive/consequential distinction, namely, the first one: to take as basic constitutive essence and define consequential essence in terms of it by means of logical closure, restrictedly or unrestrictedly.34 And while there is nothing wrong in principle with taking as basic constitutive essence, we should note that, in the context of giving an account of ontological dependence in terms of constitutive essence, Fine’s first proposed procedure for drawing the constitutive/consequential contrast does not really give us an independent handle on the notion of ontological dependence. Fine himself is well aware of this feature of his account: It is, of course, no surprise that dependence can be defined in terms of the objectually constrained form of essential truth; for the notion of dependence is already built into the constraints by which the relevant notion of essential truth is understood. But even without the constraints, a definition could still be given. For we may say that x depends upon y just in case, for some property φ not involving y, it is true in virtue of the nature of x that y φ’s and yet not true in virtue of the nature of x that every object φ’s; the dependees are the objects which cannot be “generalized out”. Thus we do not have, in the notion of dependence, an idea that is genuinely new.  (Fine (1995c), p. 243)

Thus, although nothing prevents us from defining a notion of ontological dependence in terms of constitutive essence as proposed here, we should be conscious of the fact that we have not thereby accomplished more than to state what is at bottom a single ontological relationship in two different, but interdefinable, ways, viz., either by means of a suitably restricted notion of essential truth or by means of a suitably restricted notion of ontological dependence. To illustrate, it helps to return once more to the proposed asymmetry in the relation between Socrates and Socrates’ singleton set. Thus, to say that it is a properly constitutive essential truth about Socrates’ singleton set   See also Correia (2013) for discussion.

34

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

varieties of essential dependence  159 that it has Socrates as its sole member (while it is not also a properly constitutive essential truth about Socrates to be the sole member of Socrates’ singleton set) is really just another way of saying that Socrates’ singleton set ontologically depends on Socrates (while Socrates does not also ontologically depend on Socrates’ singleton set); and vice versa. As Fine puts it in the passage just cited: “ . . . the notion of dependence is already built into the constraints by which the relevant notion of essential truth is understood”; and, given (ED4), a suitably narrow conception of essence is also already built into the constraints by which the relevant notion of ontological dependence is understood.35

5.3.3  Constitutive Definitional Dependence Fine assumes that, for the purposes at hand, essences can be identified with collections of propositions that are true in virtue of the identity of the particular object or objects whose essences they are. Such collections of propositions that are true in virtue of the identity of an object or objects, in Fine’s view, can also simultaneously be thought of as real definitions for the object or objects in question. There is not, then, on this approach, much of a distinction between essences and real definitions. But we may wish to proceed somewhat differently and leave room for a less propositional conception of essences, such as that endorsed by Aristotle, for example. As was noted in Chapter 3, the essence of a thing for Aristotle includes at least its form. (Whether the essence of a thing also includes additional components besides the form, e.g., the matter, is a controversial question which was briefly discussed in Section 3.4.2.) For example, the essence of a living being, in Aristotle’s view, encompasses at least its soul, i.e., the form of the living being. But, given Aristotle’s association of the soul with 35   As noted earlier (see Section  4.5), Aristotle, like Fine, also recognizes a distinction between what belongs to the essence proper of an object and what merely follows from its essence proper, i.e., the so-called “propria” or “necessary accidents.” But Aristotle’s “narrow/wide” distinction for essences does not exactly line up with Fine’s. For example, for Aristotle, it is part of the essence of planets that they are heavenly bodies that are near; but it is merely a necessary (but non-essential) feature of planets that they do not twinkle. The latter proposition, in Aristotle’s view, states a feature which merely follows from, but is not itself included in, the essence of planets. But the relevant notion of “following from” that is operative in this context, for Aristotle, cannot simply be that of logical consequence. The (explanatorily less basic) proposition that planets are heavenly bodies which do not twinkle is logically entailed by the (explanatorily more basic) proposition that planets are heavenly bodies that are near together with auxiliary premises (e.g., the proposition that heavenly bodies that are near do not twinkle). But the same holds also in the opposite direction: the (explanatorily more basic) proposition that planets are heavenly bodies that are near is also logically entailed by the (explanatorily less basic) proposition that planets are heavenly bodies that do not twinkle together with auxiliary premises (e.g., the proposition that heavenly bodies that do not twinkle are near). Thus, the relation of logical entailment alone (as is brought out in a more contemporary context by Sylvain Bromberger’s “flagpole” objection against Hempel’s deductive-nomological model of scientific explanation) is not sufficient to capture the asymmetry of scientific explanation. In Aristotle’s system, the relevant notion of asymmetric consequence that is operative in his model of scientific explanation is that of demonstration, as developed theoretically in his Posterior Analytics and (according to my reading of Aristotle) applied practically, for example, in his biological treatises. A scientific explanation is asymmetric, for Aristotle, because (if accurate) it is the theoretical and/or linguistic reflection of an asymmetric realworld relation of causal priority, where causation here must be construed along Aristotelian (and not Humean) lines. See Chapter 4 and Koslicki (2012b) for further discussion of these issues.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

160  Ontological Dependence certain kinds of capacities [dunameis], e.g., the capacity for nourishment, perception, and thought, it would be strange to think of the soul of a living being as a collection of propositions. It is perhaps more natural to take real definitions, which Aristotle regards as linguistic entities [logoi] of some sort, viz., formulae or statements of the essence, as collections of propositions or perhaps as only a single proposition, if there is only a single canonical way of stating the essence of a thing. The basic idea underlying Fine’s essentialist approach to ontological dependence can then be reformulated in terms of real definitions as follows: (ED5)  Constitutive Definitional Dependence: x is constitutively definitionally dependent on y ≡def y is a constituent of a real d ­ efinition of x.36

5.3.4  A Potential Difficulty I will end by mentioning a potential trouble case for the essentialist accounts of ontological dependence proposed by Lowe and Fine. (ED3), (ED4), and (ED5) have the following odd consequence. Assuming that it is in fact the case, following the Axiom of Extensionality, that sets are essentially dependent on their members, in the sense of (ED3)–(ED5), then no non-empty set would be classified as an ontologically independent entity or as deserving of substance status, which is perhaps a welcome result.37 But what about the empty set? Its identity cannot in any meaningful sense be regarded as fixed by the identity of its members, since it has none; nor, for the same reason, can its members appear as constituents in its constitutive essence or real definition. Rather, it seems that the identity of the empty set must simply be taken for granted and that its essence must be regarded as simple and non-relational, in the sense that no entities numerically distinct from itself figure as constituents in its constitutive essence or real definition. In that case, however, it seems that, by (ED3)–(ED5), the empty set qualifies as an ontologically independent entity and (assuming an essentialist independence criterion of substancehood) as deserving of substance status, even though no other set does. I take it that this would be an unfortunate consequence for an essentialist account of ontological dependence, since, given a certain taxonomic category of entities, presumably either all or none of the entities belonging to the category in question should count as ontologically independent and as deserving of substance status. It strikes me as strange to think that the empty set alone qualifies for substance status or deserves to be counted as ontologically independent, while none of the other sets do.38 36   The formulation of (ED5) in the text contains “a real definition,” rather than “the real definition,” since I want to leave open for the time being whether an entity can have more than one real definition. This possibility would obtain if two different propositions (or collections of propositions) are equally wellsuited to stating the essence of the entity in question. 37   On the relation between the Axiom of Extensionality and the claim that sets have their members essentially, see, for example, van Cleve (1985) and Forbes (1985). 38   Perhaps the culprit here is the innocuous-sounding sortal uniformity principle to which I have just appealed, according to which substance status is to be granted to entities not individually, but by sort. If we substitute “fundamental” for “deserving of substance status,” then the result just generated is perhaps less

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

varieties of essential dependence  161 A similar difficulty arises when we consider the relation between the number 0 and all the other natural numbers. If we think of the natural numbers as in some sense constructed from the number 0, together with the successor relation, then it seems that which number the number 1 is fixed by its relation to the number 0, viz., by its being the immediate successor of the number 0. The number 1, on this conception, is therefore naturally taken as having the number 0 as a constituent in its constitutive essence or real definition. But we cannot similarly conceive of the number 0 as being constructed from other natural numbers, together with the successor relation, since the number 0 is not the successor of any other natural number. Rather, in analogy with the set-theoretic case, the identity of the number 0 must be taken as fixed independently of its relationship to the other natural numbers and its constitutive essence again appears to be similarly simple and non-relational, in the sense that what it is to be that very number is not defined in terms of its relation to any other natural number. But it would be similarly puzzling if a criterion of substancehood classified the number 0 alone, and no other natural number, as ontologically independent or as deserving of substance status. An essentialist account of ontological dependence, along the lines of (ED3)–(ED5), ideally should have something to say in response to these considerations.39, 40 bewildering: (ED3)–(ED5), when construed as yielding a criterion for fundamentality, classify the empty set as fundamental (or as more fundamental than the other sets), while they classify the remaining sets as non-fundamental (or less fundamental than the empty set). Still, metaphysical realists at least will want to leave some room for a distinction between what is taken as primitive by a particular theory or conceptual system and what is ontologically fundamental as a matter of fact. See also my earlier remark (see n. 18) in response to Schnieder’s reliance on what is assumed to be primitive or complex relative to a given theory or conceptual system. 39   These and similar cases are discussed in Lowe (2012a); see also Schwartzkopff (2011). The following response to the concern raised in this section regarding the apparent inclusion of the empty set among the ontologically independent entities or substances was suggested to me by Allen Hazen (pers. comm.). The worry in question can be avoided by formulating a variant of ZF set theory which postulates only nonempty sets. Such a variant will be recognizable as the kind of theory to which mathematicians appeal under the name of “set theory” and will be structured by something recognizable as “The Iterative Conception of Set.” When such a variant is formulated, it is possible to interpret standard versions of set theory which postulate a null-set in terms of the non-null-set-postulating theory. (Here “interpretation” is understood in the standard technical sense according to which one theory is “interpreted” in a second if a “translation” of the language of the first theory is given into the language of the second theory which maps theorems of the first into theorems of the second.) For a supporting conception of the null set, see, for example, Fraenkel, Bar Hillel, and Levy (1973), p. 23. Hazen points out that for pure mathematics, pure sets suffice. For applications of set theory which include sets with independently given individuals as members, a coding is available which designates a certain suitable individual to play the role of the null set. (See Hazen (1991); Lewis (1991) for discussion.) Even if this kind of response is successful in the case of the empty set, however, we are still left to wonder how to proceed in the case of the number 0. 40   For further discussion of how essentialist construals of ontological dependence fare when evaluated with respect to the measures of explanatory success singled out in Section 5.1, see Koslicki (2012a) and (2016a). On the downside, as I argue in Koslicki (2012a), even essentialist construals of ontological dependence still lack the necessary level of explanatory fine-grainedness required to meet the fourth criterion, since they do not (at least in their current formulations) distinguish between different varieties of ontological dependence, e.g., what I call their “feature dependence” and “constituent dependence.” On the upside, however, as I argue in Koslicki (2016a), essentialist construals of ontological dependence are wellsuited to meet the third criterion: for they allow us to make sense of how non-existential disputes in ontology over questions of fundamentality can be substantive. The success of essentialist accounts in meeting the second criterion (Criterion of Substancehood) will be examined again in more detail in Chapter 6.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

162  Ontological Dependence

5.4 Conclusion In this chapter, I considered both existential and non-existential essentialist construals of ontological dependence. The main goal of Section 5.2 was to bring out why existential construals of ontological dependence, whether modal or non-modal, do not adequately reflect all that is encompassed by this notion. And while there is nothing wrong in principle with defining whatever technical concept one wishes, the question arises, in the face of the plethora of relations that go under the name of “ontological dependence,” what explanatory tasks these notions are designed to accomplish, and how well they in fact meet the desiderata that are set for them. Throughout this chapter, I evaluated particular definitions of ontological dependence by means of four potential measures of explanatory success: firstly, how well they do in classifying certain paradigmatic cases of ontological dependence in a particular desired way; secondly, whether the definitions in question contribute to the formulation of a plausible criterion of substancehood; thirdly, whether they make room for the possibility of substantive non-existential disagreements in ontology over questions of fundamentality; and, finally, whether they are sufficiently fine-grained to illuminate the nature of the connections at issue in putative cases of ontological dependence. Relative to these four measures, we found that existential construals of ontological dependence are open to persuasive counterexamples. In Section 5.3, I considered several essentialist construals of ontological dependence: the notion of essential identity dependence, proposed by E. J. Lowe; constitutive essential dependence, due to Kit Fine; and constitutive definitional dependence, a variant of Fine’s proposal. When evaluated relative to the four explanatory goals identified above, non-existential essentialist construals of ontological dependence have a more impressive track record than their existential counterparts. Nevertheless, various questions still remain to be addressed by essentialist approaches as well, in particular the question of whether and how hylomorphic compounds can be designated as substances simpliciter or as more deserving of substance status than other entities which are thought to be ontologically dependent on them. The alleged substance status of hylomorphic compounds will be our focus in Chapter 6.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

6 Independence Criteria of Substancehood 6.1  Introductory Remarks In Chapter 5, we evaluated a plethora of relations that are offered in the literature under the heading of “ontological dependence.” When measured against the four criteria of explanatory success laid out in Section 5.1, essentialist construals of ontological dependence overall turned out to outperform their existential counterparts. Nevertheless, several potentially problematic issues remained to be addressed by proponents of essentialist accounts as well. Prominent among these is the question of whether and how essentialist construals of ontological dependence, in compliance with the second explanatory goal (Criterion of Substancehood), are compatible with the idea that at least certain kinds of concrete particular objects (e.g., living organisms) deserve to be included in the category of substances. We noted that this challenge is especially significant for Aristotelians who are committed to the following three tenets: (i) a hylomorphic analysis of concrete particular objects; (ii) an independence criterion of substancehood; and (iii) the assignment of substance status to at least some hylomorphic compounds. For when we attempt to combine these three commitments, an apparent conflict emerges. Suppose, for example, as is common among Aristotelians, that living organisms are designated as substances. Following the first tenet, these entities are analyzed along hylomorphic lines as compounds of matter (hulē) and form (morphē). In that case, however, one wonders whether they will not turn out to be ontologically dependent on entities numerically distinct from themselves, viz., their form and possibly their matter as well. If, in accordance with the second tenet, the notion of substancehood is construed in terms of some preferred notion of ontological independence, such a criterion would then seemingly exclude all hylomorphic compounds from the category of substances and thereby lead to a conflict with the third tenet, viz., the assignment of substance status to at least some hylomorphic compounds. In this chapter, I will examine the apparent tension between these three commitments and will propose what I take to be the most promising strategy for Aristotelians who wish to resolve the apparent conflict in question in a way that is maximally consistent with their other commitments.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

164  Independence Criteria of Substancehood

6.2 Substancehood According to the third tenet popular among Aristotelians, at least some hylomorphic compounds are to be included among the substances. If an ontology makes room for the designation of certain types of entities as substances, then it must at least implicitly invoke some criterion or notion of substancehood in accordance with which entities are awarded this title. But how is substance status to be conferred on an entity and which entities, if any, qualify for it? Moreover, what hangs on the question of whether some type of entity is, or is not, to be included among the substances? These are among the most hotly debated questions in metaphysics at least since the time of Aristotle.1 Taxonomic vs. Non-Taxonomic Substancehood. Among the most important uses to which the concept of substance is put in philosophical contexts are what I will call a “taxonomic” and a “non-taxonomic” use. First, in its taxonomic role, philosophers employ this notion to single out certain kinds of entities (e.g., certain kinds of macroscopic concrete particular objects), without thereby simultaneously committing themselves to the idea that these entities must be assigned a position of “ontological privilege” within their respective ontologies. When the concept of substance is utilized in this first taxonomic way, the substances appear merely as one among many entries in a catalogue of beings. The resulting inventory might, for example, constitute an answer to the existential question, “What is there?,” which Quine saw as central to the discipline of ontology (see Quine (1948)). In other contexts, however, philosophers employ the concept of substance in a second non-taxonomic role, in order to indicate that certain kinds of entities (taxonomically speaking) deserve to be assigned a special place relative to the ontology in question. Much confusion has resulted over the years from a failure to distinguish between these two very different, but equally important, roles played by the concept of substance in philosophical contexts. To illustrate, consider Aristotle’s well-known list of ten categories in which “substance” appears as his first entry: Of things said without any combination, each signifies either substance or quantity or qualification or a relative or where or when or being-in-a-position or having or doing or being-affected. (Categories, Ch. 4, 1b25–27)

Here, Aristotle draws our attention to a ten-fold division among different kinds of being (taxonomically speaking). In order to set up a hierarchical ordering among these

1   See also the Introduction to Part II for a brief discussion of what hangs on the question of whether an entity is, or is not, included among the substances. I give a more detailed elaboration and defense of the continued philosophical usefulness of the concept of substance in Koslicki (2015b), in particular in response to arguments to the contrary proposed in Simons (1998). The question of how an Aristotelian substance-ontology is to be squared with contemporary science briefly came up in Chapter 2, “Brief Interlude: Metaphysics and Science,” though a proper treatment of these difficult issues concerning the relation between metaphysics and science would unfortunately take us too far afield for present purposes.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

substancehood  165 entities, Aristotle appeals to his two relations, being in a subject and being said of a subject, as follows: A substance—that which is called a substance most strictly, primarily, and most of all—is that which is neither said of a subject nor in a subject, e.g. the individual [human being] or the individual horse. The species in which the things primarily called substances are, are called secondary substances, as also are the genera of these species. For example, the individual [human being] belongs in a species [human being], and animal is a genus of the species; so these—both [human being] and animal—are called secondary substances. (Categories, Ch. 5, 2a11–19; Ackrill’s emphasis)2

Given the ultimate subject criterion of substancehood Aristotle endorses in the Categories, certain kinds of entities (taxonomically speaking), e.g., individual living organisms, turn out to be “most strictly, primarily, and most of all” deserving of the title, “substance,” in Aristotle’s Categories ontology, while others rank below them: the so-called “secondary substances” (viz., the species and genera to which the entities classified there as primary substances belong); as well as the non-substances (viz., the individuals, species, and genera belonging to the other nine taxonomic categories listed earlier). The concept of substance, then, can be, and has been, used by philosophers, among other things, in the service of at least two distinct tasks: firstly, to classify entities taxonomically; and, secondly, to impose a non-taxonomic ordering of some kind onto the entities included in a given ontology. The second project moreover requires us to engage with the further question of why certain items are to be assigned a privileged position within a particular ontology. For this reason, philosophical disputes concerning substancehood often center on the very criteria themselves by means of which substance status is awarded to certain taxonomic categories within a given ontology. Absolute, Comparative, and Relational Substancehood. The concept of substance in its philosophical applications can take on further explanatory roles depending on whether it is used in an absolute, a relational, or a comparative way: Absolute Substancehood: x is a substance simpliciter. Relational Substancehood: x is the substance of (or a substance of ) y. Comparative Substancehood: x is more deserving of substance status than y. In its first absolute role, the concept of substance is used to designate entities as substances simpliciter. In its second relational role, the concept of substance picks out a relation between pairs of entities, x and y, when x is the substance of, or a substance of, y. When used in the third comparative way, the concept of substance ranks entities by the degree to which they are deserving of substance status. All three of these explanatory roles are evident in Aristotle, but they are unfortunately not always clearly distinguished by Aristotle himself or his interpreters. For example,   I have taken the liberty of replacing “man” with “human being” in Ackrill’s rendition of this passage.

2

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

166  Independence Criteria of Substancehood in Categories, Ch. 5, 2a11–19 (cited earlier), Aristotle classifies individual living organisms, as well as their species and genera, as substances simpliciter, using an absolute notion of substancehood. In the very same passage, however, we find Aristotle identifying certain items as primary substances (e.g., individual living organisms) and others as secondary substances (e.g., the species and genera to which these individual living organisms belong). In this usage, Aristotle ranks entities in a comparative way, by the degree to which they satisfy the ultimate subject criterion at work in the Categories.3 The relational use of the concept of substance is operative in the passage cited earlier from Plato’s Euthyphro (see Part II: Introduction ); it is also employed, for example, by Aristotle in the opening lines of Met. Z.6: We must inquire whether each thing and its essence are the same or different. This is of some use for the inquiry concerning substance; for each thing is thought to be not different from its substance [tēs heautou ousia], and the essence is said to be the substance of each thing [hē hekastou ousia]. (Met. Z.6, 1031a15–18; my emphasis)

In this passage, Aristotle speaks of the essence of each thing, relationally, as the substance of that thing. If we think of the concept of substance, when employed in its non-taxonomic uses, as a marker of a certain kind of fundamentality, then comparative substancehood gives rise to a relative, rather than an absolute, fundamentality ranking of entities. Fundamentality, when construed in this relative way, is a notion that comes in degrees: it allows one to assess an entity, or type of entity, as more or less fundamental (or derivative) than some other entity, or type of entity. When we encounter alleged examples in which one type of phenomenon is purported to be more or less fundamental (or derivative) than another, we should not immediately assume that we are dealing with a case in which the more fundamental phenomenon in question can also be correctly described as absolutely fundamental. In the case of Plato’s Euthyphro dilemma, many philosophers, including Plato, would not be particularly tempted to say that something’s being pious is an absolutely fundamental matter, even if they agree that something’s being pious is more fundamental than its being loved by the gods.4 Thus, we should take care to separate questions which concern relative fundamentality from those which concern absolute fundamentality. Leaving Taxonomic Uses Behind. Armed with these distinctions, we can now return to the questions I posed earlier: are at least some matter–form compounds deserving of substance status? And, if so, according to what criterion or notion of substancehood? Assuming, as I do, that we are operating in a philosophical context in which the existence of concrete particular objects is not in dispute, I take it that the designation of concrete 3   See also, for example, Categories, Ch. 5, 2b7–8: “Of the secondary substances the species is more a substance than [mallon ousia] the genus, since it is nearer to the primary substance.” 4   Presumably, for the Plato of the middle period, only the form of the good would count as an absolutely fundamental entity. All the other entities in Plato’s middle period ontology (e.g., the other forms, the sensible particulars, or the mathematical objects) would receive only a comparative ranking depending on their relation to the form of the good.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

independence criteria of substancehood  167 particular objects as substances, in the purely taxonomic sense of the term, is primarily a matter of terminological convenience. To illustrate, a trope theorist (e.g., Campbell (1990) or Simons (1994)), who construes concrete particular objects as complex trope bundles, might nevertheless refer to these entities as “substances” and yet at the same time designate them as ontologically derivative. For, from the perspective of the tropetheoretic bundle theorist, concrete particular objects are further analyzable in terms of other entities (viz., tropes) which are designated, relative to the ontology in question, as either absolutely fundamental or as relatively fundamental, i.e., as more fundamental than concrete particular objects. In this way, then, the purely taxonomic designation of an entity as a substance can come apart from its occupation of an ontologically privileged position relative to a particular ontology. For the sake of clarity and simplicity, I will in what follows refrain, as much as possible, from using the term “substance,” taxonomically, i.e., as a mere classificatory shorthand for “concrete particular object” (unless I am paraphrasing another philosopher’s taxonomic use of the term). Instead, I will focus subsequently on non-taxonomic uses of the term, “substance,” and the proposed criteria of substancehood which are intended to go along with them. Our initial question concerning the possible assignment of substance status to at least some hylomorphic compounds can therefore be understood as asking whether at least some of these entities deserve to be assigned a privileged ontological position and, if so, according to what conception of “ontological privilege,” i.e. (absolute or relative) fundamentality.5

6.3  Independence Criteria of Substancehood There is a long tradition dating all the way back to Aristotle’s Categories of employing some notion of ontological independence in the formulation of a criterion of substancehood.6 At least as a first pass, we would expect an independence criterion of substancehood to include at a minimum the following necessary condition: (NICS)  Necessary Independence Condition for Substancehood: x is a substance only if x is ontologically independent.

5   Proponents of ground-theoretic approaches to fundamentality (e.g., Schaffer (2009)) may feel that an entity’s designation as a substance in the non-taxonomic sense can be adequately captured by means of the grounding idiom. As I have argued in Koslicki (2015a), however, the notion of ground, due to its coarsegrainedness, is not, in fact, a helpful tool to accomplish this philosophical work. 6   As noted earlier (see Section 5.2.1), when Aristotle states his ultimate subject criterion of substancehood in the Categories, he does not directly use a Greek term which is translated into English in terms of the vocabulary of dependence or independence. Nevertheless, Aristotle is commonly interpreted (and rightly so, in my view) as putting forward an independence criterion of substancehood in this text. For the two key relations, being said of a subject and being in a subject, in terms of which Aristotle formulates his ultimate subject criterion of substancehood in the Categories, should be read as indicating two ways in which all the other entities depend ontologically on those entities which are classified in the Categories as primary substances, while the primary substances do not in turn ontologically depend on anything else in either of these two ways.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

168  Independence Criteria of Substancehood For the time being, we may understand an entity’s status as ontologically independent to consist in that entity’s failure to depend ontologically on any other entity numerically distinct from itself (in accordance with some, as of yet, unspecified preferred notion(s) of ontological dependence): (IND)  Ontological Independence: x is ontologically independent ≡def There is no y numerically distinct from x such that x is ontologically dependent on y.7 (NICS) states only a necessary condition for an entity’s inclusion in the category of substances: at the very least, according to (NICS) and (IND), such an entity may not ontologically depend on any other entity numerically distinct from itself (according to some preferred construal of ontological dependence). To arrive at a criterion which yields both necessary and sufficient conditions for an entity’s inclusion in the category of substances, proponents of an independence criterion of substancehood may consider supplementing the necessary condition given so far in various ways. Below, in our discussion of proposals endorsed by E. J. Lowe and Michael Gorman, respectively, we will encounter several options for how independence theorists may wish to proceed in this connection. Moreover, a further decision must be made by proponents of an independence criterion as to whether they opt for a formulation in terms of particular entities or types of entities. I chose the former option here, because it is more in line with the formulations adopted by Lowe and Gorman and hence allows for greater continuity with our evaluation of their proposals below.8 According to (IND) and (NICS), substance status is awarded to an entity only if it does not ontologically depend on any other entity numerically distinct from itself (according to some preferred notion(s) of ontological dependence). As our discussion of Lowe and Gorman will illustrate, such independence criteria of substancehood are 7   Despite the apparent innocuousness of (NICS) and (IND), not all proponents of an independence criterion of substancehood would be happy with these formulations. For reasons which will concern us in more detail in Section 6.4, some independence theorists find themselves tempted to add various exemption clauses to the right-hand side of (IND), in order to disqualify an entity’s ontological dependence on certain entities that are numerically distinct from the entity in question as irrelevant to that entity’s status as a substance. (NICS) and (IND) are currently only intended as a starting point for our discussion; we will have occasion in what follows to consider various proposed modifications to (NICS) and (IND). 8   In addition to the supplementary conditions proposed by Lowe and Gorman, which will concern us in more detail in the next section, independence theorists may also wish to expand (NICS) by requiring that the substances, in addition to being themselves ontologically independent, must also act as a sort of “ontological anchor” for all the other entities that are included in the ontology under consideration. According to this expanded necessary condition, endorsed, for example. by Aristotle in the Categories, everything which does not qualify as a substance (i.e., every entity, y, that is not ontologically independent and hence ontologically depends on some relevant, z, numerically distinct from y) ontologically depends on something which does qualify as a substance (i.e., on something that is ontologically independent). In what follows, I argue that we have reasons to reject (NICS) as an adequate statement of a necessary independence condition for composite substances, even when it is weakened in the way that Lowe and Gorman propose. Thus, adding a further necessary condition to (NICS) in the manner just outlined does not improve the situation for proponents of an independence criterion of substancehood. Going forward, I therefore focus on the simpler necessary condition stated in (NICS).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

independence criteria of substancehood  169 typically interpreted by contemporary Aristotelians as giving rise to an absolute, rather than a relational or comparative, classification of entities.9 When read in this light, the notion of ontological dependence in terms of which the right-hand side of (IND) is interpreted has to be suitable to yield a “yes” or “no” answer to the question of whether an entity is or is not ontologically independent: either there is a y numerically distinct from x on which x ontologically depends; or there is no such y. Thus, when construed in this way, we can restate (NICS) as follows: (NICSS)  Necessary Independence Condition for Substancehood Simpliciter: x is a substance simpliciter only if x is ontologically independent. As noted in Chapter 5, a plethora of relations have been defined in the literature under the heading of “ontological dependence.” Some of the most popular definitions are formulated in modal terms; others in non-modal, e.g., explanatory or essentialist, terms; some (viz., the existential construals of ontological dependence) emphasize requirements which must be met in order for an entity to exist; others (viz., the nonexistential essentialist construals of ontological dependence) focus on requirements which must be met in order for an entity to be the very entity that it is at each time at which it exists; some are rigid, in the sense that they involve a relation between particular entities; others are generic, in the sense that they involve only a relation between an entity and some entities or other, which bear certain characteristics. I restate several of these definitions of ontological dependence here in order to facilitate our evaluation (or, in some cases, the reevaluation) of their usefulness for the purposes of formulating an independence criterion of substancehood: (ND1)  Rigid Existential Necessary Dependence: x is rigidly existentially necessarily dependent on y ≡def Necessarily x exists only if y exists. (ND2)  Generic Existential Necessary Dependence: x is generically existentially necessarily dependent on Fs ≡def Necessarily, x exists only if some Fs exist. (ED3)  Essential Identity Dependence: x is essentially identity dependent on y ≡def There is some function φ such that it is part of the essence of x that x = φ(y). (ED4)  Constitutive Essential Dependence: x is constitutively essentially dependent on y ≡def y is a constituent of x’s essence (narrowly construed). 9   In this respect, the contemporary practice of interpreting an independence criterion as yielding exclusively an absolute notion of substancehood diverges from Aristotle’s practice in the Categories and Metaphysics Z, H, Θ. As we remarked in Section 6.2, Aristotle uses the concept of substance both in an absolute and in a comparative way in the Categories. In Metaphysics Z, H, Θ, the earlier ultimate subject criterion is still an important consideration, but no longer the only factor in evaluating whether an entity deserves to be assigned substance status. In these texts, we come across absolute, relational, and comparative uses of the concept of substance as well as criteria of substancehood which go along with these.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

170  Independence Criteria of Substancehood (ED5)  Constitutive Definitional Dependence: x is constitutively definitionally dependent on y ≡def y is a constituent of a real ­definition of x. When we substitute any of these notions of ontological dependence into (IND) and (NICSS), it turns out that, given certain assumptions, no matter–form compound counts as a substance simpliciter. This result is surely startling, especially in light of the three tenets cited earlier which are widely accepted among Aristotelians: for, according to (iii), at least some matter–form compounds are to be included in the category of substances; and, assuming that the notion of substancehood at play here is that of substancehood simpliciter, the negative result just stated contradicts (iii). As we will observe below, the denial of (iii) follows from a more general result which can be established when (NICSS) and (IND) are interpreted in terms of any of the construals of ontological dependence just listed. According to the more general result in question, of which the denial of (iii) is a special case given certain assumptions, any entity which is composite and has essential constituents fails to be ontologically independent, and is therefore disqualified from counting as a substance simpliciter, assuming that ontological independence is read in terms of the definitions stated here. Suppose that A is a composite entity and B is an essential constituent of A. In that case, necessarily A exists only if B exists; therefore, by (ND1), A is rigidly existentially necessarily dependent on B. Suppose further that B is an F; then, generalizing the result just stated, necessarily A exists only if some F exists. Hence, by (ND2), A is also generically existentially necessarily dependent on Fs. Next, given our hypothesis, we may assume further that there is some function, φ, such that it is part of the essence of A that it is the result of applying φ to B;10 in that case, by (ED3), A is essentially identity dependent on B. Finally, given our hypothesis and given the conception of essence used in formulating (ED4), B is a constituent of A’s essence (narrowly construed), i.e., B is a constituent of the proposition or collection of propositions that is A’s essence. Therefore, by (ED4), A is essentially constitutively dependent on B. By the same reasoning, A is also constitutively definitionally dependent on B, in accordance with (ED5), the definitional variant of (ED4). Thus, if (IND) is read with any of these definitions of ontological dependence in mind, it turns out, given our additional assumptions that, for any A, such that A 10   A brief review of how to understand E. J. Lowe’s appeal to functions in the formulation of (ED3) will follow in Section 6.4. (For a more detailed discussion of (ED3), see Section 5.3.1.) In order for A to be essentially identity dependent on B, we suppose that it is part of A’s essence that A is the result of applying a composition operation of some sort to B. (In order to maintain maximum generality, I leave open whether the composition operation in question satisfies the so-called “Weak Supplementation Principle,” i.e., whether A must have some further constituent, C, disjoint from B.) In order for the essential relationship between A and B to be functional, we must suppose that, in any given circumstance, the application of the composition operation in question to B gives rise to a unique compound, viz., A. Thus, in the case of (ED3), the negative result derived in the main text only follows for composition operations which satisfy a principle analogous to the Axiom of Extensionality in set theory or the Uniqueness of Composition in mereology.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

alternative strategies  171 is a composite entity which has essential constituents, A fails to be ontologically ­independent; hence, by (NICSS), A is not a substance simpliciter. As a special case, assuming (as I do) that matter–form compounds are composite entities which have at least their forms as essential constituents, then no matter–form compound satisfies the necessary condition for substancehood simpliciter stated in (NICSS). Hence, no matter–form compound is a substance simpliciter, given the definitions and assumptions currently operative.

6.4  Alternative Strategies It is possible to avoid the negative result just derived by denying some combination of the assumptions to which I appealed in generating it. In particular, Aristotelians who wish to uphold their commitment to the inclusion of at least some hylomorphic compounds among the substances simpliciter may wish to take issue with one or more of the following claims: (a) that matter–form compounds and their forms are numerically distinct;11 (b) that forms are essential constituents of matter–form compounds;12 or (c) that an entity’s ontological dependence on its own essential constituents disqualifies it from counting as a substance simpliciter.13 In what follows, we will engage with various proposed independence criteria which rely on rejecting one or more of these claims. In my view, none of these attempts to uphold (iii) are satisfactory in the end. In light of these considerations, as I will argue in the remainder of this chapter, we should concede that matter–form compounds, due to their composite nature, fail to be ontologically independent in any of the senses defined earlier and therefore do not qualify for the status of substances simpliciter, when this notion is approached through the lens of ontological independence. Nevertheless, I will propose that matter–form compounds should count as ontologically privileged in other ways, despite their composite nature, since they exhibit a relatively high degree of unity compared to that exhibited by other composite entities (e.g., heaps). The resulting notion of substancehood, however, gives rise to a comparative, rather than an absolute, fundamentality ranking among composite entities.

11   As discussed earlier (see Section 3.2.3), this assumption is rejected by identity theorists (e.g., Frede (1985, 1987a); Frede and Patzig (1988); Lowe (1999); and Whiting (1984, 1986, 1991, 1992)), who hold that a matter–form “compound” is (in some sense) numerically identical to its form. 12   Of course, the second claim could be broken down further by separating the claim that forms are constituents of matter–form compounds from the claim that they are essential constituents of matter–form compounds. However, since I assume that anyone who accepts the former also accepts the latter, I focus only on the combined claim. 13   Again, as indicated in the previous note, the third claim could also be broken down more finely into a claim concerning an entity’s ontological dependence on its constituents more generally and a claim concerning an entity’s ontological dependence on its essential constituents more specifically. For reasons which will become apparent shortly, however, this further differentiation turns out to be unnecessary, since only an entity’s ontological dependence on its essential constituents will be judged to be relevant to the question of whether it should, for that reason, be disqualified from counting as a substance simpliciter.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

172  Independence Criteria of Substancehood

6.4.1  Lowe’s Account E. J. Lowe has proposed the following criterion of substancehood simpliciter: 14 (ICS1)  Independence Criterion for Substancehood Simpliciter (Lowe): x is a substance simpliciter ≡def (i) x is a particular; and (ii) there is no particular y such that (a) y is not identical with x and (b) x is essentially identity dependent on y.15 Clause (ii.b) of (ICS1) makes use of Lowe’s notion of essential identity dependence, (ED3), which was discussed in more detail in Section 5.3.1. (ED3) classifies an entity, x, as essentially identity dependent (or, for short, “(ED3)-dependent”) on an entity, y, when x’s synchronic identity at each time at which x exists is fixed by a certain (functional) relationship x essentially bears to y at that time. In contrast, an entity, x, counts as essentially identity independent of other entities, according to (ED3) (or, for short, “(ED3)-independent”) when x’s synchronic identity at each time at which x exists is not fixed by a (functional) relationship x bears essentially to any other entities numerically distinct from itself. Thus, no synchronic criterion of identity or principle of individuation which appeals to numerically distinct entities will be available at all for (ED3)-independent entities; rather, that they are the very entities they are at each time at which they exist is simply taken as a basic, non-derivative fact about these entities. Thus, if a concrete particular object, such as Socrates, is to qualify as an (ED3)-independent entity, and hence, by (ICS1), as a substance simpliciter, then it must be the case that Socrates does not owe his individuation or synchronic identity, i.e., his being the very entity that he is at each time at which he exists, to a relation he bears essentially to any other entity numerically distinct from himself. One might think, as noted in Section 5.3.1, that Lowe’s conception of the ontological independence of the substances simpliciter is incompatible with the essentiality of origins. For if it were part of Socrates’ essence, for example, to have originated from a particular zygote, then it might appear to be possible to formulate a criterion of individuation or synchronic identity after all for an allegedly (ED3)-independent entity such as Socrates, viz., one which appeals to Socrates’ origins. But whether the essentiality of origins in fact yields a criterion of individuation or synchronic identity which tracks concrete particular objects across possible worlds in which they exist is 14   In what follows, I will refer to Lowe’s and Gorman’s criterion as an “independence criterion of substancehood,” even though ontological independence is not the only component of their respective accounts. In proposing their respective independence criteria, it is clear that Lowe and Gorman have in mind substancehood simpliciter, rather than a comparative or relational use of the concept of substance. For an alternative proposal, see, for example, Hoffman and Rosenkrantz (1994, 1997). The remainder of this chapter is extracted from Koslicki (2013b), with light revisions. 15   (ICS1) is a slightly reformulated version of what is called “(SUB-2)” in Tahko and Lowe (2015). It should be noted, too, that, except for the addition of clause (i), (ICS1) embodies the conception of substancehood simpliciter which was seen to get us into trouble with the empty set and the number 0 in Section 5.3.4. In Section 6.4.3, we will encounter reasons to question the legitimacy of including (i) in a criterion of substancehood simpliciter.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

alternative strategies  173 itself doubtful (see Section 3.4.3); moreover, Lowe is in any case happy to give up on the essentiality of origins, since he regards this thesis as implausible. As the general negative result derived in Section 6.3 brings out, however, a more troubling conflict apparently emerges between Lowe’s conception of the ontological independence of the substances simpliciter and a certain natural interpretation of the hylomorphic analysis of concrete particular objects as compounds of matter and form. For if it were part of the essence of an allegedly (ED3)-independent concrete particular object, such as Socrates, to be a compound of some matter and some form, then it might appear again that Socrates’ synchronic identity, at each time at which he exists, could be fixed at least by appeal to his form, even if his matter is permitted to vary from one time to another. In this case, Lowe adopts a different escape route: he argues, in Lowe (1999)), that matter–form “compounds” should be identified with their particularized forms and therefore are not, strictly speaking, compounds of matter and form at all. We will return to this strategy in Section 6.4.5.1.

6.4.2  Gorman’s Modifications of Lowe’s Account Michael Gorman has argued that Lowe’s independence criterion for substances should be modified in the following way:16 (ICS2)  Independence Criterion for Substances Simpliciter (Gorman): x is a substance simpliciter ≡def (i) x is a particular; (ii) there is no particular y such that (a) y is not identical with x, (b) x is essentially identity dependent on y, and (c) y is not one of x’s proper parts; and (iii) x is unified in the right way. (ICS2) is just like (ICS1) with the exception that Gorman adds two clauses to Lowe’s criterion, viz., (ii.c) and (iii). The first of these, (ii.c), allows an entity to qualify as a substance simpliciter even if it is essentially identity dependent on entities numerically distinct from itself, as long as these entities are among its own proper parts. The second added clause, (iii), requires substances simpliciter to be unified “in the right way.” These additional clauses are intended to exclude the following types of cases which Gorman considers to be counterexamples to Lowe’s independence criterion, as stated in (ICS1). Firstly, to motivate the unity-clause in (iii), Gorman asks us to consider the Berlin Philharmonic Orchestra. Assuming that orchestras are particulars, it might seem that (ICS1), as it stands, classifies such entities as substances simpliciter, which Gorman takes to be an unwelcome result. Since its inception in 1882, the Berlin Philharmonic Orchestra, for example, has managed to survive all sorts of changes, e.g., with respect to its conductor or the musicians who are its members at each time at which the orchestra exists. In this way, orchestras are more similar to such entities as organisms, which are allegedly to be classified as substances simpliciter and which can   Gorman (2006a), p. 116. (ICS2) is a slightly reformulated version of what Gorman calls “RLS*.”

16

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

174  Independence Criteria of Substancehood also persist through changes with respect to their parts, than they are to such entities as sets and mereological sums, which are allegedly not to be included among the substances simpliciter and which are not capable of surviving changes with respect to their members or parts. Since the synchronic identity of the Berlin Philharmonic Orchestra at each time at which it exists is apparently not fixed by its essential relations to any other particulars numerically distinct from itself, such as its conductor or the musicians who are members of it, (ICS1) therefore seems to have the consequence that the Berlin Philharmonic Orchestra is classified as a substance simpliciter. To avoid this result, Gorman introduces the unity requirement in (iii), since he believes that what accounts for the difference between such entities as organism, which are properly classified as substances simpliciter, and such entities as orchestras, sets, and mereological sums, which ought not to be classified as substances simpliciter, is that entities which belong to the former categories are unified in a way in which entities which belong to the latter categories are not. Since he does not spell out further in what way the substances simpliciter are unified compared to their complement class, however, the exact content of the unity-requirement in (iii) at this point still remains to be determined. Secondly, Gorman’s exclusion of proper parts in (ii.c) rests on the idea that even entities which he regards as plausible candidates for inclusion among the substances simpliciter can have essential proper parts.17 To illustrate, Gorman takes it to be part of the essence of H2O molecules (which, in his view, are likely contenders for the status of substances simpliciter) to be composed of the very hydrogen and oxygen atoms of which they are in fact composed. In that case, it appears that H2O molecules would be classified by (ED3) as essentially identity dependent on particulars numerically distinct from themselves, since there would be some function, φ, e.g., the “molecule composition” function, such that it is part of the essence of an H2O molecule that it is the result of applying φ to the oxygen and hydrogen atoms that are its essential proper parts.18 The molecule’s synchronic identity at each time at which it exists, in that case, would be fixed by appeal to these atoms which are its essential proper parts, much like a set’s identity is fixed by appeal to its members. Unless the essential identity dependence 17   Only essential, rather than accidental, parts are relevant to the question of whether an entity should be disqualified from its status as a substance simpliciter due to the fact that it is ontologically dependent on entities numerically distinct from itself (namely, in this case, its own proper parts). For clause (i.b) in (ICS2) narrows down the range of entities which might pose a threat to x’s status as a substance simpliciter to those entities, y, on which x is essentially identity dependent. But no entity, x, would count as essentially identity dependent on its accidental parts, since it would not be the case that there is a function, φ (e.g., the “is mereologically composed of ” function) such that it is part of x’s essence that x is the result of applying φ to any of its accidental parts. Since x can survive through changes with respect to its accidental parts, x’s identity at any time or world at which it exists cannot be fixed by which entities, y, z, w, . . . , are its accidental parts at that time or world. In what follows, when I consider the question of whether the exclusion of proper parts from the criterion of substancehood simpliciter is admissible, I will therefore limit myself to the discussion of essential, rather than accidental, parts. 18   As mentioned in Section 6.3, in order for the relation between an H2O molecule and the oxygen and hydrogen atoms which allegedly essentially compose it to be functional, the composition operation in question must be assumed to yield a unique compound when applied to these constituents in any given circumstance.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

alternative strategies  175 of an entity on its own proper parts is explicitly excluded as irrelevant to its status as a substance simpliciter, as is done by clause (ii.c) of (ICS2), an H2O molecule would therefore be stripped of its alleged title as a substance simpliciter by (ICS1). Such entities as mereological sums, which are apparently also essentially identity dependent only on their own proper parts, in Gorman’s view, are to be ruled out as potential substances simpliciter by way of the unity requirement in (iii).

6.4.3  The Stipulative Exclusion of Non-Particulars Both Lowe’s original criterion in (ICS1) and Gorman’s modified criterion in (ICS2) contain clauses which explicitly exclude non-particulars from the range of entities which might qualify as substances simpliciter. But the stipulative exclusion of non-particulars from an independence criterion of substancehood simpliciter is problematic, because it renders apparently substantive ontological disputes over questions of fundamentality non-substantive.19 Consider, for example, two philosophers who agree that both universals and particulars exist, but disagree over which taxonomic category of entities deserves to be granted the status of substances simpliciter: one philosopher, let’s say, regards universals as occupying the ontologically fundamental role of substances simpliciter, while the other takes the substances simpliciter to be particulars. Given (ICS1) and (ICS2), the first philosopher’s thesis: “The substances simpliciter are universals,” is classified as contradictory (assuming that nothing is both particular and universal), since the criteria require that by definition something is a substance simpliciter only if it is a particular. The second philosopher’s thesis: “The substances simpliciter are particulars,” in contrast, is classified by (ICS1) and (ICS2) as trivial, since it simply follows from clause (i) of the definition that the substances simpliciter are particulars. If we now attempt to remedy this situation by interpreting the two philosophers engaged in this dispute as subscribing to distinct criteria of substancehood simpliciter, then we reach the equally unfortunate result that these two philosophers, instead of being engaged in what appears to be a substantive disagreement over questions of ontological fundamentality, are now simply talking past each other, with each of them subscribing to a different criterion of substancehood simpliciter. Given these considerations, I take it that clause (i) should be regarded as an unattractive addition to an independence criterion of substancehood simpliciter: it should turn out to be a philosophically interesting and meaningful question which taxonomic category or categories of entities (if any) satisfy a given criterion of substancehood simpliciter and whether these entities are particulars or universals. We thereby

19   The restriction to particulars is also present in an independence criterion of substancehood simpliciter proposed in Schnieder (2006), according to which x is a substance simpliciter just in case x is a particular and there is no y such that x rigidly and permanently existentially depends on y. As our discussion of rigid and permanent existential dependence in Section 5.2.3.2 brought out, this construal of ontological dependence is subject to additional, and unrelated, difficulties as well.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

176  Independence Criteria of Substancehood arrive at the following first revision of Gorman’s independence criterion in (ICS2), with the restriction to particulars deleted: (ICS3)  Independence Criterion for Substances Simpliciter (First Revision): x is a substance simpliciter ≡def (i) there is no y such that (a) y is not identical with x, (b) x is essentially identity dependent on y, and (c) y is not one of x’s proper parts; and (ii) x is unified in the right way.20

6.4.4  The Stipulative Exclusion of Proper Parts As we have seen, one of the two ways in which Gorman’s modified criterion in (ICS2) differs from Lowe’s original criterion in (ICS1) is in its addition of clause (ii.c), which eliminates part-dependence as a possible threat to an entity’s status as a substance simpliciter. I now want to consider the question of whether such a stipulative exclusion clause governing proper parts should be regarded as an admissible element in an independence criterion of substancehood simpliciter.21 In principle, considerations analogous to those I adduced in connection with the stipulative exclusion of non-particulars appear to be relevant in this context as well: for it ought not simply to be settled by fiat whether entities which are ontologically dependent only on their own proper parts can be classified as occupying the ontologically fundamental role of substances simpliciter. But instead of pursuing this line of argument, I will bring other issues to bear on the question of whether a clause excluding part-dependence should be considered an admissible component of an independence criterion of substancehood simpliciter. 20   Both Lowe’s criterion in (ICS1) and Gorman’s criterion in (ICS2) contain an additional restriction to  particulars in clause (ii), which I have also removed in this revision of Gorman’s criterion in (ICS3). According to this additional restriction, given that both Lowe and Gorman take it to be a settled question that only particulars may qualify as substances simpliciter, the only entities, y, which could pose a threat to an entity, x’s, status as a substance simpliciter are other particulars numerically distinct from x on which x is essentially identity dependent. But again we may wonder whether the exclusive focus on particulars is legitimate. If an entity, x, is essentially identity dependent on an entity, y, that is numerically distinct from it, then should x’s status as a substance simpliciter not thereby be jeopardized, even if y is a non-particular (e.g., a universal)? As far as I can see, this kind of possibility is already ruled out by other considerations and the second restriction to particulars in clause (ii) may therefore be safely removed. For suppose that it is part of Socrates’ essence to be human and that being human is here construed as a universal. Still, the universal, being human, could not have a role in fixing Socrates’ identity at every time and world at which he exists, since it is also part of the essence of many other particulars (Plato, Aristotle, etc.) that they exemplify the same universal. Thus, in order to determine which exemplar of the universal, being human, Socrates is at every time and world at which he exists presupposes that Socrates’ identity is already settled. 21   Gorman is again not alone in opting for the stipulative exclusion of proper parts from an independence criterion of substancehood simpliciter. Kit Fine, for example, states that “ . . . a substance may be taken to be anything that does not depend upon anything else or, at least, upon anything other than its parts” (Fine (1995a), pp. 269–70; my emphasis). Similarly, Peter Simons offers an independence criterion of substancehood simpliciter in terms of what he calls “strong independence,” which also explicitly excludes proper parts from its range: an entity x is strongly dependent on an entity y just in case necessarily, x exists only if y exists and y is neither identical to x nor a proper part of x (Simons (1998), p. 236; my emphasis). Simons’ notion of strong independence is formulated in terms of what I called above “rigid existential necessary dependence” or (ND1) and adds further exemption clauses to it.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

alternative strategies  177 6.4.4.1  The Possibility of Simple Substances Simpliciter In a recent discussion of this issue, Patrick Toner has objected to the stipulative exclusion of proper parts from an independence criterion of substancehood simpliciter on the following grounds (Toner (2010)). In Toner’s view, we should at least in principle allow for the possibility of substances simpliciter which are simple, i.e., entities which have no proper parts at all and which therefore, a fortiori, cannot ontologically depend on their essential proper parts. Possible examples of such simple substances simpliciter might include God, if God exists; minds, souls, or persons, according to certain conceptions of these entities; or physical simples, i.e., incomposite concrete particular objects which may be included in the inventory of fundamental physics. Toner asks: Why accept that simple substances, which are self-sufficient in one way (a way that doesn’t except dependence on their parts) are the same kind of things as “substances” that are self-­ sufficient in a very different kind of way (a way that does except dependence on their parts)? (Toner (2010), p. 40)

If there are any simple entities which are not ontologically dependent on anything numerically distinct from themselves, then (ICS2), in Toner’s mind, turns the category of substances simpliciter into a heterogeneous collection. For one thing, this category would then comprise these simple entities which are completely ontologically independent from everything else. These entities are admitted into the category of substances simpliciter by (ICS1)–(ICS3) without requiring any special exemption. But, in addition to these simple entities, the category of substances simpliciter, according to (ICS2) and (ICS3), would also include composite entities which may be ontologically dependent on their own essential proper parts, as long as these entities are not ontologically dependent on anything numerically distinct from them besides their own essential proper parts. These composite entities are admitted into the category of substances simpliciter by (ICS2) and (ICS3) only by way of the special exception clause governing proper parts. But why believe, Toner asks, that we have thereby arrived at a unified category? This stipulative exclusion strategy, so Toner argues, is analogous to allowing into the class of all flying things not only things that have the ability to propel themselves through the air by their own power, but also things that can be carried along by something else. This way of allegedly delineating the class of flying things does not yield a unified category; nor, in Toner’s view, do we arrive at a unified category by allowing entities which ontologically depend on their essential proper parts to count as substances simpliciter, along with entities which have no proper parts and hence cannot ontologically depend on their essential proper parts. 6.4.4.2  The Threat of Heterogeneity As Toner’s observations bring out, as far as the simple substances simpliciter are concerned (if there are such entities), Lowe’s original criterion in (ICS1), stated here with

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

178  Independence Criteria of Substancehood the restriction to particulars deleted and a restriction to simple entities added, would do just as well as the modified criteria in (ICS2) or (ICS3): (ICSS)  Independence Criterion for Simple Substances Simpliciter: x is a simple substance simpliciter ≡def (i) x is simple and (ii) there is no y such that (a) y is not identical with x and (b) x is essentially identity dependent on y. According to (ICSS), the simple substances simpliciter are those entities which are simple and completely ontologically independent from everything else numerically distinct from them. Since these entities are simple, the question of whether they are appropriately unified presumably does not arise for them; nor is there any danger that such entities might depend ontologically on their essential proper parts, given their simplicity. There is therefore no need for the addition of Gorman’s unity requirement in an independence criterion for simple substances simpliciter, just as there is no need for a clause exempting part-dependence.22 The question now arises as to whether there are any composite substances simpliciter and, if so, whether a revised version of Gorman’s independence criterion might be appropriate for composite substances simpliciter: (ICCS1)  Independence Criterion for Composite Substances Simpliciter: x is a composite substance simpliciter ≡def (i) x is composite and (ii) there is no y such that (a) y is not identical with x, (b) x is essentially identity dependent on y, and (c) y is not one of x’s proper parts; and (iii) x is unified in the right way. According to (ICCS1), composite entities may qualify as substances simpliciter, as long as they are not ontologically dependent on anything besides their own essential proper parts and as long as they are appropriately unified. Even if (ICCS1) carries promise as an independence criterion for composite ­substances simpliciter, Toner would no doubt object to the resulting bifurcation of the notion of substance simpliciter into the simple ones, on the one hand, and the composite ones, on the other, with each kind being governed by its own independence criterion. Toner’s challenge to a proponent of an independence criterion for substances simpliciter who endorses a bifurcated account in the style of (ICSS) and (ICCS1) is to indicate wherein the alleged unity of the category of substances simpliciter lies. What, so he might ask, gives us the right to think of both (ICSS) and (ICCS1) as criteria allegedly delineating a single ontological category, rather than two separate categories, viz., the simple entities which are completely ontologically independent, on the one hand, and the appropriately unified composite entities which are ontologically independent only in a modified way, on the other hand? In response to Toner’s worry concerning the apparent heterogeneity or disjunctiveness in the notion of substance simpliciter to which the bifurcated account indicated 22   Notice, however, that (ICSS) still succumbs to the difficulty concerning the empty set and the number 0 mentioned in Section 5.3.4.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

alternative strategies  179 here seems to lead, a proponent of such an account may at least point to the fact that there is after all a non-ad hoc and metaphysically significant distinction between simple entities and composite entities. Assuming that a criterion of substancehood serves as an indicator of a certain kind of ontological fundamentality, it is perhaps no surprise that there might be distinct roads toward ontological fundamentality, among them: one for simple entities and another for composite entities. For the time being, then, although I do feel the force of Toner’s worry, I want to set it aside and examine instead the question whether (ICCS1) might be appropriate as an independence criterion for composite substances simpliciter. In what follows, I want to focus on a different challenge which arises for (ICCS1) as an independence criterion for composite substances simpliciter: this challenge centers on the selective emphasis on proper parts, as opposed to constituents more generally.23 6.4.4.3  Proper Parts vs. Constituents Something can be a constituent of a composite entity without being a proper part of it. For example, given widely held assumptions about sets (viz., in particular, those connected with the Iterative Conception of Sets), it is plausible to take the members of sets to be constituents of the sets constructed from them. In contrast, given widely held assumptions about parthood (viz., in particular, that parthood is transitive), the members of sets are not proper parts of the sets of which they are members.24 In what follows, I use the term, “constituent,” in such a way that proper parts are to be included among an entity’s constituents; but the reverse cannot always be assumed to be true, since not all constituents are also proper parts, as the set-theoretic example just cited illustrates. Why should the stipulative exclusion of proper parts from an independence criterion of substancehood simpliciter not also extend to constituents more generally? Insofar as any justification for this exclusion is given by those who endorse the stipulative exclusion of proper parts, the reasons stated would seem to carry over to non-mereological constituents as well. Gorman, for example, adduces the following considerations in favor of the exemption in question: To say this [i.e., that composite entities may qualify as substances simpliciter even if they are ontologically dependent on their own essential proper parts] is, of course, only to follow up on Fine’s own suggestion when he says that a substance does not depend on anything “or, at least, upon anything other than its parts”. Nor is there any reason to fear that the move is ad hoc, as 23   Gorman replies to Toner’s heterogeneity worry in Gorman (2012) and argues that his account does in fact provide a unified criterion of substancehood simpliciter: the substances simpliciter are all and only those entities which are ontologically independent from all numerically distinct entities “outside” of themselves; this, according to Gorman, applies to both simple and composite entities. When we focus on what goes on “inside” the boundaries of an alleged substance simpliciter, however, we can still discern a difference between simple and composite entities. The “inside”/“outside” distinction which is invoked by Gorman and others in an attempt to give at least an informal justification of the stipulative exclusion of proper parts from an independence criterion of substancehood simpliciter itself raises interesting questions to which we will turn shortly. 24   Though see Fine (2010) for a more generalized notion of parthood, closer to what I am calling here “constituency,” which allows for members to be parts of sets.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

180  Independence Criteria of Substancehood it is a development of the pre-philosophical intuition that the theory of substance is intended to make sense of. Putting the point a bit vaguely, as pre-philosophical intuitions must be put, the things that philosophers come to call substances are not dependent on others but are instead self-sufficient in some way. Now a thing with an essential part is (of course) distinct from that part, but it does not follow that the thing is not self-sufficient, because this is not a way for the thing to be related to something outside itself. Expressed differently, the kind of independence here sought is not compromised by dependence that, so to speak, stays within the thing in question.  (Gorman (2006b), p. 151)25

In a similar vein, Peter Simons remarks as follows: An object A is strongly dependent on an object B if necessarily, if A exists, so does B, and B is neither A nor part of A. [. . .] An object is independent in the corresponding sense when it depends on nothing apart from itself and perhaps parts of itself, giving a sense to the idea of something depending on nothing “outside of itself ”.  (Simons (1998), p. 236)

Both Simons and Gorman point to the idea that an entity’s ontological dependence on its own essential proper parts should be treated differently from an entity’s ontological dependence on entities numerically distinct from itself which do not number among the entity’s own essential proper parts for the following reason. In the second case (non-part dependence), the entity in question is ontologically dependent on numerically distinct entities that lie “outside” of it, while in the first case (part-dependence), the entity in question is ontologically dependent on numerically distinct entities (viz., its own essential proper parts) which do not lie “outside” of it. But whatever exactly is meant by “outside” in this context, surely if an entity’s proper parts do not lie “outside” of it, then neither do an entity’s non-mereological constituents. Supposing then that entities which are ontologically dependent only on their own essential constituents more generally may also qualify as substances simpliciter, as long as they are appropriately unified, we arrive at the following revision of (ICCS1): (ICCS2)  Independence Criterion for Composite Substances Simpliciter (First Revision): x is a composite substance simpliciter ≡def (i) x is composite and (ii) there is no y such that (a) y is not identical with x, (b) x is essentially identity dependent on y, and (c) y is not one of x’s constituents; and (iii) x is unified in the right way. Clause (ii.c) of (ICCS2) now allows back in some of the entities, e.g., sets, which were previously supposed to be excluded from the reaches of (ICCS1) by way of the restriction to proper parts. In addition to sets, we might also cite the following cases as possible further examples of entities which are arguably ontologically dependent only on their own essential constituents: quantities, collections, propositions, sentences, events, facts, states of affairs, and the like. All of these entities, if they are to be excluded from the category of substances simpliciter, would now have to be ruled out by way   See also Gorman (2006a), p. 116, for a similar comment.

25

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

alternative strategies  181 of the unity requirement in clause (iii). This not only creates serious pressure for the as-of-yet unspecified unity requirement; it also makes us wonder whether clause (ii), i.e., the ontological independence requirement, is really doing any work at all in the so-called independence criterion for composite substances simpliciter. Given the long list of entities which are allegedly to be excluded from the category of substances simpliciter, even though they are arguably ontologically dependent only on their own essential constituents, it seems that a unity criterion for composite substances simpliciter holds more promise than an independence criterion, assuming that we can make good on the promise of spelling out in more detail in what respects the putative substances simpliciter are unified “in the right sort of way,” in contrast to other entities, such as sets, orchestras, committees, quantities, collections, mereological sums, propositions, sentences, events, facts, states of affairs, and the like. This line of reasoning seems to suggest the following significant change of direction in our attempt to provide a criterion of substancehood simpliciter for composite entities: (UCCS)  Unity Criterion for Composite Substances Simpliciter: x is a composite substance simpliciter ≡def (i) x is composite and (ii) x is unified in the right way.26, 27 I will return to the role of unity in a criterion for composite substances later in this chapter and in Chapter 7. For the time being, I want to turn instead to a loose thread which arose in connection with the revised independence criterion for composite substances simpliciter in (ICCS2).28 26   Given that ontological independence has now completely dropped out of the picture in (UCCS), Toner’s earlier heterogeneity worry arises again with a vengeance: if simple entities (supposing that there are any) qualify as substances simpliciter for one reason (their ontological independence) and composite entities qualify as substances simpliciter for a completely different reason (their unity), why should we believe that (ICSS) and (UCCS) point to a single ontological category, rather than two distinct ontological categories which have been misleadingly referred to with the same name? As noted earlier, the proponent of such a bifurcated account can at least draw on the non-ad hoc and metaphysically significant distinction between simplicity and complexity. Moreover, we may point out as well that simple entities are presumably also unified in the right way, due to their simplicity; in that sense, a unity account applies both to simple and composite substances simpliciter, but perhaps only trivially so in the case of simple substances simpliciter. 27   An interesting case to consider in connection with (ICSS) and (UCCS) is that of tropes. Lowe and Gorman take tropes to be entities which are both simple and essentially identity dependent on their bearers, viz., the concrete particular objects in which they inhere. They would therefore be excluded (correctly, in Lowe’s and Gorman’s view) from the category of substances simpliciter by (ICSS). Being simple, tropes are presumably also unified by default; but their unity is irrelevant to their alleged exclusion from the category of substances simpliciter, since they are subsumed under (ICSS), the criterion governing simple entities, rather than under (UCCS), the criterion governing complex entities. 28   Since Gorman is aiming for a criterion for substancehood simpliciter, clause (iii) of his revised independence criterion in (ICS2) puts to use the notion of unity in a non-relative way, by requiring the substances simpliciter to be “unified in the right way.” In contrast, I will employ unity as a comparative notion, according to which one entity or type of entity can be assigned a higher degree of unity than another. There is thus also the option of appealing to unity as a sufficient condition for a comparative notion of substancehood for composite entity as follows: x is more deserving of substance status than y if (i) x and y are composite and (ii) x has a higher degree of unity than y. Moving to a comparative notion of substancehood, moreover, helps with Toner’s heterogeneity worry as well, since this objection specifically targets approaches which classify a heterogeneous collection of entities as substances simpliciter or absolutely fundamental.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

182  Independence Criteria of Substancehood 6.4.4.4 Intrinsicness In the passages cited earlier, we saw that Gorman and Simons make at least an informal attempt at justifying the exemption for part-dependence by appeal to a distinction they allude to between what lies “inside” and what lies “outside” the boundaries of a given entity. According to Gorman, an entity’s ontological dependence on its own essential proper parts does not take away from the sort of “self-sufficiency” he regards as being required for an entity’s status as a substance simpliciter, since in that case the entity in question ontologically depends only on what lies “inside” its boundaries. But non-part dependence, in his view, does disqualify an entity from being awarded the status of a substance simpliciter, since an entity’s ontological dependence on what lies “outside” of its boundaries diminishes the “self-sufficiency” he takes to be required for an entity’s inclusion into the category of substances simpliciter. On the basis of these considerations, we might therefore propose the following revision of (ICCS2), which makes the justification offered for the exclusion of proper parts explicit in clause (ii.c): (ICCS3)  Independence Criterion for Composite Substances Simpliciter (Second Revision): x is a composite substance simpliciter ≡def (i) x is composite and (ii) there is no y such that (a) y is not identical with x, (b) x is essentially identity dependent on y, and (c) y lies “outside” of x’s boundaries; and (iii) x is unified in the right way. Since, for the proponent of an independence criterion for substancehood simpliciter, “self-sufficiency” is presumably merely another name for whatever is captured by the criterion of substancehood simpliciter in question, the notion of “self-sufficiency” that is appealed to informally by Gorman does not provide us with any additional information besides what is brought to the table by all the components of the independence criterion taken together. The question that is most relevant for present purposes then is what sense can be attached to the distinction between what lies “inside” the boundaries of a given entity and what lies “outside” of its boundaries. The first thing to note in this connection is that we ought to separate ourselves right away from the spatial overtones that the distinction between what lies “inside” and what lies “outside” a given entity tends to evoke. (Hence the quotation marks around “inside” and “outside.”) In Section 6.4.3, I argued that the stipulative exclusion of nonparticulars from an independence criterion of substancehood simpliciter ought to be regarded as inadmissible, since it has the unwelcome consequence that apparently substantive disputes in ontology are classified as either trivially answerable or as based on a contradiction. For similar reasons, an independence criterion of substancehood simpliciter should not be formulated in terms that can meaningfully apply only to concrete particular entities, i.e., entities which occupy regions of space-time. For then the thesis, “Only concrete particular entities are composite substances simpliciter,” would again trivially follow from the criterion of substancehood simpliciter in question, while the opposing thesis, “Some entities which are not concrete particulars

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

alternative strategies  183 are composite substances simpliciter,” could not be coherently maintained by a reasonable philosopher who subscribes to (ICCS3). Hence, a dispute between two philosophers who find themselves drawn to these two opposing theses respectively would be mistakenly classified as non-substantive. But I suspect that Gorman himself would wish to allow that the question of whether some composite substances simpliciter are not concrete particulars is substantive and can be a legitimate subject of dispute between two reasonably minded metaphysicians. If sets and other abstract entities, for example, are to be excluded from the category of substances simpliciter by (ICCS3), then it would seem that this criterion must at least be formulated in such a way that we can sensibly ask whether the entities in question satisfy or fail to satisfy the requirements stated by each of its clauses. It would be disturbing if sets and other abstract entities were denied the status of substances simpliciter only because the “outside”/“inside” distinction does not meaningfully apply to them. A natural approach to the “outside”/“inside” distinction appealed to in (ICCS3) which does not require the stipulative restriction to concrete particular entities just cited is to understand it in terms of the distinction between what is intrinsic and what is extrinsic to a given entity.29 Presumably, in whatever way exactly we construe the distinction between what is intrinsic and what is extrinsic to a given entity, any plausible account of this distinction should allow that we can just as sensibly speak of the nonmereological constituents of a non-empty set (i.e., its members) as being intrinsic to the set as we can speak of the mereological constituents of a concrete particular entity (i.e., its parts) as being intrinsic to the whole they compose. We thus arrive at the following reformulation of clause (ii.c) of (ICCS3): (ICCS4)  Independence Criterion for Composite Substances Simpliciter (Third Revision): x is a composite substance simpliciter ≡def (i) x is composite and (ii) there is no y such that (a) y is not identical with x, (b) x is essentially identity dependent on y, and (c) y is extrinsic to x; and (iii) x is unified in the right way. If (ICCS4) strikes the proponent of an independence criterion of substancehood simpliciter as attractive, he would now face the non-trivial task of having to link his criterion for composite substances simpliciter to a suitable account of the intrinsic/ extrinsic distinction. This requirement, for example, immediately rules out any appeals to the notion of substancehood simpliciter in an account of the intrinsic/extrinsic distinction, since such an appeal would then render the overall theory in question circular. Even if we grant the proponent of an independence criterion for substances simpliciter that the content of clause (ii.c) can be specified in a suitable fashion, however, we should note that (ICCS4) has some interesting consequences which may or may not 29   The question of how best to draw the intrinsic/extrinsic distinction has generated an enormous literature and I will not at present try to enter into this debate; but see, for example, Humberstone (1996); Lewis and Langton (1996); Sider (1996); and Yablo (1999) for discussion.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

184  Independence Criteria of Substancehood be found to be objectionable by those in favor of an independence criterion for composite substances simpliciter. For reasons of space, I will here only point to one such consequence to which (ICCS4) may lead, depending on the additional metaphysical assumptions with which (ICCS4) is combined. Consider artworks and artifacts more generally. Suppose Michelangelo’s David, for example, is essentially identity dependent on the artist, Michelangelo, who created the artwork in question with the intention of achieving a certain artistic representational goal. Given Lowe’s notion of essential identity dependence, in order for the sculpture in question to be essentially identity dependent on the artist who created it with a certain artistic intention in mind, there must be some function, φ, e.g., the “is the sculpture which was artistically created with a certain representational intention” function, such that it is part of the essence of the sculpture in question that it is the result of applying φ to Michelangelo. In other words, if the condition just stated in fact holds, then which sculpture the artwork in question, is at least in part fixed by reference to the artist who created it with the intention of achieving a certain representational goal. But the artist, Michelangelo, is presumably under any reasonable conception of the intrinsic/extrinsic distinction, extrinsic to the sculpture he has created. Thus, regardless of whether artworks are appropriately unified, such entities could not be awarded the status of substances simpliciter by (ICCS4), since their ontological dependence on a numerically distinct entity that is extrinsic to them constitutes a violation of clause (ii.c). If artifacts in general are essentially identity dependent on the artisans who produce, design, or invent them, perhaps with a certain creative intention in mind, then the same result follows more broadly for the entire category of artificially made objects. Those proponents of independence criteria of substancehood simpliciter who take it to be a desideratum of their account that artworks or artifacts more generally can be classified as substances simpliciter would thus have to weigh their options in the face of the possibility that these entities might be excluded from the category of substances simpliciter, due to the extrinsicness of their individuation conditions.30, 31 30   Those who find (ICCS4) attractive might respond to the consideration just raised by rejecting the central assumption used in generating it, viz., that artworks in particular and perhaps artifacts more generally are in fact essentially identity dependent on their creators, in Lowe’s sense. The special metaphysical issues raised by artifacts will be investigated in more detail in Chapter 8. For present purposes, I am content to note that, given apparently plausible assumptions which at least cannot be dismissed out of hand, (ICCS4) will exclude these objects from the status of substances simpliciter as well as any other entities (if there are any) whose synchronic identity is fixed by appeal to numerically distinct entities extrinsic to them. The same result would have followed from any of the previous formulations of Gorman’s independence criterion as well. Thus, it is not the addition of the extrinsicness clause in particular which generates the result that artworks and/or artifacts in general are excluded from the status of substances simpliciter; the most recent revision merely attempts to make explicit the motivation which presumably lies behind the exemption for proper parts in the first place. Alternatively, those in favor of (ICCS4) might also consider it to be an advantage that this criterion excludes artworks and artifacts from the category of composite substances simpliciter due to the extrinsicness of their individuation conditions. 31   In addition, those who endorse the essentiality of origins will have to worry that a certain subclass of natural things as well will be excluded from the category of substances simpliciter by (ICCS4), on the grounds that they are essentially identity dependent on numerically distinct entities which are extrinsic to them (e.g., the particular egg and sperm from which a living organism originates).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

alternative strategies  185

6.4.5  Matter–Form Compounds In the foregoing sections, we have focused on two recent attempts at providing an independence criterion of substancehood simpliciter in the Aristotelian tradition, viz., E. J. Lowe’s criterion, stated in (ICS1), and a modified version of it proposed by  Michael Gorman, stated in (ICS2). I objected to the stipulative exclusion of ­non-particulars contained in both (ICS1) and (ICS2) on the grounds that this restriction makes it difficult to account for the apparently substantive nature of ­certain ontological disputes over questions of fundamentality. Moreover, our discussion of Gorman’s stipulative exclusion-clause governing part-dependence seemed to indicate that unity might have an important role to play instead of, or at least in addition to, independence in formulating a criterion of substancehood for composite entities. I now want to bring these considerations to bear on the question raised at the very beginning of this chapter, namely whether and how it might be possible to preserve the alleged substance status of at least some matter–form compounds. As I pointed out there, an apparent conflict emerges when we try to combine three central tenets popular among Aristotelians: (i) a hylomorphic analysis of concrete particular objects; (ii)  an independence criterion of substancehood simpliciter; and (iii) the desire to include at least some matter–form compounds among the substances simpliciter. For if alleged candidates for the status of substances simpliciter, such as organisms, are analyzed in the hylomorphic fashion, as compounds of matter and form, one wonders whether they will not then turn out to be ontologically dependent on entities numerically distinct from themselves (viz., their form and possibly their matter as well), thereby jeopardizing their inclusion in the category of substances simpliciter. This question should certainly be of concern to Aristotelians who find themselves attracted to the three tenets at issue. 6.4.5.1  Lowe’s Strategy When we attempt to apply Lowe’s criterion in (ICS1) to the case of matter–form compounds, we seem to run into the following problem. Suppose that a matter–form compound is numerically distinct from its form and matter. If, following Lowe’s definition in (ED3), a matter–form compound turns out to be essentially identity dependent on its form or its matter (assuming that the form or matter associated with a matter–form compound are particulars), then (ICS1) will exclude such compounds from the category of substances simpliciter. To prevent the outcome that matter–form compounds are disqualified from the status of substances simpliciter, Lowe therefore must deny one (or more) of the following claims: (1) A matter–form compound is numerically distinct from its form. (2) A matter–form compound is numerically distinct from its matter. (3) A matter–form compound is essentially identity dependent on its form. (4) A matter–form compound is essentially identity dependent on its matter.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

186  Independence Criteria of Substancehood (5) The form of a matter–form compound is a particular. (6) The matter of a matter–form compound is a particular.32 As noted earlier, Lowe opts for the denial of (1), among other things, and endorses an approach according to which concrete particular objects are identified with their particularized forms (see Lowe (1999)).33 As discussed in Chapter 4, my own preference lies with a compositional approach to the form–compound relation (see Section 4.4.4). 6.4.5.2  Gorman’s Exemption for Part-Dependence If an independence criterion for substances simpliciter is formulated in such a way that it contains an exemption for part-dependence, constituent-dependence, dependence on what lies “inside” the boundaries of an entity, or dependence on what is intrinsic to the entity in question, along the lines of Gorman’s modified criterion in (ICS2) and the series of revisions we have considered in (ICS3) and (ICCS1) through (ICCS4), then further possibilities are opened up for Aristotelians who wish to resolve the apparent conflict identified earlier between their commitment to hylomorphism, their sympathy for independence criteria of substancehood simpliciter, and their desire to count at least some matter–form compounds as substances simpliciter. For given the modified criterion and its subsequent revisions, a matter–form compound, assuming that it is appropriately unified, would be able to qualify as a substance simpliciter as long as those numerically distinct entities (if any) on which it is essentially identity dependent are either among its proper parts (in accordance with (ICS2), (ICS3), and (ICCS1)); or among its constituents (in accordance with (ICCS2)); or “inside” the boundaries of the matter–form compound in question (in accordance with (ICCS3)); or intrinsic to the matter–form compound in question (in accordance with (ICCS4)).34 But I take it that, on any reasonable formulation of the hylomorphic position, the form and/or matter of which a matter–form compound consists would, at least in certain cases, satisfy one, or possibly more than one, of these conditions.35 According to the mereological reading of the hylomorphic position, compounds of matter and form, strictly 32   In Section  6.4, I mentioned that Aristotelians might be tempted to reject at least one of the three claims, (a)–(c), cited there, if they wish to uphold (i)–(iii). (a) reappears here as (1). (b) and (c) will be considered shortly, in connection with the strategies open to Gorman. 33   Lowe would no doubt be happy to deny (4) as well: since matter–form compounds can apparently survive changes with respect to their material parts, it cannot be the case, given (ED3), that the synchronic numerical identity of a matter–form compound is fixed by the matter which composes it at any time or world at which it exists. But the denial of (4) alone is not enough to escape our current quandary: for, as long as (1) and (3) still hold, a matter–form compound nevertheless threatens to be essentially identity dependent on an entity that is apparently numerically distinct from itself (namely its form) and its (alleged) status as a substance simpliciter would thus still be in jeopardy, even if (4) is rejected. 34   The possibilities reviewed here present different ways of developing the denial of claim (c) cited earlier in Section 6.4. 35   The qualification, “at least in certain cases,” is meant to take into consideration the case of artworks and artifacts more generally considered in Section 6.4.4.4, which (assuming that these entities are nevertheless considered matter–form compounds) might present us with an exception to the generalization that all matter–form compounds would fall under at least one, and possibly more than one, of the exemption clauses added by Gorman.

OUP CORRECTED PROOF – FINAL, 07/27/2018, SPi

alternative strategies  187 and literally speaking, contain their form and matter as proper parts, thus qualifying as substances simpliciter under any of the versions of Gorman’s criterion.36 A modified version of this position is also available according to which the form and/or matter of which a matter–form compound consists are at least regarded as constituents, if not proper parts, of the entity in question. Even those who find neither of these interpretations of the hylomorphic position palatable may avoid the extreme measure taken by Lowe by endorsing one of the revised versions of Gorman’s modified criterion I offered for composite substances simpliciter in (ICCS3) and (ICCS4).37 Hence, even if matter– form compounds turn out to be essentially identity dependent on their form and/or their matter, they would not thereby be automatically excluded from the category of substances simpliciter given either Gorman’s exemption for part-dependence or any of the revised formulations I offered subsequently.38 But the objections I raised earlier speak against these sorts of approaches. For, as I pointed out in the preceding sections, once a Gorman-style exemption clause is added to an independence criterion of substancehood simpliciter, then all manner of composite entities which are apparently ontologically dependent only on their own proper parts (or constituents or what lies “inside” their boundaries or what is intrinsic to them) must be excluded from the category of substances simpliciter on other grounds (e.g., their lack of unity). The additional condition on the basis of which unwanted composite entities are excluded from the category of substances simpliciter then does all the work in the so-called “independence” criterion of substancehood simpliciter. In this case, the independence component of this criterion becomes powerless and therefore redundant. For these reasons, an alternative criterion of substancehood for composite entities is called for. 6.4.5.3  Form as Principle of Unity But there is a further and, to my mind, preferable option available to Aristotelians who already accept a hylomorphic analysis of concrete particular objects for independent reasons and who wish to preserve the substance status of at least some matter–form compounds. After all, according to the Aristotelian perspective, what distinguishes matter–form compounds from other, less unified, composite entities (e.g., sets, mereological sums, committees, and the like) is precisely that matter–form compounds contain within themselves a principle of unity which these other, less unified composite entities lack, namely, their forms. Traditionally, forms are assigned the special role of acting as the principle of unity within the matter–form compound, i.e., as that active principle within the matter–form compound which somehow ties together its material components into a single unified whole, as opposed to, for example, a mere heap, aggregate, or plurality. Aristotelians thus would seem to be missing out on an important 36   This is the version of hylomorphism for which I argue in Koslicki (2008a) and which is also endorsed in Fine (1982) and (1999). 37   Adherents to this variety of hylomorphism would be committed to the denial of claim, (b), cited earlier in Section 6.4. 38   For alternative ways of spelling out the hylomorphic position, see, for example, Harte (2002); Johnston (2002, 2006); Rea (2011).

OUP CORRECTED PROOF – FINAL, 07/27/2018, SPi

188  Independence Criteria of Substancehood advantage they gain through their commitment to hylomorphism if they did not also capitalize on the special role of form as the principle of unity within the compound in their quest to formulate an adequate criterion of substancehood for composite entities. If we are to take this option seriously, as providing us with a credible route toward a unity criterion of substancehood for composite entities, we need to know more about what it means to designate form as that active principle which plays the role of the unifying the matter–form compound. It will be my aim in Chapter 7 to investigate further in what sense matter–form compounds deserve to be designated as more unified than other complex entities which lack the presence of form within them.

6.5  The Status of Forms In this final section, I turn briefly to the question of how forms themselves fare with respect to the criteria of substancehood we have considered in this chapter. Interestingly, if we apply the definitions of ontological dependence reviewed above to the case of forms, we arrive at the result that, while forms are in some respects ontologically independent of matter–form compounds, they are also in other respects ontologically dependent on matter–form compounds. Firstly, given that the form of a matter–form compound in some way figures into a matter–form compound’s criteria of identity (see Section 3.4.3), essence or real definition (see Section 3.4.2), it cannot also be the case that the criteria of identity, essences, or real definitions of forms in turn make reference to the matter–form compounds in which these forms are present, since otherwise a vicious circularity would result. To illustrate, suppose that (in accordance with the individual forms hypothesis) what it is to be Socrates is to be some particular living organism composed of a particular human soul and some suitable body. Then, according to (ED4) and (ED5), Socrates is constitutively essentially dependent and constitutively definitionally dependent on his soul, since Socrates’ soul is a constituent of Socrates’ real definition or essence (narrowly construed). Moreover, Socrates is also essentially identity dependent on his soul, since, as required by (ED3), there is a function, φ, viz., the “hylomorphic composition” function (assuming the Uniqueness of Hylomorphic Composition), such that it is part of Socrates’ essence that he is the hylomorphic compound which results from the  presence of Socrates’ soul in some suitable body. Since Socrates is (ED3)–(ED5)-dependent on his soul, we must assume that a statement of the criteria of identity, essence, or real definition of Socrates’ soul does not in turn make reference to Socrates. Thus, it cannot also be the case that what it is to be Socrates’ soul is to be, say, that active principle which results from “hylomorphically subtracting” Socrates’ body from the particular matter–form compound that is numerically identical to Socrates. We must therefore proceed on the assumption that forms are ontologically independent of matter–form compounds, when ontological dependence is construed along the lines of (ED3)–(ED5).39 39   Suppose that (following the universal forms hypothesis) what it is to be a human being is to be a living organism composed of some human soul and some suitable body. Then, the kind of thing, human soul,

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

alternative strategies  189 Secondly, however, when ontological dependence is construed in the modal-existential way, along the lines of (ND1) or (ND2), forms will turn out to be ontologically dependent on matter–form compounds. For one of the marks distinguishing Aristotelian from Platonic forms is that the former cannot, while the latter can, exist apart from the concrete particular objects with which they are associated. Whether Aristotelian forms are rigidly or generically existentially necessarily dependent on matter–form compounds depends on where we stand with respect to the dispute discussed in Chapter 3 over the ontological category to which forms should be assigned. If forms are construed as individual or particular entities, then necessarily each such form exists only if the very matter–form compound with which it is associated exists and, by (ND1), forms are therefore rigidly necessarily existentially dependent on these particular matter–form compounds. In contrast, if forms are taken to be universal or general entities, then the existence of each such form requires only that some matter– form compound or other of a certain kind exists and, by (ND2), forms are therefore generically necessarily existentially dependent on some matter–form compounds or other of a certain kind. Either way, however, forms will turn out to be ontologically dependent on matter–form compounds, when ontological dependence is construed with either (ND1) or (ND2) in mind. Unless we somehow discount the modal-existential dependence of forms on matter– form compounds, we must therefore conclude that forms, like the matter–form compounds with which they are associated, do not qualify for the status of substances simpliciter, as long as we approach this notion through the lens of ontological independence. At the same time, however, when we compare forms and matter–form compounds with respect to the number of ways in which each type of entity turns out to be ontologically dependent on entities numerically distinct from it, then forms overall are classified as ontologically dependent on matter–form compounds in fewer ways than matter–form compounds are classified as ontologically dependent on the forms that are associated with them. Thus, relatively speaking, even if forms do not gain admission into the category of substances simpliciter, they nevertheless ought to be recognized as more deserving of substance status or more fundamental than matter– form compounds, according to a comparative notion of substancehood or ontological fundamentality, when these notions are viewed with an ontological independence criterion in mind. Based on the considerations advanced in this and subsequent chapters,

figures in the real definition or statement of the essence (narrowly construed) of the kind of thing, human being; and hence human beings in general are ontologically dependent on their souls, in the sense of (ED4) and (ED5). Proponents of the universal forms hypothesis are not committed to the claim that matter–form compounds are (ED3)-dependent on their universal forms, since this construal of ontological dependence concerns some particular entity’s being the very thing that it is at each time at which it exists. Thus, in light of the conflict identified above concerning (i)–(iii), one possible strategy for proponents of independence criteria of substancehood who also adhere to the universal forms hypothesis is to deny (5) (viz., the claim that matter–form compounds are essentially identity dependent on their forms; see Section  6.4.5.1). Nevertheless, the results derived earlier for (ED4) and (ED5) hold for both the universal and the individual forms hypothesis alike.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

190  Independence Criteria of Substancehood we thus arrive at the following ranking of entities with respect to their relative or absolute fundamentality status: (I) Matter–form compounds are not substances simpliciter (in the non-taxonomic sense). (II) Matter–form compounds are less deserving of substance status (non-­ taxonomically) than the forms associated with them. (III) Matter–form compounds are more deserving of substance status (non-­ taxonomically) than other types of composite entities (e.g., heaps, mereological sums, sets, collections, quantities, etc.).40 (IV) Like matter–form compounds, forms are also not substances simpliciter (in the non-taxonomic sense).

6.6 Conclusion In this chapter, I considered attempts by E. J. Lowe and Michael Gorman at defending an independence criterion of substancehood simpliciter and argued that the stipulative exclusion of non-particulars and proper parts (or constituents) from such accounts raises difficult issues for their proponents. The results of the present discussion indicate that, at least for the case of composite entities, unity has at least as much, and perhaps more, to offer than ontological independence and therefore ought to be explored further by Aristotelians in search of a defensible notion of substancehood. In Chapter 7, we will examine how unity might be used by Aristotelians in connection with a comparative conception of substancehood, to support the ranking of matter–form compounds as more deserving of substance status than other complex entities, by virtue of their highly unified nature. The proper conclusion to draw from these considerations, in my view, is that neither matter–form compounds nor their forms should be regarded as substances simpliciter. In fact, at least for non-theists, it is not obvious that we have any reason at all to believe that there are substances simpliciter, at least as long as we approach substancehood in the manner of (IND) and (NICSS). This negative conclusion, however, should not prevent us from classifying matter–form compounds or their forms as more deserving of substance status than other types of entities. But such a relative fundamentality ranking is not to be confused with a classification of matter–form compounds or their forms as belonging to the category of the absolutely fundamental or the substances simpliciter. At the same time, the importance of assigning matter–form compounds or their forms to their proper position within the hierarchy of the relatively fundamental should not be underestimated, since many of the most interesting metaphysical projects, especially those that are dear to Aristotelians, take place within the sphere of what, in absolute terms, would count as ontologically derivative.   The justification for (III) still awaits our discussion of unity in Chapter 7.

40

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

7 Unity 7.1  Introductory Remarks In Chapter 6, I argued that matter–form compounds, due to their composite nature, fail to be ontologically independent according to any of the construals of ontological dependence considered in Chapter 6. As a result, as long as the notion of substancehood simpliciter is approached through the lens of ontological independence, matter–form compounds will be excluded from the category of substances simpliciter. Despite their composite nature, however, matter–form compounds do deserve to count as ontologically privileged in other ways, since, among other things, they are more unified than other composite entities (e.g., heaps, non-empty sets, etc.). A unity criterion of substancehood thus holds more promise than an independence criterion to deliver the sought-after fundamentality ranking for composite entities, despite a general preference among Aristotelians for the latter. Since I take unity to be a comparative notion, the resulting notion of substancehood yields a relative, rather than an absolute, fundamentality ranking of (types of) composite entities based on their being more or less unified than other (types of) composite entities. Our next job, then, is to develop a serviceable account of unity which allows us to make the required comparative judgments between more or less unified composite entities. In what follows, I propose a conception of unity which is inspired by and modifies an approach originally defended in Nicholas Rescher and Paul Oppenheim’s “The Logical Analysis of Gestalt Concepts” (Rescher and Oppenheim (1955)). According to the modified account I offer here, what unifies a structured whole is not, as Rescher and Oppenheim have it, that its parts functionally depend on each other for the values they take with respect to their contingent physical attributes (since all physical systems meet this condition). Nor, as I argue in what follows, can the unity of structured wholes be said to be traced to the fact that their parts ontologically depend on each other for their existence, identity, essence, or real definition (since this requirement is not generally satisfied even by highly unified structured wholes). Instead, I propose that a structured whole derives its unity from the way in which its parts interact with other parts to allow both the whole and its parts to manifest those of their capacities that require “team work” among the parts. Interesting differences with respect to the source of their respective principles of unity will emerge between paradigmatic matter–form compounds belonging to natural (e.g., physical, chemical, or biological) kinds on the one hand, and composite entities belonging to social (e.g., artifactual) kinds on the other

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

192  unity hand. In the latter cases, we will find that the interactional dependencies that connect the components of a system can be traced to mind-dependent factors that are extrinsic or external to the system in question (e.g., the mental states of agents who invent, design, produce, or use an artifact).

7.2  Wholeness, Structure, and Unity In Koslicki (2008a), I argue that we should conceive of wholes as structured entities, namely, in particular as compounds of matter and form, subject to the following restricted composition principle: (RCPT) Restricted Composition Principle:  Some objects, m1, . . . , mn, compose an object, O, of kind, K, at a time t just in case m1, . . . , mn satisfy at t the constraints dictated by some formal components [. . .], f1, . . . , fn, associated with objects of kind, K. (Koslicki (2008a), p. 190) I construe the constraints that are dictated by an object’s formal components as ­primarily structural (see also Section 4.4.4.2): Structural constraints: Requirements concerning (1) the type or variety of material components; (2) their configuration, i.e., manner of arrangement; and, in some cases, (3) the number of material components of which objects of kind, K, may be composed. I take (1) and (2) to be necessary conditions which must be met by an entity if it is to count as a whole and hence as structured. (3) is met only in some, but not all, of the cases in which an entity satisfies (1) and (2) (e.g., (3) is satisfied by H2O molecules, but not generally by living organisms), and is therefore not a general requirement for an entity’s status as a whole. The restricted composition principle just cited carries with it a commitment to an ontology of kinds. For, according to (RCPT), a plurality of objects is said to compose a whole of a particular kind, when the objects (material components) in question satisfy the selection requirements set by the formal components associated with wholes of that particular kind, i.e., requirements concerning, among other things, the variety, configuration, and sometimes the number of material parts out of which wholes of that particular kind may be composed. Such a restriction on composition, of course, only has plausibility if there are independent reasons for thinking that objects really do belong to kinds and that kinds really do pose constraints on the mereological composition of their members. In my earlier account, I assumed that the question of which kinds there are is not answered by the mereologist proper, but by the ontologist at large, in conjunction with considerations coming from other domains, such as science and common sense, which turn out to have something to contribute to the question, “What is there?” or, more specifically, to the question, “What kinds of objects are there?”. For the special case of natural kinds, for example, our commitment to these kinds is typically

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Wholeness, Structure, and Unity  193 supported by an appeal to their role in prediction and explanation; particularly noteworthy in this respect is the weight borne by scientific natural kinds (e.g., physical, chemical, and biological kinds) in (i) inductive reasoning, (ii) the laws of nature, and (iii) causal explanations. Once such independent, extra-mereological reasons for believing in the existence of a certain kind of object have been given, we find in general that objects must satisfy more or less stringent mereological constraints in order to count as instances of the kind. Given this account, can we link up structure and unity in such a way that a whole counts as unified just in case it is structured, i.e., in the sense specified by (RCPT) together with (1) and (2)?1 In that case, a whole would be unified just in case it satisfies the structural constraints imposed by its formal components governing the variety of material components of which wholes of the particular kind in question may be composed as well as the configuration exhibited by these material components. But according to (RCPT) mereological composition only occurs in the first place when the conditions specified in this principle are met. Therefore, according to this approach, all entities which qualify as wholes at all would be classified as both structured and unified, and as unified to the same extent, since in order for something to be a whole at all (i.e., in order for mereological composition to take place successfully) the entity in question must be structured in the sense specified by (RCPT) in conjunction with (1) and (2). I will refer to this proposal as (WSU), short for “Wholeness = Structure = Unity.” As it stands, then, (WSU) apparently does not make room for the possibility of wholes which are less unified than matter–form compounds. Where does this leave us with respect to other types of composite entities which are apparently less unified than matter–form compounds (e.g., heaps, non-empty sets, committees, and the like)? If we link up wholeness with structure and unity in the way indicated by (WSU), non-empty sets, for example, can be classified as non-unified, because they lack the kind of structure required by (RCPT) together with (1) and (2). We cannot in any meaningful sense attribute formal components to non-empty sets whose job it is to place constraints on these entities which govern the variety and configuration of constituents of which they may consist. Assuming that a non-mereological construction operation (viz., set formation) is involved in “building” sets from their members, the question of whether sets exist does not hang on whether the set-building operation satisfies (RCPT), in conjunction with (1) and (2). Still, (WSU) yields a way of categorizing nonempty sets as non-unified, since they fail to satisfy the structural constraints specified in (1) and (2). On the assumption, then, that an entity must exhibit the sort of structure required by (1) and (2), in order for it to count as a whole and as unified, non-empty sets would fail to qualify for this status.2 1   This is what I gravitate toward in Koslicki (2008a), in the final pages of Chapter 7; however, for reasons which will become apparent shortly, I have since come to believe that my remarks concerning unity there were too quick. 2   In what follows, I will employ the following vocabulary in order to describe construction or composition operations in a neutral way, leaving open whether the operation in question satisfies (RCPT) together

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

194  unity Heaps constitute an interesting intermediate case: while they fail to satisfy the first structural constraint (viz., variety), they do at least in a loose way meet the second structural constraint (viz., configuration).3 Few, if any, type restrictions govern the constituents from which heaps may be formed: in principle, one may pile almost anything on top of anything; and, if there are restrictions, they seem to be of a more practical, rather than metaphysical, nature. But there are some configuration requirements which a plurality of objects must satisfy in order for these objects to form a heap, viz., they must at least be arranged in certain spatiotemporally proximate positions so that the overall result of their arrangement has the approximate shape of a heap. In this way, heaps are somewhat more structured than non-empty sets, say, but less so than matter–form compounds. The proposal currently under consideration thus opens up a way of assigning an intermediate status to heaps which would place them between matter–form compounds on the one hand, and non-empty sets on the other hand, since they satisfy some, but not all, of the structural constraints which would confer on an entity the highly unified status characteristic of matter–form compounds. Heaps do not quite make it into the (WSU) category; but they come closer to doing so than non-empty sets do, since they at least satisfy one of the two structural requirements an entity must satisfy for full (WSU) status. Not surprisingly, natural matter–form compounds meet the criteria necessary for inclusion into the (WSU) category with flying colors. For Aristotle, the contrast between heaps and natural matter–form compounds comes down to whether an entity’s constituents are held together by an internal principle of unity whose presence in the object is due to nature, rather than art. In the case of a heap, the constituents (e.g., the sticks in a bundle) are held together, if at all, only by an external and artificially imposed principle of unity (e.g., a band tied around the bundle of sticks). But the principle of with (1) and (2): “construction operation,” “composition operation,” “collection,” “constellation,” “plurality,” “system,” “assembly,” “constructed entity,” “composite entity,” “member,” “constituent,” “component,” “element,” “consists of,” “composes,” “is composed of,” and “forms” (used as a verb). Thus, we may speak, for example, of a collection or plurality of floating magnets (an example from Rescher and Oppenheim (1955) I will go on to discuss ) as a system, a constellation, or an assembly of entities; of a non-empty set as a collection, or a constructed, or composite entity; of Socrates and Plato as being members, constituents, components, elements of, or as composing or forming the set containing Socrates and Plato; of the set in question as being composed of or consisting of Socrates and Plato; and of the operation of set-formation as a construction or composition operation. For the reasons just noted, however, it would be inappropriate to describe Socrates and Plato as parts of the set containing Socrates and Plato or of the set in question as a whole that is mereologically composed of Socrates and Plato, since the set-building operation fails to satisfy the structural requirements set by (1) and (2). While I am committed to the view that mereological composition for concrete particular objects is governed by (RCPT) together with (1) and (2), I want to leave open that there may be other types of composition operations which apply in other cases and which do not satisfy these structural constraints. (See Fine (2010), however, for a more general understanding of mereological composition and parthood, according to which the members of a set can be said to be part of the set.) 3   According to the positive account of unity I will develop here, such entities as ice cubes or columns of smoke, for example, are classified as being roughly on a par with heaps of sand, piles of stones, or bundles of sticks with respect to the extent to which they are unified. I intend to use the term, “heap,” in a sufficiently broad manner to include all of these cases. (Thanks to Friederike Moltmann, pers. conv., for pressing me on this point.)

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Wholeness, Structure, and Unity  195 unity, in this case, is not one which is internal to the bundle of sticks or which arises naturally from the sticks in the bundle; rather, it is an element that is added to the sticks, by art rather than by nature, for the specific purpose of giving some cohesion to the bundle of sticks which it would otherwise lack. In this case, the principle of unity (the band) which holds together the remaining constituents (the sticks in the bundle) belongs to the same ontological category as the constituents on whom it confers cohesion. If the category of matter–form compounds includes artifacts, then each of the constituents of the heap would already in itself count as more highly unified than the entire constellation into which these constituents enter when they form a heap. The case of artifacts will concern us in more detail in this chapter and Chapter 8. In contrast, at least in the case of natural matter–form compounds, the unity of the whole, in Aristotle’s view, can be traced to the presence of a principle of unity (its form) that is both natural and internal to the matter–form compound. In this case, the form, which acts as the principle of unity for the matter–form compound, belongs to a different ontological category from that occupied by the material parts it serves to unify. As empirical evidence for the unifying power of form in natural matter–form compounds, Aristotle points, for example, to the contrast between the body of a living organism and the corpse it leaves behind. While the organism is alive, its form, viz., the soul, holds together the organism’s body and the material parts which are present in it only potentially. But after the organism has died, the corpse it leaves behind slowly begins to decay and disintegrates into the multiplicity of material constituents which are now actually contained within the corpse. With no soul present in it to act as principle of unity, there is nothing to prevent the earth, air, fire, and water which actually compose the corpse, in Aristotle’s view, from slowly separating from each other, following their natural tendencies. Thus, based on the considerations rehearsed so far, (WSU) already allows us to establish at least the following ranking among composite entities with respect to the extent to which they are unified: Natural matter–form compounds:  highly structured and unified wholes which satisfy both (1) and (2). Heaps:  entities that have constituents, but are less highly structured and unified than matter–form compounds; they do not satisfy (1) and satisfy (2) only in a loose and undemanding way.4 4   An interesting case to consider (suggested to me by Marko Malink) is that of a circle composed of curved segments. (For the sake of specificity, I construe the case at hand as involving a concrete token of the abstract geometrical type, circle, which is not created by an intentional agent for a particular purpose.) We are here apparently presented with a case in which both sorts of structural constraints are satisfied: a concrete instance of the geometrical type, circle, must be composed of parts of the type, curved segment, and these parts must be arranged in a circular way. However, the configuration constraints which must be satisfied in order for a collection of curved segments to compose a concrete token of the geometrical type, circle, are merely shape-related, as in the case of heaps. This example thus points to the possibility of drawing more fine-grained distinctions than those offered here among composite entities based on the idea that configurational constraints themselves can come in more or less demanding varieties. Although this direction

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

196  unity Non-empty sets:  entities that have constituents, but lack any kind of structure and unity; they satisfy neither (1) nor (2).5 As it stands, (WSU) does not yet make room for comparisons among more or less unified composite entities which satisfy both sorts of structural constraints listed in (1) and (2). The following types of composite social (e.g., artifactual or institutional) entities, however, put some pressure on the account given by (WSU), since it is at least not prima facie obvious how they ought to be classified in relation to paradigmatic matter–form compounds which belong to natural kinds. To illustrate, a screwdriver, for example, in order for it to be suitable for its intended purpose (viz., to loosen and tighten screws), should ideally consist of a handle (i.e., a part which allows the user to hold the screwdriver while tightening and loosening screws), a tip (i.e., a part which has the right shape to fit into the head of the type of screw on which the screwdriver is intended to be used), and a shaft (i.e., a part which connects the handle and the tip). The screwdriver’s tip and shaft should be made of sufficiently hard material relative to the material composing the screws on which the screwdriver is intended to be used, so that the tip and the handle can maintain their physical integrity even after repeated use. The material of which the handle consists, as well as its shape, in turn are chosen in such a way as to fit comfortably into the user’s hand and to maximize the user’s ability to employ the device for its intended purpose. Screwdrivers, and artifacts more generally, thus appear to present us with examples of composite entities whose material parts must satisfy various type constraints and configuration constraints in order to compose a whole of the kind under consideration.6 Another interesting case to consider in this connection, is that of composite social entities belonging to institutional kinds, e.g., football teams. As long as we have in mind a sufficiently tolerant understanding of “configuration,” it appears that instances of institutional kinds as well can be made to fit the structural constraints cited earlier. An American football team, for example, consists of a certain number of offensive and strikes me as in principle a promising one to pursue further, I will in what follows focus on a different set of issues which require the introduction of additional apparatus not relevant to this particular case. 5   This raises the question: are there entities which satisfy the type constraints in (1), but not the configuration constraints in (2)? And, if so, how should these items be ranked with respect to their status as more or less unified than the other types of composite entities considered so far? We might think that certain kinds of collections, e.g., committees, fit this category. To illustrate, consider the US House Ways and Means Committee which is governed by very tight restrictions as to who can serve on it. Does the fact that the members of this committee satisfy certain type constraints warrant that the committee is designated as more unified than other types of composite entities whose components fail to satisfy analogous type constraints? If so, where should such a committee rank with respect to its unity compared to that exhibited by other types of composite entities, e.g., heaps? In principle, the ranking I have offered allows for the possibility of recognizing an additional category for entities satisfying (1), but not (2). Social kinds more generally, of which committees are a subspecies, turn out to be an interesting test case for my own favored approach to unity, and will concern us in more detail subsequently. 6   The reader may have noticed how natural it is to slip into normative language when describing the material components of which an artifact ideally should be composed and the configuration these material components ideally should exhibit if the artifact is to be used for its intended purpose. We will have occasion to return to this important characteristic of artifacts later in this chapter.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Wholeness, Integrity, and Unity  197 defensive players who must be at least of the type, human being, in addition to whatever other requirements might be imposed on those who qualify for membership in a particular team. Due to their roles on the team (e.g., quarterback, line backer, etc.), these players occupy various positions within a configuration whenever they are engaged in carrying out a particular play. In this way, football teams, with their players as parts, seem to satisfy the structural constraints placed on wholes by (WSU), and therefore apparently gain admission into the category of highly unified matter–form compounds. In what follows, I focus on the case of artifacts; a detailed discussion of institutional entities would take us too far afield for present purposes.7 Such cases pose a challenge to (WSU) at least on the assumption that some distinction ought to be drawn between composite entities belonging to social kinds on the one hand, and paradigmatic cases of matter–form compounds coming from the natural realm (e.g., living organisms or chemical compounds) with respect to the extent to which they are recognized as unified. On the basis of what has been said so far, it is not yet obvious on what grounds composite entities belonging to social kinds could be assigned a different status from that granted to paradigmatic cases of highly unified structured natural wholes belonging to (WSU)’s top category.

7.3  Wholeness, Integrity, and Unity The treatment of “Gestalt concepts” in Rescher and Oppenheim (1955) bears directly on the question of how the notions of wholeness, structure, and unity are interconnected. Rescher and Oppenheim are interested in proposing criteria which would, for example, disqualify a system consisting of three stones, selected at random, located in different continents, from counting as a genuine whole (Rescher and Oppenheim (1955), p. 90). As a starting point for their discussion, Rescher and Oppenheim offer three informal conditions as “intuitive requirements or conditions of adequacy which underlie our talk of wholes” (Rescher and Oppenheim (1955), p. 90; their emphasis throughout): (I) The whole must possess some attribute in virtue of its status as a whole—an attribute peculiar to it, and characteristic of it as a whole. 7   One might think that composite social entities, such as institutional entities or artifacts, can be immediately disqualified from the top (WSU)-category of highly unified structured wholes, on the grounds that the components of the system in these cases need not be continuously configured in a certain characteristic way at all times at which the system in question exists. For example, once a particular football game is over, the players who participated in the game can go their own separate ways without their teams therefore ceasing to exist. A similar characteristic seems to be prevalent among the members of many artifact kinds as well (e.g., bicycles, watches, trumpets, etc.), since they can apparently survive being disassembled and reassembled. In what follows, I will not place a heavy explanatory burden on the feature just cited, however, since it is possible for the constituents of a composite social entity to be permanently, and not just temporarily, configured in a certain way characteristic of the kind in question. A case in point might be, for example, social groups related by familial relations, such as being a parent of, being a child of, being a sibling of, etc., assuming that such systems in fact meet the structural constraints cited in (1) and (2).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

198  unity (II) The parts of the whole must stand in some special and characteristic relation of dependence with one another: they must satisfy some special condition in virtue of their status as parts of a whole. (III) The whole must possess some kind of structure, in virtue of which certain specifically structural characteristic pertain to it.

We may refer to systems which satisfy Rescher and Oppenheim’s three criteria as integrated wholes, and to their components as parts that are integrated into these wholes.8 Since Rescher and Oppenheim’s third criterion is already contained in (WSU), I will in what follows simply take for granted that we should accept some version of (III), in particular one which interprets this criterion along the lines of the structural constraints in (1) and (2), as Rescher and Oppenheim do. But the first and second criterion go beyond (WSU) and it is thus worth asking whether these criteria move us forward in our search for a meaningful and non-arbitrary way of ranking composite entities as more or less unified than some relevant comparison class. I will refer to the new proposal under consideration as (WIU), short for “Wholeness = Integration = Unity.” (WIU) coincides with (WSU) with respect to (III), but supplements (WSU) with the addition of two new criteria, (I) and (II). I begin with the question of how composite entities which meet the structural requirements set by (1) and (2) fare with respect to Rescher and Oppenheim’s first criterion. Is there in general some attribute which these entities possess that is “peculiar” to them or “characteristic” of them as composite entities? When Rescher and Oppenheim speak of an attribute that is “peculiar” to and “characteristic” of a whole, one that it possesses “in virtue of its status as a whole,” they have in mind the following (stated here in a simplified manner): Rescher and Oppenheim’s First Criterion:  An integrated whole must possess an underivable attribute, namely, an attribute which is (I.1) not also possessed by its parts; and (I.2) whose possession by the whole cannot be fully explained, relative to some theory, by appeal to the attributes possessed by its parts.9

8   For other relevant discussion, see also Fine (2010) and Simons (1987), especially Ch. 9, which examines Rescher and Oppenheim’s proposal in detail and develops Simons’ own notion of “integral whole.” In what follows, I will employ the terms “integral” and “integrated,” when used in connection with wholes, as synonyms. The notion of an integrated whole is also employed by Friederike Moltmann: Moltmann (1990, 1997a,  1997b,  1998,  2005). Moltmann’s notion of integrated whole is much less restrictive than the one I will be developing here: according to her account, heaps of sand and football teams, for example, would count as integrated wholes. Schaffer (2013) employs the concepts of unity and integration in an argument in favor of priority monism, on the grounds that “to be one is to act as one” and only the entire cosmos, in his view, is a truly unified and integrated nomic system, governed by the fundamental laws. 9   The appeal to derivability in (I.2) may be read as leaving open the possibility of full explanations which are not statable in the form of logically valid inferences. Rescher and Oppenheim also note (see p. 93) that their definition of an underivable attribute is adapted from the definition of emergence given in Hempel and Oppenheim (1948). For a recent discussion of the different ways in which the concept of emergence is employed, see, for example, Wilson (2016).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Wholeness, Integrity, and Unity  199 To illustrate, suppose a pile consisting of three stones weighs 3 kg, with each stone weighing 1 kg. Then, the pile of stones satisfies (I.1) by virtue of possessing the attribute of weighing 3 kg, since this attribute is possessed by the pile, but not by each stone individually, or by any two of the stones taken together. But the pile of stones does not thereby satisfy (I.2), since, given the right physical theory, the weight of the pile can be fully explained by appeal to the weight of the individual stones. Whether wholes ever really have attributes whose possession by the whole cannot be fully explained by reference to the attributes possessed by the parts is a controversial matter. Philosophers who are attracted to certain versions of physicalism or naturalism, for example, may wish to deny that any of the attributes possessed by a whole are ever truly underivable in the sense of (I.2). However, for the purposes of illustration, we may use as an example of an integrated whole which satisfies (I.1) and (I.2) a human being who possesses the attribute of being a moral agent: the parts of a human being individually, or smaller collections of these parts (arguably), do not themselves possess the attribute of being a moral agent and there is (arguably) no theory on the basis of which the possession of this attribute by the whole can be fully explained by appeal to the attributes possessed by the parts. The first criterion constitutes Rescher and Oppenheim’s attempt at making more precise what might be meant by the saying, “The whole is more than the sum of its parts.”10 Do composite social entities belonging to artifactual kinds, for example, satisfy (I.1) and (I.2)? We have at least the following two options in responding to this question. The first option assumes that we are serious about our ontological commitment to the entities in question, as numerically distinct from their constituents. According to this option, these entities belong to kinds to which we are already committed for independent reasons, and which therefore deserve separate entries in our ontological ledgers in addition to the entries we reserve for their constituents. In contrast, the second option assumes that, when pressed, we are inclined to opt for some sort of reductive analysis of the entities in question which somehow makes these entities disappear from our ontological ledgers and permits us to proceed with just the entries reserved for their apparent constituents. Choosing the second option leads to the result that these cases no longer constitute a potential challenge for the plausibility of (WSU). For if we feel drawn to the position that we need not really be committed to the existence of artifacts, for example, as entities in their own right numerically 10   If a certain entity possesses the attribute of being a moral agent, then, one might think, subtracting a few molecules from this entity still leaves us with something that is itself a moral agent. Does it follow from this that an entity which possesses the attribute of being a moral agent has proper parts which are themselves moral agents? This difficulty is an instance of what is known as the “Problem of the Many.” Any view which allows that moral agents can have proper parts that are themselves (numerically distinct) moral agents faces the unwelcome result that apparently a staggering number of numerically distinct moral agents occupy highly overlapping regions of space-time, all of whom seem to care about the welfare of a single person. Since I take mereological composition to be restricted in the manner indicated earlier, I am not in general committed to the idea that large proper parts of any given whole of kind, K, themselves compose a further whole of kind, K. In what follows, I assume that we can bracket the Problem of the Many for the purposes of the present discussion.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

200  unity distinct from their apparent constituents, then (given their non-existence) we can also stop worrying about whether, or to what extent, artifacts are unified. In that case, the question of whether these apparent entities satisfy Rescher and Oppenheim’s first criterion becomes a non-issue.11 Given the first option, Rescher and Oppenheim’s criterion, at least on its own, cannot effect a meaningful and non-arbitrary ranking of composite entities which meet the structural constraints in (1) and (2) with respect to the extent to which they are unified, at least given the following assumption which I take to be plausible. Philosophers who are serious about their ontological commitment to the composite entities in question, in my view, should also want to ascribe to these entities some attributes which are possessed by them but not by their components and whose possession by these entities cannot be fully explained by appeal to the attributes possessed by their components. A case in point is modal attributes. It follows from the non-modal conception of essence to which I expressed my sympathies in earlier chapters that those who are serious about their commitment to the existence of screwdrivers, for example, as entities in their own right numerically distinct from the natural matter– form compounds which apparently compose them, would also be well advised to hold that screwdrivers possess certain modal attributes, e.g., being capable of tightening or loosening screws, which their individual material components, or smaller collections of them, do not possess, and whose possession by a screwdriver cannot be fully explained by appeal to the attributes possessed by their material components. That screwdrivers are the kinds of entities that are capable of tightening or loosening screws does not follow from the attributes (including the modal attributes) possessed by their material components (viz., the handle, tip, and shaft), or smaller collections of them, though whether a screwdriver manifests this capacity on any particular occasion is at least partly determined by how its material components conduct themselves in relation to one another. Perhaps a screwdriver’s possession of such attributes can be fully explained by reference to the acts and intentions of agents who are involved with their invention, design, production, or use. (In fact, given that screwdrivers have been invented and created by human beings to serve certain purposes, I take it that some such account of how screwdrivers come to possess the modal attributes they apparently do possess is highly plausible; we will consider several options for how such an account might be developed in Chapter 8.) But these agents (e.g., inventors, designers, producers, or users) are not themselves among the material components of a screwdriver on which, 11   We may, however, nevertheless wonder why, given the second option, mereological composition does not in fact take place in such a case, even though by all appearances the structural requirements in (1) and (2) are satisfied. The second option would answer this question by denying that the apparent kind with which we are presented by the case at hand is in fact a genuine kind to which we already take ourselves to be committed for independent reasons. Thus, on this view, a plurality of natural matter–form compounds arranged in such a way as to allow agents to engage in behavior we would describe as “tightening screws” or “loosening screws” fails to compose a further entity, viz., a screwdriver, because strictly speaking our ontology does not include artifacts as genuine kinds of entities in their own right, numerically distinct from the natural matter–form compounds which apparently compose them.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Wholeness, Integrity, and Unity  201 through certain of their acts and intentions, they confer such modal attributes as being capable of tightening or loosening screws. We have after all assumed that if a screwdriver is a genuine structured whole, then its material components would be its handle, tip, and shaft. It would be unfortunate if our conception of artifacts led to the consequence that screwdrivers are mereologically composed of all those agents who have the power to contribute to the conferral of modal attributes on them. Instead, we should adopt the position that artifacts, assuming that we are serious about our ontological commitment to them, possess certain attributes whose possession cannot be fully explained in terms of the attributes possessed by their individual material components, or by smaller collections of them. If this point generalizes to other types of composite entities, on the assumption that we are to take our ontological commitment to these entities seriously as well, then we reach the conclusion that some underivable attribute can always be found by virtue of which these entities would satisfy Rescher and Oppenheim’s first criterion. As a result, this criterion, at least on its own, cannot serve to rank composite entities which satisfy the structural constraints in (1) and (2) as more or less unified than some relevant comparison class. The spotlight therefore is on the second criterion: we must now examine to what extent this criterion, together with our other apparatus, contributes to our ability to draw meaningful and non-arbitrary distinctions between composite entities which meet the structural constraints in (1) and (2) but differ with respect to their status as being more or less unified than some relevant comparison class. In their informal statement of the second criterion cited earlier, Rescher and Oppenheim formulate this criterion in two different ways: the first half of (II) requires the parts of a genuine whole to stand in some “special and characteristic relation of dependence” to one another, while the second half of (II) speaks of the parts of a whole as satisfying “some special condition in virtue of their status as parts of a whole.” But there is an interesting and important difference between requiring the constituents of a composite entity to stand in “some special and characteristic relation of dependence” to one another and requiring these constituents to meet “some special condition” in virtue of being constituents of the composite entity in question: for not every such “special condition” dictates that the constituents of the composite entity stand in a dependence relation to one another. If the “special condition” requirement is construed too liberally, then (II) simply collapses into (I). For unless some restrictions are placed on how the “special condition” which figures in (II) is to be met by the constituents of a composite entity, whatever attribute the composite entity in question possesses in virtue of which it satisfies (I) can immediately be translated into a “special condition” met by the constituents which will allow these constituents to satisfy (II). For example, suppose that among the special attributes possessed by a screwdriver, in virtue of which it satisfies (I), is the attribute of being capable of tightening or loosening screws. Then, the tip, handle, and shaft, in virtue of the fact that they compose the screwdriver, also possess the corresponding attribute of being able to contribute to the screwdriver’s ability to tighten and loosen screws. For this reason, I regard Rescher and Oppenheim’s first formulation of (II) as

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

202  unity stronger, more specific, and ultimately more helpful for our present purpose than their second formulation. In their more developed statement of (II), Rescher and Oppenheim in fact single out dependence relations as the relevant way in which the parts of an integrated whole must meet the “special condition” requirement. To state Rescher and Oppenheim’s second criterion, several technical notions are needed. “Definition 4” defines an attribute’s being “φ-dependent” on a class of attributes for the special case of quantitative attributes as follows: A quantitative attribute f of the part p1 in a configuration of kind R of the n objects p1, p2, . . . , pn is φ-dependent upon the class G of quantitative attributes of these objects pi, if φ is a relationship such that in every configuration of kind R the f-value of the first member of this configuration is related by φ to the values which the quantitative attributes in G assume for the parts p1, p2, . . . , pn.  (Rescher and Oppenheim, (1955) pp. 95–6)

A “configuration of kind R” is an ordered set of objects, p1, p2, . . . , pn, which stand in the relation R to each other, i.e., R(p1, p2, . . . , pn) (p. 95). To assert a φ-dependence of the type defined in Definition 4 (often a functional relationship) is to assert that the connection between the value taken by p1 with respect to f and the values taken by p1, p2, . . . , pn with respect to G is law-like, i.e., generalizes across all configurations of kind R. Rescher and Oppenheim hold that the extension of Definition 4 to the non-quantitative case is readily available. “Definition 5” defines a “φ-dependence system” as follows: A configuration is a φ-dependence system relative to a set G of attributes if each part of the configuration has some G-attribute which is φ-dependent upon (some or all of) the G-attributes of (some or all of) the remaining parts.  (Rescher and Oppenheim, (1955) p. 98)

With these definitions in place, Rescher and Oppenheim’s second criterion can now be stated as follows: Rescher and Oppenheim’s Second Criterion:  The parts of an integrated whole must form a φ-dependence system relative to some class of attributes, G. We shall say, in what follows, that the components of a system which satisfy the second criterion are unified under “lawful interdependence” with respect to some class of attributes, G, and relative to some set of laws or law-like connections. To illustrate, consider the following system which, in Rescher and Oppenheim’s view, satisfies the second criterion: . . . magnetic needles of equal strength are inserted in pieces of cork and floated, with all like poles upwards, in a basin of water. A strong magnet of unlike pole is placed in a fixed position overhead. The floating magnets arrange or rearrange themselves in a symmetric pattern of one or more concentric circles, depending on the number of magnets.  (Rescher and Oppenheim, (1955) p. 95)

In this system, as Rescher and Oppenheim point out, the distance of one cork from the nearest adjacent corks is a function of the strengths and number of magnets involved

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Wholeness, Integrity, and Unity  203 in the system. Thus, when the attributes in question include the distance of one cork from another as well as the number and strength of the magnets involved in the system, then the components of the floating magnet system are unified under the relation of lawful functional interdependence. The relationship in question is lawful, since the laws of physics dictate that similar systems behave in similar ways with respect to the functional determination of the distance of one magnet from the nearest adjacent magnet by the number and strength of magnets involved in the system. Additional cases which are used by Rescher and Oppenheim to illustrate the workings of the second criterion include the following: the price of a commodity is dependent upon the distribution of supply and demand over the commodities in a financial system, according to the laws (or law-like connections) of classical economic theory; the temperature of a body part is dependent upon the temperature of the other body parts in a homeostatic systems, according to the laws (or law-like connections) of biology; the pressure of a gas is dependent upon its volume and temperature, according to the Van der Waals law of physical chemistry; the perceived color of an area is dependent upon the actual colors of the areas in the system, just as the perceived length of a line segment is dependent upon the actual lengths and angles of the line segments in the system, according to the laws (or law-like connections) invoked by Gestalt psychology; finally, the timing of the movement of one insect leg is dependent upon the number of legs and the timing of movement of the remaining legs of the insect, according to the laws (or law-like connections) of physiology. Several important questions are currently still left open by the statement of the second criterion and the accompanying definitions cited earlier, among them questions concerning the nature and scope of the quantifiers implicit in (II). To avoid unnecessary clutter, the following disambiguations abstract away from the second criterion’s appeal to the values taken by the components of a system with respect to some class of attributes. Let D be a division of a whole into parts; then at least the following readings of (II) are available, beginning with the strongest and ending with the weakest:12 (IIStrong1) There is some dependence relation, R, such that every member of D bears R to every other member of D and no member of D bears R to anything that is not a member of D. (IIStrong2) There is some dependence relation, R, such that every member of D bears R to every other member of D. (IIStrong3) There is some dependence relation, R, such that every member of D bears R to some other member of D and no member of D bears R to anything that is not a member of D. 12   D is a division of a whole, w, just in case every member, x, of D is a part of w and for every part, x, of w there is a y that is a member of D, such that x overlaps y. A division whose members are disjoint is a partition (see Simons (1987), p. 327). I leave open whether the dependence relations in question are irreflexive.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

204  unity (IIStrong4) There is some dependence relation, R, such that every member of D bears R to some other member of D. (IIWeak1) Every member of D bears some dependence relation or other, R1, . . . , Rn, to every other member of D and only to members of D. (IIWeak2) Every member of D bears some dependence relation or other, R1, . . . , Rn, to every other member of D. (IIWeak3) Every member of D bears some dependence relation or other, R1, . . . , Rn, to some other member of D and only to members of D. (IIWeak4) Every member of D bears some dependence relation or other, R1, . . . , Rn, to some other member of D.13 The main differences between these readings of the second criterion are the following. The strong readings, (IIStrong1)–(IIStrong4), require that there be a single dependence relation, R, which relates the members of D to one another. (IIStrong1) and (IIStrong3) differ from (IIStrong2) and (IIStrong4), in that (IIStrong1) and (IIStrong3) require that the single dependence relation, R, in question must also differentiate the members of D from surrounding entities, while (IIStrong2) and (IIStrong4) allow the members of D to bear the single dependence relation in question to surrounding entities as well. (IIStrong3) and (IIStrong4) weaken (IIStrong1) and (IIStrong2), in that (IIStrong3) and (IIStrong4) only require the members of D to bear the single dependence relation in question to some, but not all, members of D. The weak readings, (IIWeak1)–(IIWeak4), allow there to be some variation in the dependence relations which members of D bear to other members of D. Otherwise, the weak readings exactly replicate the relationships described in the strong readings.14, 15 13   Simons informally defines an “integral whole” as one which fulfils or at least approximates the following condition: “Every member of some division of the object stands in a certain relation to every other member, and no member bears this relation to anything other than members of the division” (Simons (1987), p. 327). Simons’ condition amounts to the strongest reading stated above, viz., (IIStrong1), with the restriction to dependence relations removed. Simons is well aware that, unless some such restriction is imposed on the relations under consideration, the criterion in question threatens to trivialize. He is in principle sympathetic to the idea that the parts of integral wholes are unified under an interdependence relation of some sort, an idea which also prominently figures in the work of Edmund Husserl and Roman Ingarden (see Husserl (1900–1); Ingarden (1925, 1960, 1965, 1974)). In Section 7.4, we will have occasion to examine some of the dependence relations Simons considers as candidate substitutions into a plausible and nontrivial criterion of integral wholes (see Simons (1987), especially Ch. 9; as well as Simons (1994) and Mulligan, Simons, and Smith (1984) for further relevant discussion). 14   Further weakenings of the second criterion are available, for example, by changing “every” to some appropriate weaker-than-universal quantifier (e.g., “most,” “many,” etc.); further strengthenings of the second criterion are available, for example, by requiring the dependence relation(s) in question to be mutual. 15   Rescher and Oppenheim also employ a distinction between “strong” and “weak” dependence, but they have in mind here the distinction between deterministic and indeterministic systems. Rescher and Oppenheim propose that the degree of dependence exhibited by a system can be defined based on the extent to which the value taken by a component of the system with respect to a G-attribute is either fully (i.e., non-probabilistically) or less than fully (i.e., probabilistically) determined by the values taken by the components of the system with respect to the G-attributes. Accordingly, indeterministic systems in general will exhibit a lower degree of lawful interdependence than deterministic systems, assuming that the characterization of the system includes all the relevant information necessary to capture the determination relation in question. Merely

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Wholeness, Integrity, and Unity  205 To illustrate, substituting functional dependence for “R,” Rescher and Oppenheim’s floating magnet system satisfies (IIStrong2), (IIStrong4), (IIWeak2), and (IIWeak4); but not (IIStrong1), (IIStrong3), (IIWeak1) or (IIWeak3). (IIStrong2) holds with respect to this system, since the distance between any two members of D is functionally determined by the strength and number of magnets involved in the system. Since (IIStrong2) entails (IIStrong4), (IIWeak2), and (IIWeak4), the floating magnet system also satisfies these conditions. But the system fails to satisfy (IIStrong1), (IIStrong3), (IIWeak1), or (IIWeak3), because the same physical forces which are at work in determining the pattern in which the corks and magnetic needles are arranged also hold between the members of this system and every other physical object in the universe (however weakly). As philosophers of science like to point out, no system, with the exception of the entire universe, is truly isolated and the same physical forces act on everything that is located in space-time. Thus, when the second criterion is read with functional dependence in mind and relativized to the contingent physical attributes of a system, the result is an overly weak interpretation of Rescher and Oppenheim’s second criterion, according to which every physical system whatsoever counts as unified under lawful interdependence.16 Our next step, therefore, is to inquire whether there are viable alternative interpretations of Rescher and Oppenheim’s second criterion which can be used to draw meaningful distinctions between more or less unified systems.17 partial determination relations may also be due to epistemic factors, e.g., if our current state of knowledge concerning systems of the kind in question delivers only partial information concerning the relevant class of G-attributes. 16   Of course, not every such system will also qualify as an integrated whole, by the lights of Rescher and Oppenheim’s account. For example, even though the floating magnet system satisfies the second criterion, in Rescher and Oppenheim’s view (I.2) fails in this case, since the attributes possessed by the system as a whole (e.g., central symmetry) are derivable, i.e., they can be fully explained, relative to some theory, by appeal to the attributes possessed by the components of the system. In the remainder of this chapter, we will take up the important question of how the second criterion should be read in light of the restrictions placed on integrated wholes by the first criterion. 17   In addition to the disambiguations brought out by the weak and strong readings distinguished here, Rescher and Oppenheim’s illustrations of the second criterion cited earlier raise several further questions which are not conclusively settled in the relevant section of Rescher and Oppenheim (1955) (see “Dependence Systems,” pp. 95–9). According to Definition 4, the value taken by one component, p1, of a given system with respect to a φ-dependent attribute, f, depends upon the values taken by all the components of the system (including p1) with respect to the G-attributes (where it is assumed that f is not included in G). This definition is satisfied in the financial case, under the assumption that the price (f ) of a commodity is not included among the G-attributes (supply and demand) and that the distribution of G-values (supply and demand) over all the commodities in the system (including p1), i.e., p1, p2, . . ., pn, determines the price (f) of any one commodity in the system (e.g., p1). The relationship identified here is in effect a version of a mereological supervenience thesis, where the supervenience base includes the distribution of G-attributes over all the components in the system, p1, p2, . . ., pn, and the supervenient feature is the value of the φ-dependent attribute, f, taken by any one component in the system, e.g., p1. Definition 5, in contrast, states that a φ-dependence system is one in which the value taken with respect to some G-attribute, g, by one of the components, p1, is dependent upon the value taken by some or all of the remaining components (excluding p1), i.e., p2, . . ., pn, with respect to some or all of the G-attributes (possibly including g). This definition is satisfied, for example, in the case involving homeostatic systems, assuming that the temperature (g) of one body part, p1, is dependent upon the temperature (g = G) of some or all of the remaining body parts, p2, . . ., pn; or in the case involving perceived and actual colors (G), assuming that the perceived color (g) of one area in the system, p1, is dependent upon the actual colors of

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

206  unity

7.4  Ontological Dependence It is tempting to turn to ontological dependence relations as plausible candidates for substitution into Rescher and Oppenheim’s second criterion. As noted in Chapters 5 and 6, a plethora of relations have been defined in the literature under the heading of “ontological dependence.” I restate here a representative sample of the definitions of ontological dependence discussed earlier: (ND1)  Rigid Existential Necessary Dependence: x is rigidly existentially necessarily dependent on y ≡def Necessarily x exists only if y exists. (ND2)  Generic Existential Necessary Dependence: x is generically existentially necessarily dependent on Fs ≡def Necessarily, x exists only if some Fs exist. (ED3)  Essential Identity Dependence: x is essentially identity dependent on y ≡def There is some function φ such that it is part of the essence of x that x = φ(y). (ED4)  Constitutive Essential Dependence: x is constitutively essentially dependent on y ≡def y is a constituent of x’s essence (narrowly construed). (ED5)  Constitutive Definitional Dependence: x is constitutively definitionally dependent on y ≡def y is a constituent of a real definition of x. In effect, (ND1)–(ED5) narrow down the class of “attributes” relative to which (II) is to be evaluated to just the existence, identity, essence, or real definition of the parts of a given whole. The parts of such a whole are thus unified under ontological interdependence, in the sense defined by (ND1)–(ED5), just in case these parts depend for the “value” they take with respect to their existence, identity, essence, or real ­definition on the “value” taken by the other parts of the same whole with respect to the same “attributes.” Moreover, if it were in fact the case that the parts of an integrated whole are unified under ontological interdependence in any of the ways described earlier and this result furthermore generalizes across all composite entities which meet the first and third criteria, then the relationship in question would be lawful as well, as required by the second criterion. some or all of the remaining areas, p2, . . ., pn, in the system. Here, if the relationship identified in Definition 5 were to be stated in the form of a mereological supervenience thesis, the supervenience base would include the distribution of G-values (possibly including g) over p2, . . ., pn (excluding p1), while the supervenient domain would consist of p1’s g-value. Unless otherwise noted, when I speak of the parts of an integrated whole as forming a system that is unified under lawful interdependence, I have in mind a φ-dependence system as defined in Definition 5.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Ontological Dependence  207 To see whether (ND1)–(ED5) yield a workable interpretation of Rescher and Oppenheim’s second criterion, we must determine whether it is plausible to think of the parts of paradigmatic matter–form compounds as being unified under ontological interdependence, according to any of the definitions of ontological dependence just cited. For the sake of concreteness, consider a particular H2O molecule whose material components are its two hydrogen atoms and its oxygen atom. (I will refer to these atoms as “H1,” “H2,” and “O,” respectively.) Do H1, H2, and O ontologically depend on each other for their existence, identity, essence, or real definition in one or more of the ways defined by (ND1)–(ED5)? The answer unfortunately is negative, as can be illustrated by the following applications of (ND1)–(ED5) to two of the molecule’s material components, viz., O and H1: (A) O is not rigidly existentially necessarily dependent on H1, since it is not the case that necessarily O exists only if H1 exist. (B) O is not generically existentially necessarily dependent on some hydrogen atoms, since it is not the case that necessarily O exists only if some hydrogen atoms exist. (C) O is not essentially identity-dependent on H1, since it is not the case that there is some function φ such that it is part of the essence of O that O = φ(H1). (D) O is not constitutively essentially dependent on H1, since it is not the case that H1 is a constituent of O’s essence (narrowly construed). (E) O is not constitutively definitionally dependent on H1, since it is not the case that H1 is a constituent of a real definition of O. As (A)–(E) bring out, the material components of an H2O molecule cannot be taken to be unified under ontological interdependence, according to any of the definitions of ontological dependence just listed . The existence of any particular oxygen atom does not require the existence of any particular hydrogen atom; nor does the existence of any particular oxygen atom require the existence of hydrogen atoms in general. Moreover, it is not plausible to think that a particular hydrogen atom (or hydrogen atoms in general) would figure in a criterion of identity for a particular oxygen atom; nor that a particular hydrogen atom (or hydrogen atoms in general) would figure in the essence or real definition of any particular oxygen atom (or of oxygen atoms in general). (A)–(E) suffice to establish that substituting an ontological dependence relation into Rescher and Oppenheim’s second criterion does not provide a plausible interpretation of (II), even on the weakest reading of (II), viz., (IIWeak4), which requires only that every member of some division, D, of the whole in question bear some dependence relation or other to some other member of D. The particular case just considered constitutes a counterexample to this claim, since O does not bear any of the dependence relations cited above to either of the molecule’s other two material components. (The choice of H1, relative to O, was arbitrary; hence, the same results generalize to H2 as well as to other cases.) We must therefore look elsewhere to find a viable alternative

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

208  unity interpretation of Rescher and Oppenheim’s second criterion which accounts for the unity of integrated wholes.18, 19, 20

7.5  Interactional Dependence Let’s take stock. So far, we have established two negative conclusions, viz., that both of the attempts at providing a substantive account of unity considered earlier are unsuccessful. The first, put forward by Rescher and Oppenheim, was found to be too weak, since all physical systems are unified under lawful functional interdependence with 18   The weak and strong readings formulated earlier, which disambiguate the second criterion, are directly applicable only when the dependence relation(s) at issue is (are) rigid, rather than generic, since rigid dependence relations take particular entities as relata. However, modified versions of the weak and strong readings may be formulated for generic dependence as well. Such readings omit the variations in the quantifiers as issue and instead focus only on the remaining two contrasts brought out above: firstly, whether the dependence relations are permitted to vary when considering different members of a single system (viz., as they are in the weak, but not the strong, readings); and secondly, whether the dependence relation or relations in question manage to differentiate the members of a single system from surrounding entities (viz., as required by the odd-numbered, but not the even-numbered, readings). I will return to the question below of how Rescher and Oppenheim’s second criterion is best disambiguated when generic dependence relations are considered for substitution into (II). 19   Simons (1994) argues that the parts of integrated wholes are unified under ontological interdependence, in accordance with (ND1) and (ND2), a view that is also to be found in Husserl and Ingarden. To reach the result that the parts of an integrated whole are unified under rigid or generic existential necessary interdependence, Simons adopts the following (controversial) assumptions: concrete particular objects are bundles of tropes; these trope bundles can be divided into a “nucleus” (consisting of the object’s essential tropes) and a “periphery” (consisting of the object’s accidental tropes); the tropes in the nucleus are unified under rigid existential necessary interdependence; the tropes in the periphery are tied to the rest of the system through generic existential necessary interdependence. Simons’ proposal in effect yields a version of the weakest reading of the second criterion, viz., (IIWeak4), according to which every member of D bears some dependence relation or other to some other member of D, with (ND1) and (ND2) specifying the range of dependence relations in question. Such a system does not satisfy the stronger readings of the second criterion, since (firstly) more than one dependence relation is required to capture the relationships in question, (secondly) the members of one system only bear the dependence relation(s) to some but not all other members of the system, and (thirdly) the members of this system also bear these dependence relations to surrounding entities. This is easy to see in the case of (ND2); but certain essentialist theses (e.g., the essentiality of origins) lead to the same result in the case of (ND1) as well. Whatever its other merits, Simons’ proposal cannot be used to establish that an H2O molecule’s material components are unified under ontological interdependence: for each of these concrete particular objects is already an integrated whole, and Simons’ proposal does not speak to the question of how these individual integrated wholes hang together and form a further integrated whole when they are assembled in the form of an H2O molecule. In addition, I also take issue with some of the other trope-theoretic assumptions on which Simons’ conception of the unity of integrated wholes relies. 20   It is possible to reach the desired outcome (that the material components of an H2O molecule are unified under ontological interdependence in one of the ways indicated earlier) by taking on board Aristotle’s so-called “Homonymy Principle,” according to which the parts of a matter–form compound cannot survive being separated from the whole whose parts they are. As we saw in Chapter 4, in the context of our discussion of hylomorphic monism, a version of this position is endorsed by some contemporary hylomorphists (see, for example, Jaworski (2016b); Marmodoro (2013); Oderberg (2007); Toner (2006, 2008, 2010, 2011a, 2011b,  2011c, 2013)). According to my own hylomorphic conception of matter (see Chapter 2), however, the material parts of a hylomorphic compound can survive being separated from the whole of which they are part; thus, reading Rescher and Oppenheim’s second criterion with an ontological dependence relation in mind is not an option for me, given my other commitments.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Interactional Dependence  209 respect to their contingent physical attributes and relative to the laws of physics. The second proposal restricts the first by taking the parts of an integrated whole to be unified under lawful ontological interdependence with respect to their existence, identity, essence, or real definition and relative to the “laws of metaphysics.” This account, in contrast, turned out to be overly strong, since even the parts of highly unified structured wholes do not in general ontologically depend on other parts of the same whole with respect to their existence, identity, essence, or real definition. What is needed, therefore, is an interpretation of Rescher and Oppenheim’s second criterion that is intermediate in strength, i.e., one that is situated between those offered by the two rival accounts considered earlier. In what follows, I will attempt to develop such a proposal by focusing not on the contingent physical attributes or on the existence, identity, essence, or real definition of the parts of an integrated whole, but rather on their capacities to interact with one another in the particular mereological setting in which they find themselves as parts of an integrated whole.21 I will refer to the members of a system in which the conditions specified below are met as unified under lawful interactional interdependence. To motivate the account I offer in what follows, let’s first consider an example: an eye that is part of a living organism which is capable of engaging in visual perception. While the eye is part of the living organism, it is able to contribute to the living organism’s ability to see by manifesting certain of its capacities. But the eye cannot accomplish this alone. Rather, in order for the eye to make its contribution to the living organism’s visual system, the eye’s activities must be coordinated in the right sort of way with the activities carried out by other components of the organism’s visual system (e.g., the brain). This sort of calibration between what the eye does and what the other components of the organism’s visual system do requires that the eye is connected in the right sort of way to these other components of the organism’s visual system. But mere spatiotemporal connection between, say, the eye and the brain is not enough to set the stage for the sort of interaction at issue. The eye’s ability to interact with the brain in the right sort of way also requires that the eye and the brain (along with the other components of the organism’s visual system) act and interact in certain specific and complex ways whose precise characterization is best left to scientists rather than metaphysicians. In this way, both the eye and the brain are integrated into the organism by playing a vital role in its visual system. If the eye were to be separated from the living organism, it would not thereby automatically cease to exist (contra Aristotle’s Homonymy Principle). For it is plausible to think, given modern medicine, that under 21   I intend my use of the term, “capacity,” to be understood broadly in such a way as to encompass what is also sometimes referred to as “power,” “potentiality,” “potency,” “propensity,” “disposition,” and the like. In the literature on dispositional essentialism in the philosophy of science, the ascription of a disposition to an object is often understood roughly as follows (see, for example, Bird (2007), p. 3): an object which has a dispositional property (e.g., being soluble in water) is such that it would give rise to some characteristic manifestation (e.g., being dissolved in water) in response to a certain kind of stimulus (e.g., being exposed to water). For present purposes, I do not mean to commit myself to any specific way of understanding the properties in question. For a recent account of potentiality, see Vetter (2015).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

210  unity the right circumstances this very eye could for example be preserved or transplanted into another living organism (or even potentially into an artificial system) where the very same eye can then take up its activities again, now as a contributor to the visual system of its new “host.” But an eye which exists in a state separated from any such suitable surroundings is not able to manifest those of its capacities by means of which it can contribute to a living organism’s ability to engage in visual perception. In turn, a living organism which is missing one of its eyes is not able to manifest its capacities for visual perception to the same extent as it would if the missing eye were present. Generalizing from this example, we may informally characterize the interactional dependence relation required for the new interpretation of the second criterion as follows: an integrated whole derives its unity from the way in which its parts are able to interact with other parts of the same whole; the interplay between these activities carried out by the parts of an integral whole in turn allows the whole as well to manifest certain of its capacities, namely, those whose manifestation by the whole requires “team work” among its parts. I will refer to these capacities of the whole as “TWR-capacities” (short for “‘team work’ requiring capacities”). The relevant capacities of the parts are those by means of which the parts of an integrated whole contribute to the whole’s ability to manifest its “team work” requiring capacities. I will refer to these capacities of the parts as their “contributing capacities” (short for “capacities by means of which a part contributes to the whole’s manifestation of its TWR-capacities”). This informal characterization of the unity of integrated wholes highlights the way in which Rescher and Oppenheim’s three criteria should be viewed as building on one another: (III) sets minimal structural requirements governing all wholes, whether they are integrated or not, while the unification of the parts of an integrated whole under (II) creates the right environment for certain successful instances of (I). In particular, the permissive reading of (II) witnessed earlier, according to which every physical system counts as unified under lawful functional interdependence, is now restricted by enforcing a connection between the whole’s possession of certain underivable attributes, relevant to the satisfaction of (I), and the attributes of the parts that are at play in the satisfaction of (II). The former is achieved by way of the whole’s manifestation of certain of its TWR-capacities, while the latter is due to the parts’ manifestation of certain of their contributing capacities. To illustrate, a living organism which possesses the underivable modal attribute of being capable of engaging in visual perception (thereby satisfying (I)) is able to manifest this capacity by virtue of the complex interactions among certain of its parts (in compliance with (II)) which are configured in the right way and are of the right type to play a role in the organism’s visual system (in accordance with (III)). Our next task is to offer a more precise formulation of the second criterion, under the reading proposed, followed by an evaluation of how well this proposal performs with respect to our familiar test cases. Relative to the disambiguations offered previously, we must clarify, firstly, whether the interactional dependence relation needed for the proposed reinterpretation of the second criterion is rigid or generic; and, secondly,

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Interactional Dependence  211 where this relation fits with respect to the weak and strong readings distinguished earlier. To bring out, firstly, why the dependence relation in question need only be ­generic and not rigid, let’s return once more to the specific case of visual perception considered earlier. To illustrate, if an eye, which is part of a living organism, were to stand in a rigid interactional dependence relation to the organism’s brain, then the eye would require the very brain, with which it is, in fact, interacting, to be present and available for interaction, in order for the eye to be able to manifest those of its capacities by means of which it contributes to the organism’s ability to engage in visual perception. In fact, however, the interactional dependence which obtains between the eye and the brain is only generic, insofar as the eye only requires some suitable brain or other (or some suitable analogue of a brain) with the right capacities to be present and available for interaction, in order for the eye to contribute in its respective ways to the organism’s ability to engage in visual perception (together with the other components of the organism’s visual system). We can extrapolate from this specific case that the manifestation of the relevant contributing capacities by the parts of an integrated whole in general requires only that some part or other of a certain kind with suitable contributing capacities be present and available for interaction, but not that any particular member of the kind in question be present and available for interaction within the same whole. Secondly, by examining the specific case at hand, we can determine further which of the odd- or even-numbered strong or weak readings distinguished earlier provides the best overall fit for present purposes. Recall that the strong readings differ from the weak readings, in that the strong readings (IIStrong1)–(IIStrong4), require that there be a single dependence relation, R, which relates the members of a division, D, to one another, while the weak readings (IIWeak1)–(IIWeak4), allow there to be some variation in the dependence relations which members of D bear to other members of D. Moreover, the odd-numbered readings differ from the even-numbered readings in that the oddnumbered readings require that the dependence relation(s) at issue must also differentiate the members of D from surrounding entities, while the even-numbered readings allow that the members of D bear the dependence relation(s) at issue to surrounding entities as well. Finally, the third and fourth readings weaken the first and second readings, in that the third and fourth readings require only that the members of D bear the dependence relation(s) at issue to some, but not all, members of D, while the first and second readings require that the members of D bear the dependence relation(s) at issue to all members of D. Since our reformulation of the second criterion makes use of a single dependence relation, viz., generic interactional dependence, we can exclude the weak readings and focus instead on the strong readings. Generic interactional dependence may be defined as follows: (GID)  Generic Interactional Dependence: x, which belongs to kind K, generically interactionally depends on some member of kind K’ ≡def (i) x, which belongs to kind K, has some capacities,

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

212  unity CK1, . . . , CKn, and members of kind K’ have some capacities, CK’1, . . . , CK’m, and (ii) the manifestation of CK1, . . . , CKn by x requires the manifestation of CK’1, . . . , CK’m by some member of kind K’. Among the strong readings, the first and second (IIStrong1)–(IIStrong2), are overly strong in that they require each of the parts of an integrated whole to stand in a generic interactional dependence relation to every other part of the same whole. But an eye, for example, need not interact with the organism’s left toe, say, or even with all the other components of the organism’s visual system, in order for it to manifest those of its capacities by means of which it contributes to the organism’s ability to engage in visual perception. The third strong reading (IIStrong3), requires that the parts of an integrated whole interact in a certain distinctive way with other parts of the same whole in which they do not also interact with surrounding entities which are not integrated into the same whole as parts, to give rise to successful manifestations of the whole’s TWRcapacities. The applicability of this reading to the case at hand, while perhaps ultimately defensible, requires some method of distinguishing, for example, between the eye’s two-way interactions with the organism’s brain on the one hand, and the eye’s one-way interactions with features in the organism’s environment (e.g., objects in the organism’s visual field) on the other hand. For the time being, I will therefore make use only of the fourth strong reading (IIStrong4), reformulated here as (II’StrongGID4), for the purposes of restating Rescher and Oppenheim’s second criterion: (II’StrongGID4) Every member of D is generically interactionally dependent upon some member of D. With these clarifications in place, we may reformulate Rescher and Oppenheim’s second criterion as follows, where (II’) is to be read with (II’StrongGID4) in mind and the relevant class of attributes possessed by the parts of an integrated whole is restricted here to their contributing capacities, i.e., to those capacities by means of which the parts of an integrated whole contribute to the whole’s manifestation of its TWR-capacities: (II’)  The Revised Second Criterion: The parts, p1, p2, . . . , pn, of an integrated whole are lawfully generically interactionally interdependent with respect to the manifestation of their contributing capacities, C, i.e., (II’.1) the value taken by any one part (e.g., p1) with respect to the manifestation of some contributing capacity, c, is generically interactionally dependent upon the value taken by (some or all of) the remaining parts (viz., p2, . . . , pn) of the same whole with respect to the manifestation of (some or all of) their contributing capacities, C; and (II’.2) the relationship described in (II’.1) is governed by general laws or law-like connections. When (II’1.) is read with (II’StrongGID4) in mind, what holds for any one part of an integrated whole must hold for every part of the whole in question: that is, every part of an integrated whole must be generically interactionally dependent for the value it takes

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Interactional Dependence  213 with respect to the manifestation of some of its contributing capacities on the value taken by at least some of the parts of the same whole with respect to their manifestation of some of their contributing capacities. Moreover, given the lawfulness requirement in (II’.2), the relationship between the values taken by the parts of an integrated whole with respect to the manifestation of their contributing capacities must generalize to other similar systems which are governed by the same laws or law-like connections. Returning now to our familiar test cases, we may ask to what extent composite social entities belonging to artifactual kinds satisfy the revised second criterion under the proposed reading. To what extent, for example, are the material components of a screwdriver unified under lawful generic interactional interdependence? To evaluate whether, and to what extent, (II’) holds in this case, let’s consider a particular TWRcapacity, CS, e.g., the capacity of a Phillips screwdriver to tighten or loosen a Phillips screw, whose manifestation by the screwdriver requires the manifestation of certain contributing capacities by its material components, viz., the screwdriver’s tip, shaft, and handle. The satisfaction of (II’.1) read with (II’StrongGID4) in mind, requires that the value taken by each material component with respect to its capacity to contribute to the screwdriver’s manifestation of CS must generically interactionally depend on the value taken by some (though not necessarily all) of the other material components with respect to their manifestation of some capacity by means of which these other material components contribute in their respective ways to the screwdriver’s ability to manifest CS. Since the dependence relation in question is generic, and not rigid, every material component need only be related in this way to some material component of the screwdriver or other, but not necessarily to any particular material component. Further, the satisfaction of (II’.2) requires that the generic interactional interdependence among the screwdriver’s material components generalizes to other similar systems governed by similar laws or law-like connections. (II’.1) does appear to be satisfied in this case. To illustrate, consider the tip’s contributing capacity, CT, to fit into the head of a Phillips screw. While the tip is part of the screwdriver, it is able to contribute to the manifestation of CS by the screwdriver by manifesting its contributing capacity, CT. But the tip, by manifesting CT, is only able to contribute to the manifestation of CS by the screwdriver, if the tip’s manifestation of CT is properly calibrated with the manifestation of contributing capacities by the screwdriver’s other material components, e.g., the shaft’s ability to maintain its physical integrity even when a force within a certain range is applied to the screwdriver. It is, moreover, straightforward to see that the interactional dependencies which connect a screwdriver’s material components are only generic, and not in general rigid: switching out one handle for another, for example, need not hinder, and may in fact improve, the tip’s ability to contribute, by manifesting CT, to the manifestation of CS by the screwdriver. Finally, (II’StrongGID4) allows that the screwdriver’s material components must interact in particular ways not only with one another, but also with surrounding entities, e.g., a Phillips screw with the right sort of shape to be tightened or loosened by a Phillips screwdriver. In this respect, (II’StrongGID4) contrasts with the third strong reading

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

214  unity (IIStrong3), which requires that the parts of an integrated whole bear the dependence relation in question only to other parts of the same integrated whole, but not also to other entities which are not themselves parts of the integrated whole in question. An interesting contrast does emerge, however, between the case presently under discussion and the cases previously considered when we turn to the lawfulness requirement in (II’.2): for the domain of artifacts is not governed by laws or law-like connections in the same way, and to the same extent, as the domain of natural things. In the particular case at hand, for example, the hypothesis that each material component of a screwdriver must interact in particular ways with the screwdriver’s other material components in the ways outlined here requires taking as fixed, among other things, the intentions of agents (e.g., inventors, designers, producers, or users) to engage in particular types of actions (viz., to tighten or loosen screws) contributed not by the screwdriver itself, or by its material components, but by factors external to the screwdriver itself and its material components. The assumption that an agent intends to put to use the artifact in question in order to engage in a certain type of action, in turn, gives rise to certain principles as to how a screwdriver’s material components ought to interact with one another if the agent’s practical goal is to be achieved. But such generalizations can be formulated only normatively and conditionally, on the assumption that the artifact in question functions as it ought to and is employed by a competent user in order to perform a particular type of action. These conditional normative principles by themselves do not dictate how, other things being equal, the components of a system must interact under certain conditions irrespective of the mental states of agents that are external to the system in question. Generalizing from the particular case at hand, then, composite social entities belonging to artifactual kinds do not appear to satisfy (II’.2) in the same way, or to the same extent, as paradigmatic matter–form compounds belonging to natural (e.g., physical, chemical, or biological) kinds. For the satisfaction of (II’.1) by the material components of an artifact in a particular case gives rise to a generic interdependence relationship which generalizes across similar systems only if the notion of lawfulness operative in (II’.2) is extended to include conditional normative principles governing the members of social kinds, in addition to the laws or law-like connections governing natural kinds.

7.6 Conclusion In sum, we may conclude, then, that Rescher and Oppenheim’s three criteria do yield a promising account of the unity of integrated wholes, provided that we interpret the second criterion in a certain way. What unifies the parts of an integrated whole is not that they functionally depend on each other for the values they take with respect to their contingent physical attributes (since all physical systems meet this condition). Nor can the unity of integrated wholes be said to be traced to the fact that their parts ontologically depend on each other for their existence, identity, essence, or real definition (since this requirement is not generally satisfied even by highly unified structured

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Conclusion  215 wholes). Instead, integrated wholes, so I have argued, derive their unity from the way in which their parts lawfully depend on each other for the manifestation of certain of their capacities, viz., those capacities by means of which the parts of an integrated whole contribute to the whole’s ability to manifest those of its capacities which require “team work” among the parts. An interesting consequence of the account of unity provided here is that it can also explain why highly unified wholes are capable of persisting through changes with respect to their parts and characteristics over time. For notice that we have only required that the parts of an integrated whole are generically, but not necessarily rigidly, interactionally dependent on other parts of the same whole. As long as an eye, for example, is able to interact with some brain or other (or some suitable analogue of a brain) with the right capacities, the eye can make its respective contribution to the organism’s ability to engage in visual perception, assuming that the other components of the organism’s visual system with which the eye must cooperate are also operating smoothly. The eye’s ability, as well as the capacity of the organism as a whole, to engage in these carefully calibrated activities therefore does not turn on the numerical identity of the other components of the organism’s visual system with which the eye is interacting, but only on there being some components or other of the right sort present and available for interaction. For this reason, switching out one suitable component for another within the organism’s visual system does not affect the overall viability of its perceptual apparatus. In the course of our discussion of unity in this chapter, some notable differences have emerged between paradigmatic matter–form compounds belonging to natural (e.g., physical, chemical, or biological) kinds and composite social entities, in particular artifacts. In the case of natural kinds, generic interactional interdependencies among the parts of an integrated whole generalize across similar systems which are governed by the same set of laws or law-like connections. In the case of artifacts, in contrast, we noted that the generic interactional interdependencies among the components of a system are at best governed by conditional normative principles which state how the components of these systems ought to interact, assuming that they are to contribute to the performance of certain actions in which agents external to the system in question intend to engage. We will continue our discussion of artifacts in Chapter 8 by considering the challenges posed by this particular domain for existing essentialist as well as antiessentialist frameworks.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

8 Artifacts 8.1  Introductory Remarks Artifacts have come up several times during the course of this study. Firstly, we ­considered artifacts in Chapter  3, in the context of discussing the assignment of forms to their proper ontological category. In this context, I brought up the account developed by Evnine (2016a) who proposes a version of origin-essentialism, according to which it is essential to an artifact that it is the product of a certain creative act (see Section 3.2). In this chapter, we will, among other things, take up the question of whether this kind of origin-essentialism is sufficiently fine-grained to settle questions concerning the numerical identity of artifacts of the sort discussed in Section 3.4.3. Secondly, artifacts became a subject of discussion again in Chapter  6, when the question arose as to whether forms are intrinsic to the matter–form compounds with which they are associated (see Section 6.4.4.4). In this connection, artifacts pose the following challenge: if an artifact is essentially the product of a creative act in which an agent successfully manifests his or her creative intentions, then it seems that the essences of artifacts include factors that are extrinsic to them, at least on the seemingly plausible assumption that an artifact’s maker and the maker’s intentions are extrinsic to the artifact that is produced in the relevant creative act. Given the close connection between forms and essences (see Section 3.4.2), this position would call into question the thesis that forms are intrinsic to the matter–form compounds with which they are associated, at least on the assumption that artifacts are accepted as genuine matter– form compounds. Thirdly, I considered artifacts as a potential test case for the account of unity ­proposed in Chapter 7. In this connection, I noted that artifacts do not seem to be governed by laws or law-like connections in the same way or to the same extent as paradigmatic natural matter–form compounds, e.g., those belonging to physical, chemical, or biological kinds (see Section 7.5). Rather, in the case of artifacts, the principles governing the interactions among their material components seem to be of a normative conditional nature and require us to make reference to the mental states of agents who intend to make use of the artifacts in question in order to engage in certain types of actions. More generally, given their apparent mind dependence, artifacts pose challenges to any realist account of the metaphysics of concrete particular objects, whether hylomorphic or not. Structure-based hylomorphic accounts, in particular, must

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

artifact kinds and natural kinds  217 wrestle with the question of whether artifacts should be assigned the same status within their ontologies as paradigmatic natural matter–form compounds, given that artifacts appear to be just as highly structured as the members of paradigmatic natural kinds. In what follows, I will not attempt to develop conclusive answers to all of these important questions concerning the status of artifacts in a realist hylomorphic ontology, since such an endeavor would require a book-length treatment in its own right. Rather, I will point to what I take to be interesting directions for future research which call for a more detailed discussion than what I can provide in the present context. In particular, in what follows, we will consider several prominent essentialist accounts of artifacts, all of which trace the essences of artifacts, directly or indirectly, to the intentions of an artifact’s original author (e.g., its inventor, maker, producer, or designer). We will discover that these “author-intention-based” accounts are subject to the concern that, when stated in its most general form, human creative intentions are not nearly as powerful and discriminating as proponents of these accounts make them out to be. This result might seem to indicate that artifacts might be more naturally accommodated within the confines of an anti-essentialist account of some sort: in many cases, after all, artifacts are created by human beings to serve human interests. Perhaps surprisingly, however, we will find that existing anti-essentialist frameworks are also, in their current formulations, not able to account for the special characteristics of artifacts as straightforwardly as we might have expected them to do so. For these approaches tend to trace our attribution of modal features to objects to our semantic, inferential, or explanatory practices; but our engagement with the realm of artifacts seems to be primarily practical and action-based. Thus, as it stands, the domain of artifacts leaves us with some open questions which call for the further development and refinement of essentialist and anti-essentialist frameworks alike.

8.2  Artifact Kinds and Natural Kinds There is no uncontroversial way of distinguishing between natural things and artifacts, or between natural kinds and artifact kinds. Bird and Tobin (2012) characterize a natural kind as “a grouping that reflects the structure of the natural world rather than the interests and actions of human beings.” Since human beings are themselves part of nature, however, it is at least not immediately obvious why divisions of things into kinds which reflect “the interests and actions of human beings” could not themselves also yield divisions of things into kinds which reflect “the structure of the natural world.” Alternatively, a kind may also be construed as natural, not because it is found in nature, but because it presents us with a grouping of particular items into a category or  taxonomic classification that is, in some sense, not arbitrary, heterogeneous, or gerrymandered.1 1   I adopt the latter conception of natural kinds and discuss the question of how a division of things into natural kinds may be motivated in Koslicki (2008a), Chapter 8, and Koslicki (2008b).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

218  Artifacts It is sometimes said that artifacts (or artifact kinds) can be distinguished from natural things (or natural kinds) by considering the reasons or processes that are responsible for bringing them into existence. Thus, according to Hilpinen (2011), “[a]n artifact may be defined as an object that has been intentionally made or produced for a certain purpose.” Although Hilpinen’s characterization of artifacts no doubt initially appears very plausible, any attempt to delineate artifacts and natural things on the basis of the sorts of considerations suggested by Hilpinen is immediately confronted with a wide range of tricky cases.2 (1)  Artificially Produced Members of Natural Kinds. Consider, for example, seedless grapes or synthetically produced chemical compounds (i.e., chemical compounds which are created and perhaps can only be created in the lab). Cases such as these seem to present us with objects which we might otherwise assign to natural (e.g., biological or chemical) kinds and yet they appear to be “intentionally made or produced for a certain purpose” by human beings. Domestic animals, such as dogs, furthermore raise the question of what sorts of processes should count as instances of intentionally “making” or “producing” an object for a certain purpose. For while the mechanisms by which dogs are “produced” from other dogs (viz., by sexual reproduction) appear to be straightforwardly natural, we must also note that dogs would not have evolved from wolves in the first place, if it had not been for human interference. At the same time, we would seem to be casting our net too widely, if objects were counted as artifacts on the basis of whether they would not exist, were it not for our contribution. At this point, almost everything in our environment in some way bears the marks of intentional purpose-driven human interference, and this broader method of classification would thus, apparently incorrectly, classify many objects as artifacts which would otherwise be assigned to natural (e.g., physical, chemical, or biological) kinds. (2)  “Ready-Mades” and “Found Objects.” The latter question of what conditions a process must satisfy in order for it to qualify as an intentional “making” or “producing” is also brought to the fore, though in a different way, by cases involving so-called “Ready-Mades” or “Found Objects.” Did Marcel Duchamp, for example, succeed in creating a new artifact, viz., a sculpture called Fountain, from a previously existing artifact, viz., a urinal? Or is it possible to create a new artifact simply by “converting” a piece of driftwood into a bench? If so, then apparently a process can qualify as an intentional making or producing, even if no physical work is required on the part of the agent responsible for the apparent creation of a new artifact and the underlying material from which the new artifact is apparently made (viz., the urinal or the piece of driftwood) need not undergo any intrinsic change.

2   The examples cited in the remainder of this section are not original to me, but familiar and widely discussed in the literature on artifacts (see, for example, Evnine (2016a), Margolis and Laurence (2007) and the references therein).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

artifact kinds and natural kinds  219 (3)  Byproducts, Residue, and Other Unintended Outcomes. The intentional production of an object for a certain purpose can sometimes result in unintended byproducts or residue, e.g., pollution, sawdust, or scrap metal. In addition, an entity can also apparently come about as the result of unintentional human activity (e.g., a drawing that is the result of mere “doodling”) or as an unintended consequence of intentional human activity, but not as the direct byproduct or residue of a process in which an object is “intentionally made or produced for a certain purpose.” For example, a village may be the unintended result of a collection of houses that are put up in close vicinity to one another; or a trail may be the unintended result of many walkers reaching their destination via the same route. Should these unintended outcomes of human activity themselves be classified as artifacts? If they are so classified, then apparently artifacts need not always themselves be the direct target of an intentional purpose-driven creative activity. (4)  Products of Non-human Intentional Activity. Some objects (e.g., beaver dams, ant hills, spider webs, or birds’ nests) appear to be “intentionally made or produced for a certain purpose,” but the agents responsible for their production are non-human animals. If these objects are included in the category of artifacts as well, then the ability to produce artifacts is not confined to our own species. Given this broader categorization, one wonders how cognitively sophisticated an agent must be in order to be able to partake in an intentional purpose-driven process that results in the creation of an artifact. (5)  Artworks. According to Hilpinen’s characterization, artifacts are objects which are “intentionally made or produced for a certain purpose.” At least in some cases, the purpose for which an artifact is intentionally made or produced appears to be intimately connected to the object’s intended function or use. Thus, it is natural to think that screwdrivers are, in a typical case, intentionally made or produced for the purpose of creating something which can be used to tighten or loosen screws, given that their intended function or use is to tighten or loosen screws. Although artworks are also often regarded as a special class of artifacts, it is at least not obvious that artworks have an intended function or use or, if they do, what their intended function or use would be. The case of artworks thus puts some pressure on the characterization of artifacts as the products of an intentional purpose-driven creative activity. In the preceding paragraphs, we have encountered cases which pose various sorts of challenges to Hilpinen’s intuitively plausible characterization of artifacts as objects that are “intentionally made or produced for a certain purpose.” (1) In some cases (e.g., seedless grapes, synthetically produced chemical compounds, or domestic animals), it appears to be possible that objects are “intentionally made or produced for a certain purpose,” even though they appear to be members of natural (e.g., chemical or biological) kinds. (2) In some cases (e.g., “Ready-Mades” or “Found Objects”), the apparent creative process in question hardly seems to qualify as a “making” or “producing” at all, given that the agent need not physically work on, or intrinsically change, the underlying material from which the new artifact is apparently created. (3) In some

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

220  Artifacts cases, human activity can leave in its wake various unintended byproducts, residue, or other outcomes (e.g., pollution, sawdust, or scrap metal; drawings that are the result of  mere “doodling”; or unintentionally created villages or trails) which, if they are themselves regarded as artifacts, do not straightforwardly fit the characterization of being “intentionally made or produced for a certain purpose.” (4) In some cases (e.g., beaver dams, ant hills, spider webs, or birds’ nests), objects can apparently be the products of intentional purpose-driven activity carried out by agents whose degree of cognitive sophistication is quite far removed from that manifested by humans. (5) In some cases (e.g., artworks), objects which do not obviously have an intended function or use raise the question of how, or whether, the intentional human activity involved in their production can nevertheless be regarded as purpose-driven. These five groups of cases present us with only a small selection of the many difficult cases confronting any attempt to offer a general characterization of what it takes to be an artifact. In what follows, we shall encounter many more troublesome phenomena which pose additional challenges to attempts at delineating artifacts from natural things by means of considerations such as those suggested by Hilpinen.

8.3 Artifact-Essences 8.3.1  Maker’s Intentions According to the account of artifacts developed by Amie Thomasson (see Thomasson (2003a, 2003b, 2007, 2009)), it is essential to artifacts that they are the intended products of human activity. As we noted in Section 8.2, human activity can also result in various unintended byproducts, residue, or other outcomes (e.g., pollution, sawdust, or scrap metal; drawings that are the result of mere “doodling”; or unintentionally created villages or trails). Given Thomasson’s use of the term “artifact,” these unintended products of human activity do not count as artifacts.3 Thomasson also restricts the application of the term to the intended products of human activity (as opposed to, say, activities engaged in by non-human animals), so that, for example, beaver dams, ant hills, spider webs, birds’ nests, and the like, are also not classified by her account as artifacts, even if such items might plausibly be regarded as the intended products of intentional non-human activity. 3   What is initially thought of as a mere unintended byproduct or residue could potentially acquire a different status and come to be something that is itself “intentionally made or produced for a certain ­purpose.” The possibility of such transitions would seem to spell trouble for any account, such as Thomasson’s, which is committed to the following combination of claims: (i) anything that is in fact an artifact is essentially an artifact and essentially belongs to some artifact kind; and (ii) the unintended byproducts or residue of the intentional purpose-driven production of artifacts are not themselves classified as artifacts and hence do not belong to any artifact kind. For, given (i) and (ii), either all or none of the members of the kind in question are classified as artifacts; and the very same kind cannot at some point include members which are not classified as artifacts and at a later point include members which are classified as artifacts.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

artifact-essences  221 In Thomasson’s view, intentional human activity is not only causally responsible for the production of artifacts, but constitutive of the nature or essence of artifacts.4 Since artifacts, for Thomasson, are therefore mind-dependent, they differ in this respect ontologically from the members of natural kinds whose existence and essence she regards as not constitutively dependent on intentional human activity. Nevertheless, despite the mind-dependence of artifacts, Thomasson cautions us not to conclude that these entities are for that reason mere mental constructs or any less real than members of natural kinds. Correspondingly, she stresses that, in our discourse about artifacts, we are not just engaged in pretense or make-believe. More specifically, an artifact of kind K, according to Thomasson, is the product of a largely successful intention to create something of kind K (see Thomasson (2003b), pp. 592–602). A maker has the relevant intention (viz., a largely successful intention to create something of kind K) if and only if she has a substantive concept of the nature of Ks which largely matches that of prior makers of Ks (if any) and she intends to realize that concept by imposing K-relevant features. The K-relevant features in question can be functional, structural, historical, aesthetic, or of various other sorts or combinations. The parenthetical qualification, “(if any),” is added to accommodate scenarios in which an inventor produces a prototype of a new artifact kind K. In such cases, the artifact’s maker cannot intend her own substantive concept of what she takes herself to be producing to be largely continuous with the intentions of previous makers of instances of the same artifact kind, since (as far as she knows) there are no previous instances of kind K. When a maker invents a new kind of artifact, the features that are relevant to an item’s membership in the artifact kind in question are stipulatively determined by the concept guiding the inventor’s creative activity. In cases in which later makers produce subsequent instances of an already-established artifact kind K, the intentions governing their creative activities must largely, though not exactly, match those of previous makers. Since an inventor’s concept and a later maker’s concept of what they take themselves to be producing need not completely coincide, Thomasson’s account allows for a certain degree of vagueness and malleability in our artifact concepts. Given Thomasson’s account, the familiar scientifically realist Kripke–Putnam approach to natural kinds and natural kind terms does not straightforwardly transfer to the realm of artifact kinds and artifactual kind terms. Among the main ontological, epistemic, and semantic contrasts Thomasson draws between the domain of the natural and the domain of the artifactual are the following. Ontologically speaking, while the members of natural kinds are not generally taken to be existentially or essentially 4   Thomasson’s distinction between entities that are essentially the products of intentional human activity and entities that are merely accidentally caused by intentional human activity can be invoked in response to the first class of tricky cases cited above, viz., cases in which entities that appear to be members of already accepted natural kinds nevertheless satisfy Hilpinen’s characterization of being “intentionally made or produced for a certain purpose.” To illustrate, if it is merely accidental to a chemical compound, for example, that particular instances of it are produced in a lab, then members of the kind in question would not be classified as artifacts, according to Thomasson’s account, even if particular exemplars of the chemical compound in question are in fact the products of purpose-driven intentional human activity.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

222  Artifacts dependent on intentional human activity, Thomasson, as noted earlier, does regard the members of artifact kinds as being mind-dependent in a particular way, viz., by being essentially the intended products of human activity. Epistemically speaking, while Thomasson allows that there are some objective facts about artifacts (hence opening up the possibility for certain kinds of error and ignorance), she also argues that the artifact’s creator occupies an epistemically privileged position with respect to the question of what (if anything) she creates: since, for Thomasson, the creator’s intentions stipulatively determine which specific and general features are relevant to membership in the kind in question, these elements are therefore not open to revision and insulate the artifact’s creator from the possibility of massive error or ignorance concerning the product of her intentional activity. Finally, semantically speaking, Thomasson proposes that a purely causal theory of reference is not acceptable for artifact terms. Since the artifact’s creator grounds the reference of an artifact term, the features which are stipulatively specified by way of the creator’s intentions as being relevant to an object’s membership in the artifact kind at issue are therefore conceptually tied to the meaning and use of the artifact term in question. Hence, the proper semantics for artifact terms, in Thomasson’s view, must be at least in part descriptive.5

8.3.2  Creative Acts The account of artifacts defended by Simon Evnine (see especially Evnine (2016a)) appears, on the surface, to be quite similar to Thomasson’s: both regard artifacts as not just causally, but constitutively, mind-dependent entities whose essences must be understood at least in part in terms of the creative intentions of their makers. Artifacts, as Evnine likes to say, are the results of minds impressing themselves on matter with certain creative intentions (see, e.g., Evnine (2016a), p. 69). For Evnine, however, a maker’s creative intentions enter into an artifact-essence only indirectly through the act of creation which (if successful) produces the artifact in question: it is the act of creation, in Evnine’s view, which directly figures in the essence of the resulting artifact; the presence of the maker’s intentions in the essence of an artifact, in turn, is mediated by their presence in the essence of the corresponding act of creation.6 5   For more detail on how Thomasson conceives of the ontological, epistemic and semantic differences between the natural and artifactual domains, see especially Thomasson (2003b, 2007). 6   Evnine also takes his own more general approach to ontology to be less permissive, and more genuinely realist, than Thomasson’s neo-Carnapian “ontological minimalism” or “easy ontology” (see, e.g., Thomasson (2015)). The disagreement between Evnine and Thomasson, for the particular case of artifacts, ultimately comes down to Evnine’s more restrictive conception of what it takes for an agent to exercise, or fail to exercise, a genuinely “creative power” (see especially Evnine (2016a), pp. 116–17, for discussion). To determine whether Evnine in fact manages to implement the intended restrictions on what counts as a successful, or failed, attempt at exercising a genuinely creative power would require, among other things, a detailed examination of his general account of actions as artifactual events (Evnine (2016a), Ch. 7). In addition, Evnine has a great deal of interest to say about particular cases in which agents, in his view, exercise, or fail to exercise, different kinds of genuinely creative powers: such cases include, for example, the creation of natural organisms through sexual reproduction (Evnine (2016a), Ch. 5); cases of artistic creation in which artworks are brought into existence in such domains as music, literature, or the fine arts (Evnine (2016a), Sections 4.4.1–2); as well as cases of linguistic creation in which languages are brought into existence by speakers or

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

artifact-essences  223 Evnine thus endorses a certain kind of essentiality of origins thesis for artifacts, since he takes these entities to be essentially the products of certain creative acts performed by agents. Unlike Kripke, however, Evnine views the act of creation which produces the artifact, rather than the original matter from which an artifact is fashioned, as essential to the resulting artifact. Acts of creation, and actions in general, according to Evnine, are themselves artifacts: they belong to the category of artifactual events, rather than to the category of artifactual objects, which is reserved for the intended products of successful acts of creation. Like artifactual objects, Evnine holds that actions are themselves “made” by agents out of some underlying matter; but the matter underlying an action is either other actions or bodily movements (e.g., an agent “makes” an action of switch-flipping out of a certain deliberate movement of her finger). The essence of an action, so Evnine argues, should be construed as including the agent who performs the action, the intention with which the agent performs the action, the kind to which the action belongs (when the action is singled out under a canonical description as being performed intentionally), as well as the action’s matter, i.e., the underlying events (viz., bodily movements or other actions) which materially compose the action and by means of which the agent performs the action in question. In contrast, the time at which an act of creation occurs or the matter of the artifact which is produced in a successful act of creation, in Evnine’s view, are not essential to an act of creation. Thus, the very same act of creation could have occurred earlier or later than it did; and the very same artifactual object could have been created, in the same act of creation, from different matter. But the very same act of creation could not have been performed by a different agent, with different intentions, or by performing a distinct series of underlying bodily movements or actions; nor could the very same act of creation have belonged to a different canonical action-kind. Depending on the case at hand, an artifactual object, according to Evnine’s account, can inherit further essential features (e.g., membership in a kind or a kind-associated function) from the essence of the act of creation which produces it, by way of the maker’s intentions. Thus, in cases in which an agent intends to produce a member of an already existing artifact kind K, she must be guided by a concept of what kind of artifact she takes herself to be producing. In these cases, the artifact kind in question makes its way into the essence of the resulting artifact through the creative intention of the agent who produces it and the resulting artifact essentially belongs to kind K. Moreover, an artifact which belongs to a kind with a kind-associated function also has the relevant function essentially. In Evnine’s view, however, it is compatible with an artifact having a kind-associated function essentially that the artifact in question is not actually able to carry out the function: for a broken K, for Evnine, is still a K. language-communities (Evnine (2016a), Section  4.4.3). Evnine furthermore pursues the contrast he sees between his own and Thomasson’s approach to ontology in greater detail in Evnine (2016b). Although I cannot hope to do justice to this important set of issues in the present context, I will briefly return to the question of whether Evnine has successfully distinguished his own account from Thomasson’s ontological minimalism in the way that he intends.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

224  Artifacts In addition, it is also compatible with an artifact’s having a kind-associated function essentially that the agent who produces the artifact does not actually intend to produce something which will be used to carry out the kind-associated function: for the agent may instead intend to produce an artifact with an idiosyncratic function (e.g., a show exemplar). Nevertheless, since (in order for the corresponding act of creation to be successful) the agent must intend to produce something of kind K, the resulting artifact will have the kind-associated function essentially, simply by virtue of being essentially a member of kind K. Further exceptions to the generalizations mentioned here are required in order to account for cases in which an agent intends to produce a prototype as well as cases in which an artifact does not obviously have a (kind-associated or idiosyncratic) function at all. Thus, in cases of prototype production, for example, an agent does not intend to produce something which belongs to an already existing artifact kind K, since (as far as she knows) the relevant kind does not (yet) exist. Further, Evnine allows that some artifacts (e.g., artworks) may not inherit a function from their essential membership in a certain kind or from a certain idiosyncratic use their makers have in mind for them. Thus, even though Evnine in general takes artifacts to inherit their essential features from the intentions of the agents who produce them, by way of the relevant act of creation, his account does not legislate a uniform conception of artifact-essences according to which all artifacts are essentially members of an already existing kind or essentially have a certain (kind-associated or idiosyncratic) function. Most minimally, then, an artifact, for Evnine, is just essentially the product of a certain act of creation in which an agent successfully exercises her intention to create a certain thing by working in an appropriate fashion on some suitable matter. As Evnine (2013) describes for the controversial case of “Ready-Mades” (e.g., Marcel Duchamp’s Fountain), very little physical work may in fact be required on the part of an agent in order for her to count as having successfully exercised her intention to create a certain thing by working in an appropriate fashion on some suitable matter. Evnine’s position concerning “Ready-Mades” is that a metaphysical account of artifacts and artworks should not rule out the possibility that an artist might succeed in creating a new artwork (e.g., a sculpture) merely by “selecting” an existing artifact (e.g., a urinal) for an exhibition. A theory which excludes such a possibility, in Evnine’s view, suffers from the disadvantage that it is unable to explain why artists might be interested in producing “Ready-Mades” in the first place or why audiences find them so provocative (see Evnine (2013), p. 423).

8.3.3 Functions According to the account of artifacts developed by Lynne Rudder Baker (see especially Baker (2004), (2007)), artifacts have a certain “proper function” essentially. The proper function of an artifact is the purpose or use intended for the artifact by its “author(s),” viz., the artifact’s designer(s) and/or producer(s). (Depending on the case, either a single author or multiple authors may be responsible for the design and/or production

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

artifact-essences  225 of an artifact.) Baker’s emphasis on functions certainly appears to distinguish her account from the two approaches outlined earlier, since both Thomasson and Evnine make room for the possibility that entities (e.g., artworks) may count as artifacts even if they do not obviously have (kind-associated or idiosyncratic) functions. Once we take into consideration, however, that Baker uses the term “artifact,” in a more restricted sense than our previous two authors, to refer to what she calls “technical artifacts,” the differences between Baker’s account and those defended by Thomasson and Evnine turn out to be less pronounced than they might at first appear. A “technical artifact,” for Baker, is an artifact that is designed and/or produced to attain a certain practical goal; artworks are explicitly excluded from the scope of Baker’s account. In addition, Baker agrees with Thomasson and Evnine in assigning a central role to the authors’ intentions in determining an artifact’s essence, viz., its proper function. Like Thomasson and Evnine, Baker views artifacts as in no way ontologically deficient or any less real than the members of natural kinds, despite their mind-dependence (or “intention-dependence,” as Baker prefers to say). Baker conceives of artifact-functions in a normative way, as being conceptually tied to the notion of malfunction: an artifact’s proper function is the function it is supposed to carry out, assuming that the artifact operates in the way intended by its authors. It is compatible with an artifact having its proper function essentially, in Baker’s view, that the artifact is not in fact able to carry out its proper function. But Baker is careful to note that not all cases in which an artifact fails to operate as intended by its authors should be accepted as cases in which the artifact malfunctions. Thus, a case in which an artifact fails to carry out its proper function counts as a malfunction, in Baker’s view, just in case, firstly, it is physically possible to perform the artifact’s proper function and, secondly, the artifact is being operated by a competent user and in conditions in which the artifact is intended to be operated. The first condition is violated, for example, by a perpetual motion machine or by an amulet whose intended function is to ward off evil spirits. The second condition fails to obtain, for example, when a car stops running because it is out of gas or when a machine fails to operate as intended because the person operating it lacks the proper training. Other authors, who have also emphasized the centrality of functions in their accounts of artifacts, employ a notion of function that is quite different from Baker’s. For example, Crawford Elder, following the work of Ruth Millikan in the philosophy of biology (see, e.g., Millikan (1984)), conceives of artifact-functions in a causal-historical way. Thus, in Elder’s view, the members of artifact-kinds (and so-called “copied kinds” more generally) are “produced by a process or mechanism which copies them from previous members similarly shaped, and does so as a causal consequence of performances, by those previous members, of certain functions—productions by them of certain effects” (Elder (2007), p. 38). Thus, earlier members of the kind, screwdriver, for instance, produced the effect of tightening and loosening screws; and the fact that later members of the same kind were copied from earlier members is a causal consequence of the fact that earlier members of the kind, screwdriver, produced the effect of tightening

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

226  Artifacts and loosening screws. In this case, the tightening and loosening of screws are identified by Elder as the “proper function” associated with members of the kind, screwdriver. More generally, when this causal-historical connection obtains between the production of effects by earlier and later members of a copied kind, Elder identifies the production of the effect in question as the proper function associated with members of the copied kind. Elder takes his account of artifacts to have a distinct advantage over its competitors in that it allows, in his view, for a mind-independent conception of the reality of artifacts. Human beings and their mental states, for Elder, are only causally, but not constitutively, responsible for the production of artifacts; and hence author-intentions in particular do not figure either directly or indirectly into artifact-essences.7

8.4  Challenges for Author-Intention-Based Accounts of Artifact-Essences So far, we have encountered three different styles of accounts of artifact-essences. According to the first, associated with Amie Thomasson, an artifact of kind K is essentially the product of a largely successful intention on the part of a maker to create something of kind K. According to the second account, associated with Simon Evnine, an artifact of kind K is essentially the product of a certain act of creation performed by an agent who intends to create something of kind K. According to the third account, an artifact of kind K essentially has a certain proper function associated with objects of kind K. This third account can be developed in different ways, depending on whether the operative notion of function is understood, e.g., as determined by author-intentions (as on Baker’s account) or as resulting from a causal-historical relationship of some sort (as on Elder’s account). Even though Thomasson’s, Evnine’s, and Baker’s accounts all differ from each other in important and interesting ways, all three accounts agree in tracing the essence of an artifact either directly or indirectly back to the intentions of those agents (inventors, makers, authors, designers, or producers) who are primarily responsible for the creation of the artifact. Despite their differences, I will in what follows refer to all three accounts with the short-hand, “author-intention-based accounts.” Author-intention-based accounts are open to serious challenges, among them the following.8 7   I will not discuss the Millikan/Elder account of artifacts further here; but see Thomasson (2003b, 2007) for objections to this style of account. See also Kornblith (1980, 2007) for a defense of the view that membership in an artifactual kind is determined by sameness of functions. Other than what is suggested by Kornblith’s examples, however, he does not specify what notion of function he takes to be operative in his account. For a helpful survey and discussion of the plurality of ways in which the notion of function is employed by biologists and philosophers of biology, see Wouters (2005). 8   In the remainder of Section  8.4, I focus on four objections which can be raised against authorintention-based accounts of artifact-essences. This selection is by no means meant to be exhaustive; rather, I single out objections which strike me as especially important, not least in that I find them to be indicative of the vulnerabilities characteristic of author-intention-based accounts. In this way, our discussion gives us some indication of how to design a positive account of artifacts down the road which is sensitive to the challenges faced by author-intention-based accounts.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

author-intention-based accounts of artifact-essences   227

8.4.1 User-Intentions It appears to be possible, under certain circumstances, for the intentions of the later users of an artifact to override the original author-intentions in determining what features are relevant to an artifact’s membership in a certain kind K. Thus, Kornblith (2007), for example, remarks: . . . [T]he person who first made carabiners may have had quite a different intention in making them than the users do in using them, and if the maker, now long gone, is the only one who ever had that intention, and all of the users have a different intention, then arguably the intentions which are connected to making the kind what it is are probably those of the users rather than the maker. [. . .] The maker could not insist, “I know what these things are; after all, I made them”, since the term is part of a public language which the maker cannot constrain through a sheer act of will.  (Kornblith (2007), p. 145)

In the scenario Kornblith imagines, artifacts of the kind, carabiner, were originally  intended to have a certain function, e.g., F, by their author; but later users came to associate members of the kind in question with a different function, e.g., F’.  Moreover, Kornblith supposes that only the original author, at a time in the ­distant past, ever thought of carabiners as having the function, F, while everyone else who has since used carabiners thought of them as having the function, F’. Under these circumstances, Kornblith suggests, it is plausible to hold that the intentions of carabiner-users may succeed in overriding the intentions of the original author in determining which features are relevant for an item’s membership in the kind in question. The possibility that it is, under certain circumstances, possible for user-intentions to override author-intentions can be further illustrated by considering the following case. Suppose that Alexander Graham Bell, the inventor of the telephone, initially intended his new device to be used as an aid for the hearing-impaired, while later users came to think of the telephone as a certain kind of long-distance communication device which allows two or more users (whether they are hearing-impaired or not) to carry on a conversation even when they are far apart.9 Given author-intention-based accounts of artifact-essences, assuming that Alexander Graham Bell in fact intended the device he invented to have a certain function, viz., to aid the hearing-impaired, and assuming that there is no obvious reason to think that Bell’s original author-intentions misfired during the production of the first prototype, then the device Bell invented is in fact a hearing-aid (and essentially so); and the same applies to every subsequent device which is successfully produced with the intention of being of the same type as

9   I borrow this example from Carrara and Vermaas (2009), p. 135, who suggest that the description just cited is historically accurate. I remain neutral on the question of historical accuracy and wish to use the scenario only as a further illustration of the possibility imagined by Kornblith, that user-intentions, under certain circumstances, can override author-intentions in determining which features are relevant to an artifact’s membership in a certain kind.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

228  Artifacts the device Bell invented.10 Proponents of this view are committed to holding that the intentions of later users cannot override Bell’s original author-intentions, according to which the device he invented is a kind of hearing-aid, and lead to a reclassification of the telephone as a certain kind of long-distance communication device. But the scenario under consideration suggests that it is in fact possible, under certain circumstances, for the intentions of later users to override the intentions of the original author as to how the device he or she has invented, designed, or produced is to be used. The following three types of responses come to mind when faced with scenarios in which user-intentions appear to be able to override original author-intentions in determining the features that are relevant to an artifact’s membership in a certain kind. According to the first response, when the scenario in question is appropriately redescribed, the original author-intentions turn out to be sufficiently general to accommodate the apparent shift in the function ascribed to the artifact kind in question. Following this line of thought, we might attribute to Bell only the intention that his new device be used to convert sound into electronic signals, leaving open whether these electronic signals are then amplified (as in the case of hearing-aids) or transmitted over long distances (as in the case of long-distance communication devices). But this redescription seems to mischaracterize the more specific original author-intention we attributed to Bell in the scenario in question, according to which Bell in fact intended his invention to be used as a device which, upon converting sound into electronic signals, amplifies these electronic signals in such a way as to aid the hearing-impaired, with no special focus on the transmission of these electronic signals over long distances. According to the second response, we are to assume that, in cases of apparent conflict, the function ascribed to an artifact by later users merely supplements (but does not replace) the function ascribed to the artifact in question by the original author. Thus, following this response, the function associated with the kind, telephone, initially specifies that these devices are to be used as an aid for the hearing-impaired and is later broadened to include their use as long-distance communication devices. But this second response has the counterintuitive consequence that all those devices which are in fact classified as hearing-aids (i.e., devices which, upon converting sound into electronic signals, only amplify but do not transmit these electronic signals over long distances) would then be included in the kind, telephone, and therefore classified as 10   Moreover, we may further precisify the scenario at hand in such a way that Bell’s original authorintention and the intentions of later users are not mediated by an intervening “gray area” in which the function Bell associated with the telephone gradually “morphs” into the function later users associate with the same device by a series of small steps. To exclude this reading, let us suppose that the scenario at hand is closer to Kornblith’s case involving carabiners, so that only Bell ever intended his device to be used as an aid for the hearing-impaired, and all subsequent users (and makers) intend the device Bell invented (and devices that are of the same type as the device Bell invented) to be used as a device which enables two or more users to carry on conversations over long distances. Given this hypothesis, the dramatic divergence in question between Bell’s original author-intention and those of later users cannot be traced to any sort of vagueness or malleability inherent in our artifact-concepts.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

author-intention-based accounts of artifact-essences   229 being of the same type as the device Bell invented. In fact, given the second response, we would be forced to conclude that these latter devices (i.e., devices which we now classify as hearing-aids) reflect Bell’s original author-intentions more accurately than the devices we now classify as telephones (i.e., devices which, upon converting sound into electronic signals, transmit these signals over long distances). According to the third response to the scenario considered earlier, we are to retain the idea that the features relevant to an artifact’s membership in a kind are in fact determined by the intentions of agents, but allow that this responsibility in some cases falls to the users of an artifact rather than to its original author. Endorsing this third response is tantamount to giving up on an exclusively author-intention-based account of artifact-essences. In addition, the third response also offers no general recipe as to how to adjudicate, in cases of conflict, between the intentions of the original author and the intentions of later users as to which features are relevant to an artifact’s membership in a given kind. Moreover, an account of artifact-essence which exclusively assigns the responsibilities of determining kind-relevant features to the users of an artifact runs into difficulties in cases in which a new type of artifact is invented, designed, or produced by an author but either never develops a following among users or does so only after a certain period of time. For an exclusively user-intention-based account, in that case, would lack the resources to attribute essences or kind-relevant features to artifacts for which appropriate user-intentions are not available.11 In sum, the considerations presented in this section seem to suggest that, under certain circumstances, user-intentions can indeed override original author-intentions in determining an artifact’s kind-relevant features. As it stands, such a possibility is incompatible with author-intention-based accounts of artifact-essences, since these approaches maintain that the essence of an artifact is, directly or indirectly, determined by the intentions of an artifact’s original author, assuming that the author successfully exercises her intention to create a certain thing by working in an appropriate fashion on some suitable matter. I discussed three responses to the challenge presented here. All three responses, however, were found to lead to unwelcome results and should therefore be resisted. In Section 8.4.4, I will briefly return to the challenge posed in this section and offer my own thoughts on how best to diagnose scenarios in which 11   The following possible fourth response to the carabiner and telephone/hearing-aid scenarios was suggested to me by Simon Evnine (pers. comm.). Perhaps the terms, “carabiner” and “telephone,” as they are used in the descriptions of these scenarios, are actually ambiguous and designate two different artifact kinds: one kind that has the intended proper function associated with them by the original author and a second kind that has the intended proper function associated with them by later users, who have in effect themselves become authors of a new artifact kind. Following this response, when Alexander Graham Bell uses the term, “telephone,” he designates an artifact kind whose intended proper function is to serve as an aid to the hearing impaired; but when later users/makers use the term “telephone,” they designate a different artifact kind whose intended proper function is that it be used as a long-distance communication device. This response seems to lead to an unnecessary multiplication of distinct artifact kinds and distinct meanings that are associated with our established artifact terms. An account which allows that the intentions of later users can override the intentions of the original author as to how his or her intention is best put to use avoids the ontological and semantic proliferation incurred by this fourth response.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

230  Artifacts user-intentions appear to be able to override author-intentions in determining an artifact’s kind-relevant features.

8.4.2  Easy Ontology Author-intention-based accounts of artifact-essences seem to be susceptible to the worry that it might just be too easy to create new artifacts and new artifact kinds, given their perspective.12 For example, no matter how hard I try, it appears that I simply cannot create a thermometer out of a pen merely by thinking certain thoughts or by uttering certain words. Unless the pen undergoes a fairly radical transformation in the process of my attempt to make a thermometer out of a pen, my would-be creative intention will be thwarted, given that a pen (unless dramatically altered) just does not have what it takes to detect and indicate changes in temperature, regardless of what mental or linguistic attitude I or anyone else adopts towards the pen. Thus, proponents of author-intention-based accounts will presumably want to set some limits on the circumstances in which an agent’s adoption of a certain mental or linguistic attitude towards some pre-existing objects by itself is sufficient for the creation of a new artifact. Such approaches therefore owe us an account of how to delineate circumstances in which an author succeeds in exercising her creative intentions from circumstances in which an author’s would-be creative intentions misfire in some way. When considering the prospects for proponents of author-intention-based accounts in formulating such success conditions, we should keep in mind that these approaches are quite liberal in many respects when it comes to an agent’s ability to create new artifacts. To illustrate, recall the following three types of possibilities which may arise according to an author-intention-based model of artifact production. Firstly, proponents of author-intention-based accounts allow for the possibility that, in a successful case of prototype production, an agent can under the right circumstances succeed in producing a new artifact which does not belong to an already-established artifact kind. In such cases, an agent manages to bring into existence not only a particular exemplar, but also (where successful) a whole new artifact kind. Secondly, in cases of malfunction, an item may count as belonging to an artifact kind with a kind-associated function, even if the particular exemplar under consideration is not able to carry out the function associated with the artifact kind to which it belongs. (Recall Evnine’s slogan: “A broken K is still a K.”) Thirdly, in cases involving “Ready-Mades” or “Found Objects,” author-intention-based accounts make room for the possibility that an agent 12   For discussion, see, for example, Evnine (2016a), Section 3.5, pp. 110–18. Evnine discusses a version of this objection raised in Zimmerman (2002) against Lynne Rudder Baker’s account, as set out in Baker (2000, 2002). Evnine argues that, while his own account can successfully defeat the objection at issue, rival accounts (in particular, Thomasson’s) are susceptible to it. We will return to Evnine’s defense of his own account against this objection shortly. I am here using the phrase, “easy ontology,” somewhat differently from how it is used by Thomasson: when Thomasson uses this phrase, she has in mind the ease with which apparently disputed ontological questions can be settled; when I use this phrase, I have in mind the ease with which new artifacts and new artifact kinds can apparently be created, according to author-intention-based accounts.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

author-intention-based accounts of artifact-essences   231 may succeed in exercising her creative intentions to produce a certain artifact from some underlying matter, even when the agent contributes no physical work and the underlying matter does not undergo any intrinsic change during the process of “production.” In cases involving “Found Objects,” the agent appears to be simply lucky in that a pre-existing object (e.g., a piece of driftwood) is already in the right shape intrinsically to serve as the matter for the alleged newly created artifact (e.g., a bench) without requiring any additional physical work on the part of the agent. Alternatively, the alleged newly created artifact may also belong to an artifact kind (e.g., the kind, “Ready-Mades”) whose members (perhaps not unlike marriages or promises) can apparently be created in suitable contexts simply by uttering certain words or thinking certain thoughts, or by inducing changes with respect to the relational properties exhibited by the pre-existing objects in question.13,14 Given these three possibilities, we can now imagine several different scenarios which appear to spell trouble for author-intention-based accounts. The first type of scenario involves the apparent creation of a new artifact belonging to an already-existing artifact kind; the second type of scenario involves the apparent creation of a new artifact which (if successful) would constitute a prototype of a not-yet-existing artifact kind and which (if successful) would therefore bring with it the creation of a whole new artifact kind. In both types of scenarios, the newly created artifact (given the possibility of malfunction) need not actually be able to carry out the function the agent has in mind for the artifact she intends to create. Moreover, in both types of scenarios (given the possibility of “Ready-Mades” or “Found Objects”), the agent need not physically work on, or intrinsically alter, the pre-existing objects which (assuming that the act of creation is successful) will serve as the underlying matter for the newly created artifact. Instead, the agent may, in the right circumstances, create a new artifact simply by thinking 13   As noted earlier, Baker’s conception of artifacts, unlike Thomasson’s and Evnine’s, is restricted in such a way as to exclude artworks from the category of artifacts. Nevertheless, the case of “Found Objects” serves to bring out the possibility envisioned here, in a way that is compatible with Baker’s account. 14   No doubt, as Evnine (2016a) suggests (see pp. 111–12), it would be helpful in this context for proponents of author-intention-based accounts to appeal to a distinction between being a K and being (merely) used as a K. Thus, proponents of this view, so Evnine argues, need not agree that a piece of driftwood is a bench, simply because it is used as a bench. But we may wonder whether proponents of author-intention-based account are really entitled to help themselves to this distinction without begging the question. Suppose we are wondering, in a particular scenario, whether an agent has succeeded in creating a new member of the already-existing artifact kind, bench, from a piece of driftwood, without intrinsically altering the piece of driftwood. On Evnine’s account, a bench is the product of a certain act of creation in which an agent successfully exercises her intention to create a bench by working in an appropriate fashion on some suitable matter. Given Evnine’s receptiveness towards the case of “Ready-Mades,” the “working in an appropriate fashion on some suitable matter” component that is built into Evnine’s conception of artifact creation can in some instances be satisfied even when the agent does not contribute any physical labor and the underlying matter from which the alleged newly created artifact is “produced” does not undergo intrinsic change. Thus, there seems to be no reason in principle why, on Evnine’s account, an agent would not be able to create a new artifact, viz., a bench, from a piece of driftwood without intrinsically altering the piece of driftwood, as long as the piece of driftwood is already in the right shape intrinsically to serve as the matter for a bench. In such a case, the agent would simply appear to find herself in the fortunate position that she can exercise her intention to create a bench from a piece of driftwood successfully without physically working on and intrinsically altering the piece of driftwood.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

232  Artifacts certain thoughts or speaking certain words, or by relationally (but not intrinsically) altering the pre-existing objects which (in a successful case of artifact-production) will serve as the underlying matter for the newly created artifact. Beginning with the first type of scenario, we may ask what resources proponents of author-intention-based accounts have at their disposal to impose constraints on an agent’s ability to exercise her would-be creative intentions successfully when she intends to create a new member of an already-existing artifact kind. Suppose, for example, an agent intends to “convert” a pen into a thermometer: under what circumstances, we may ask, does she count as having succeeded or failed to succeed in creating a new artifact (viz., a thermometer) from an already-existing artifact (viz., a pen)? Given the possibility of “Ready-Mades” or “Found Objects,” the agent need not physically work on, or intrinsically alter, the pen in order to “convert” it into a thermometer. Moreover, given the possibility of malfunction, the agent need not produce a device which is able to carry out the function she intends the outcome of her would-be creative intentions to perform (e.g., to measure the temperature in a  certain environment). No doubt, proponents of author-intention-based accounts would balk at the idea that, under the circumstances just described, a new artifact, viz., a broken or malfunctioning thermometer, has in fact been created from a pre-existing artifact, viz., a pen. Given the possibilities outlined earlier, however, it is unclear what is to prevent us from concluding that, in the defective case just imagined, a new member has been added to an already-existing artifact kind. Secondly, we may ask what sorts of constraints may be imposed on scenarios of the second type in which an agent intends to produce a prototype of a not-yet-established artifact kind, and thereby (where successful) also simultaneously creates a whole new artifact kind. The ability to impose such constraints is particularly important from the point of view of proponents of author-intention-based accounts who (like Evnine and Baker) see their own approach as less ontologically permissive and more substantively realist than Thomasson’s neo-Carnapian “ontological minimalism” or “easy ontology” (see, e.g., Thomasson (2015)). When discussing this concern, Evnine remarks: Chairs exist [according to Thomasson] when people work with certain intentions on appropriate matter not, as I have it, because those people are exercising some creative power to bring chairs into existence thereby, but simply because the conditions contained in the concept obtain. If we define a concept not-thair in such a way that it applies just when a person ignores some potential chair-matter and does nothing to it, then we can say that not-thairs exist (and have always existed) in the same way, under conditions that clearly are not genuinely creative in the firstorder sense. Thus the agreement between ontological minimalism and my own view over CH [a principle stating the existence conditions for chairs], and other principles like it pertaining to other artifactual concepts, is a coincidental convergence between two theories that are quite different in spirit.  (Evnine (2016a), p. 117)15 15   Here, “CH” stands for the following principle stating the existence-conditions for chairs: “[i]t is sufficient for the existence of a chair that some matter be worked on in the appropriate fashion (i.e., that the matter be arranged chair-wise) and this is true in virtue of what it is to be a chair (i.e., in virtue of the essence of chairs)” (Evnine (2016a), p. 116).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

author-intention-based accounts of artifact-essences   233 The disagreement between Evnine and Thomasson, for the particular case of artifacts, thus comes down to Evnine’s more restrictive conception of what it takes for an agent to exercise, or fail to exercise, a genuinely “creative power.” But the very question before us is precisely what distinguishes a genuinely “creative power” from one that only appears to be “creative,” e.g., one which issues from an agent’s apparently defective would-be creative intention to produce a “not-thair” by appropriately ignoring and doing nothing to some potential chair-matter. What, in Evnine’s account, blocks an agent from having or successfully carrying out a “pseudo-creative” intention to bring into existence a “not-thair?” As far as I can see, Evnine’s account does not, as it stands, deliver a conclusive answer to these questions, since doing so would require him to take a stand on a question he leaves open for future research, viz., how to conceive of the essences of mental states generally and those of intentions specifically (see Evnine (2016a), pp. 243–5 for discussion). Evnine’s attempt to distinguish his own approach to ontology from Thomasson’s must therefore await the fulfilment of this promissory note. In the meantime, though, we should note that however exactly the essences of mental states are construed, the success conditions for an agent’s attempt at exercising a would-be creative intention cannot appeal to the existence or non-existence of the kinds of objects which allegedly form the subject matter of these intentions: for such an account would render Evnine’s attempt to differentiate his own version of ontological realism from Thomasson’s neo-Carnapian approach circular. Evnine thus cannot help himself to the non-existence of “not-thairs” in the course of formulating an account of why an agent cannot succeed in having or carrying out a would-be creative intention to create a “not-thair.” Nor can he, without further explanation, appeal to the faulty nature of the concept, “not-thair,” since the question which needs to be answered by his account is precisely what (if anything) makes this concept faulty. Given that neither of these escape routes are open to Evnine in order to explain the defective nature of the relevant mental state, however, we want to know what else might be wrong with the idea that an agent can have, or attempt to carry out, a would-be creative intention to produce a “not-thair.”16 16   When Evnine takes up the question of what distinguishes his own account from Thomasson’s “ontological minimalism” or “easy ontology” in more detail elsewhere (see Evnine (2016b)), he argues that Thomasson’s account falls prey to what he calls “the problem of too much content.” Evnine’s objection is intended to draw our attention to the way in which Thomasson conceives of the application condition for concepts. As an illustration, Evnine considers the particular case of mereological sums or fusions (according to Classical Extensional Mereology) and argues that the ontological minimalist confronts the following dilemma: either (i) the application conditions governing the concept, mereological sum, require nothing more for the existence of a sum than the mere existence of a plurality of objects; or (ii) the application conditions of the concept, mereological sum, do require something more for the existence of a sum besides the mere existence of a plurality of objects (e.g., that the objects in question stand in some particular relation, viz., parthood, to the sum). Only if (i) is the case (so Evnine argues) may the ontological minimalist hold on to her contention that the question of whether mereological sums exist is a trivial matter. In contrast, if (ii) is the case, then (so Evnine argues) the question of whether mereological sums exist becomes metaphysically substantive, since the existence of sums is then no longer analytically tied by way of the application conditions associated with the concept in question to the mere existence of a plurality of objects. Evnine aims to show that in fact (ii) is the case, as the ontological minimalist implicitly recognizes. In that case, however, Thomasson’s approach falls prey to “the problem of too much content,” since she

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

234  Artifacts In this section, we considered the “easy ontology” objection to author-intention-based accounts of artifact-essences. According to this objection, such accounts are susceptible to the worry that it might just be too easy, given their perspective, for an agent to create a new artifact as well as potentially (in cases of prototype production) a whole new artifact kind, e.g., simply by adopting a mental or linguistic attitude in the right context or by relationally (but not intrinsically) altering some pre-existing objects. In order to put this concern to rest, proponents of these approaches must be able to offer a method by which to distinguish between successful and unsuccessful attempts by agents to exercise an intention to create a new artifact and/or a new artifact kind. A satisfying response to the “easy ontology” objection, moreover, cannot (without further explanation) presuppose that we already have in place criteria by which to distinguish between genuinely creative and pseudo-creative powers or intentions; or between defective and non-defective concepts which might be guiding agents in their attempt to exercise such powers or intentions. To provide such a response to the “easy ontology” objection is especially important for those proponents of authorintention-based accounts who (like Evnine and Baker) want to adopt an approach to ontology that is (in their view) less permissive and more substantively realist than Thomasson’s neo-Carnapian “ontological minimalism.” But whether or how these details are best filled in, without generating a circularity for these accounts, is at this point still an open question.17

8.4.3  Further Objections: Mass-Production and Scope I will here only briefly mention two further objections to which author-intentionbased accounts appear to be vulnerable. For reasons of space, I will not explore these objections, or possible responses to them, further in the present context. The first of these objections arises from the phenomenon of mass-production, while the second concerns the scope of the theories under consideration. In cases of mass-production, a single act of creation results in the production of multiple artifacts. Given the proposed specification of artifact-essences in terms of author-intentions, one wonders how the essence of one mass-produced artifact could be distinguished from that of another. Evnine (2016a) discusses this objection (see must build more content into the application conditions governing concepts than is compatible with her ontological minimalism. However exactly Thomasson responds to this particular challenge involving mereological sums, Evnine’s objection unfortunately does not help him further in the case of artifacts, since his own account agrees with Thomasson’s on the application conditions for artifact concepts, i.e., the conditions which must be satisfied in order for an artifact to come into existence (viz., that an agent must have successfully exercised her creative intention to produce a certain thing by working in an appropriate fashion on some suitable matter). 17   When Baker responds to a version of the “easy ontology” objection launched against her account in Zimmerman (2002), she proposes that we cannot simply think or speak things into existence for which “our conventions and practices do not have a place” (see Baker (2007), p. 44). (See Evnine (2016a), pp. 113ff, for a discussion of Baker’s reply to Zimmerman.) In order for this response to be effective, however, we would need to know more about how our conventions and practices support the creation of some artifacts and artifact kinds, while disallowing the attempted creation of others.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

author-intention-based accounts of artifact-essences   235 Section 3.4.2, pp. 97–103) and proposes that, in cases in which an agent produces multiple artifacts in a single act of creation, the objects in question share a single “collective essence,” but lack distinct individual essences. In such scenarios, it is thus impossible (on metaphysical, and not merely on epistemic, grounds) to distinguish between the individual artifacts that are created in a single act of mass-production, though collectively it is possible to distinguish these artifacts which are mass-produced in a single act of creation from other artifacts which are produced in a distinct act of creation. As Evnine recognizes, this response to the problem of mass-production will strike many as puzzling and he pursues the challenges posed for his account by the phenomenon of mass-production further in Evnine (2018). Secondly, in connection with the scope of author-intention-based theories of artifacts, recall that the theorists considered earlier restrict their respective use of the term “artifact,” in various ways. Thomasson and Evnine, on the one hand, only count the intended products of certain kinds of human activities as artifacts, thereby excluding entities of type (3)–(4) considered earlier in Section 8.2 from the category of artifacts proper (viz., by-products, residue, or other unintended outcomes as well as the products of non-human intentional activity). Baker’s account, on the other hand, is tailored specially to “technical artifacts,” i.e., devices which are designed or produced to attain a certain practical goal, thereby excluding “Ready-Mades” and other artworks from the targeted range of entities (see Section 8.2, (2), and (5)). Given these restricted uses of the term “artifact,” it remains to be seen what ontological status should be assigned to these entities which are not included in the category of artifacts proper. Do they belong to genuine kinds and, if so, to what kinds do they belong? Do they have essences and, if so, what determines their essence and their membership in the relevant kind? Since these entities do not appear to fall under already recognized natural kinds and they also lie outside the scope of the relevant author-intention-based accounts of artifacts, these questions concerning their ontological status are, as of yet, not settled.

8.4.4  The Limits of Human Creative Intentions To sum up, we have in Section 8.4 come across four challenges to author-intentionbased accounts of artifact-essences: (i) the possibility that user-intentions, under the right circumstances, might override author-intentions in determining the features relevant to an artifact’s membership in an artifact kind; (ii) the apparent absence of acceptable limits set by author-intention-based accounts on the creation of new artifacts and artifact kinds by agents who merely think certain thoughts or utter certain words in the right context, or relationally (but not intrinsically) alter some pre-existing objects; (iii) the apparent lack of acceptable methods by which to distinguish between multiple artifacts that are mass-produced in a single act of creation; and (iv) the apparent ontological “no-man’s land” occupied by a host of entities which fit neither into already accepted natural kinds nor into the restricted category of artifacts proper as it is conceived of by proponents of author-intention-based accounts.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

236  Artifacts I take away the following main lesson from the preceding discussion of challenges faced by author-intention-based accounts of artifact-essences: author-intentions are neither as powerful nor as discriminating as proponents of these accounts make them out to be. Scenarios of the type considered in Section 8.4.1 suggest that the intentions of later users can, in some cases, be at least as authoritative as, or more authoritative than, the intentions of the original author in determining what features are taken to be relevant to an artifact’s membership in an artifact kind. The “easy ontology” objection considered in Section 8.4.2 brings out that author-intention-based accounts have trouble setting reasonable limits to the creative powers ascribed to agents. Thirdly, as illustrated briefly in Section 8.4.3 by way of the phenomenon of mass-production, author-intentions are not sufficiently fine-grained to yield a method by which to discriminate between multiple apparently distinct artifacts, when these are mass-produced in a single act of creation. Lastly, in connection with the fourth challenge touched on in Section 8.4.3, we noted that, in a host of cases, author-intentions do not extend far enough to be able to determine essences or kind-relevant features for entities that are currently excluded from restricted uses of the term “artifact,” thereby leaving these entities “dangling” between already accepted natural kinds and the category of artifacts proper. Given that the creative and discriminating power of human intentions is limited in these ways, we ought therefore to take seriously the possibility, disallowed by authorintention-based accounts, that original authors can sometimes be mistaken about, or fail to appreciate fully, how the devices they have invented are best put to use. This alternative perspective suggests that the kind-relevant features associated with a newly invented artifact should not be taken to be stipulatively determined or analytically tied to the concept or intention guiding the artifact’s inventor during prototype production. Rather, the question of how a newly invented artifact is best put to use might be resolved at least in part by empirical methods that are in principle just as, or possibly more, accessible to the later users of the device than they are to the device’s inventor. Given this perspective, the original author who introduces a new artifact or artifact kind should not be assumed across the board to occupy an epistemically privileged position with respect to his or her own invention. Moreover, I draw the following further skeptical conclusion from our preceding discussion of the challenges faced by author-intention-based accounts. Presumably, the natural choices to consider for those who want to develop essentialist accounts of artifacts are just the candidates discussed earlier. These accounts trace the essence of an artifact either directly to the intentions of the original author; or they do so indirectly by way of the creative acts which lead to the production of an artifact or by way of the primary intended function that is associated with an artifact by its original author. As we observed, the first of these routes toward artifact-essences is taken by Thomasson, while Evnine and Baker choose the second and third options, respectively. We briefly considered a fourth possibility, viz., an account of artifact-essences that is formulated in terms of the intentions of later users (see Section 8.4.1); but we rejected this option due to the obvious difficulties to which it gives rise.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

challenges for existing anti-essentialist frameworks  237 We have thus had a chance to examine accounts of artifact-essences which appeal to the following four types of features: (i) original author-intentions, (ii) creative acts, (iii) primary intended functions, or (iv) user-intentions. Since all four of these options, as we noted earlier, give rise to serious difficulties, the question thus arises as to what other options are available for those who are drawn to essentialist accounts of artifacts. I submit that (i)–(iv) in fact represent the most plausible choices for proponents of essentialist accounts of artifacts. Given that we have identified shortcomings with all four of these essentialist approaches, it is at this point tempting to entertain the possibility that artifacts might find a more suitable home within the confines of an anti-essentialist theory. In many cases, after all, artifacts are created by human beings to serve human interests; and we might therefore expect that their integration into existing anti-essentialist frameworks would be an exceedingly straightforward matter. As I will indicate briefly in Section 8.5, however, popular anti-essentialist strategies are, perhaps surprisingly, not particularly well-equipped to handle the special features of artifacts. Thus, we must, for the time being, concede that the domain of artifacts presents us with an ontological problem case which still awaits a fully satisfactory resolution.

8.5  Challenges for Existing Anti-Essentialist Frameworks While anti-essentialism comes in many flavors, what is common to these approaches is that they interpret ascriptions of essential features to objects as, in some way, reflective of human interests or activities, rather than of the world “as such,” i.e., as it would be without human beings in it. Thus, if it were possible somehow to subtract from the world our characteristically human (e.g., linguistic, mental, social, conventional, explanatory, or other) practices, the result would be, from the point of view of the anti-essentialist, a fundamentally a-modal reality. When we speak of an individual object or a kind of object as being essentially thus-and-so, the anti-essentialist regards such a characterization as being indicative merely of the fact that the feature in question strikes us, from our human perspective, as being especially noteworthy, relative to some standard of noteworthiness. Different anti-essentialists trace our modal vocabulary to different sources. To illustrate, some proponents of an anti-essentialist framework, e.g., Alan Sidelle and Theodore Sider, emphasize our ability to choose, more or less arbitrarily, a set of basic semantic or quasi-logical postulates by reference to which our practice of ascribing necessary features to objects can be explained. Thus, according to Sidelle’s modal conventionalism (Sidelle (1989, 2009)), a necessary truth, such as “Necessarily, bachelors are unmarried,” is made true by a semantic convention adopted by speakers of English according to which the noun “bachelor,” is used to apply to unmarried adult males; but how these semantic conventions governing the meanings of our words are chosen,

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

238  Artifacts in Sidelle’s view, is completely arbitrary and up to us (see, e.g., Sidelle (2009), pp. 235–6). In a similar vein, following Sider’s “Humean strategy” (Sider (2011)), necessary truths are simply theorems which can be inferred from a more or less arbitrarily selected set of axioms (i.e., certain chosen true sentences) via a certain set of truth-preserving rules; but any such choice of axioms and inference rules, so Sider remarks (Sider (2011), p. 271), results in a version of Humeanism. In contrast, other versions of anti-essentialism, such as those offered by Meghan Sullivan or C. Kenneth Waters, see the relevant set of practices which gives rise to our essence attributions or to our classificatory schemes as primarily explanatory. Thus, according to Sullivan’s “explanation-relative essentialism,” the attribution of essential properties to objects is taken to be implicitly relativized to explanatory frameworks governed by objective norms (Sullivan (2017)). Sullivan’s conception allows, for example, that we can truly ascribe to gold the property of being essentially a unit of financial value or the property of being essentially an element with atomic number 79, but only when these ascriptions are relativized to the explanatory practices of economists or chemists, respectively. Any attempt, however, to characterize the essence of gold in a framework-independent way, in Sullivan’s view, would be misguided. C. Kenneth Waters’ “no general structure” thesis (Waters (2017)) advocates an implicit relativization of central scientific concepts to parameters which can be filled in differently in different explanatory contexts. Unlike Sullivan, Waters’ direct concern is not so much with the ascription of essential properties to objects, but with the interpretation of classificatory schemes as they are proposed and employed by practicing scientists. For example, the contemporary molecular gene concept, so Waters argues, does not yield a single correct division of a domain consisting of segments of DNA into a (purported) natural kind, gene, since a gene is always taken to be a gene for a certain trait. When the parameters implicit in the contemporary molecular gene concept are filled in differently in different contexts, depending on the explanatory interests of molecular biologists, different segments of DNA are singled out as genes. Generalizing from this example, Waters proposes that the world lacks a “general structure,” i.e., a structure which would suggest that central scientific concepts yield a single correct division of particulars into natural kinds. This view therefore stands in direct opposition to the conviction of natural kind essentialists, according to whom objects are essentially members of certain natural kinds, independently of the explanatory interests and priorities of scientists. I take our preceding discussion of the special features of artifacts to indicate that neither the conventionalist/Humean strategy chosen by Sidelle and Sider nor the explanation-relative strategy pursued by Sullivan and Waters is especially conducive to an anti-essentialist account of artifacts. If the first anti-essentialist strategy were feasible, our linguistic and inferential practices surrounding artifacts would have to support the formulation of a single consistent and plausible set of basic postulates governing our artifact terms and concepts, by reference to which apparent necessary truths concerning artifacts (e.g., “Necessarily, telephones are long-distance

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

conclusion  239 communication devices”) could be explained. But in their attempt to formulate such postulates governing our artifact terms or concepts, the conventionalist or Humean anti-essentialist would presumably look towards just the same candidates to which proponents of essentialist accounts of artifacts are drawn as well, viz., (i) original author-intentions, (ii) creative acts, (iii) primary intended functions, or (iv) userintentions. Given our preceding discussion of the special features of artifacts, however, we have no more reason to believe that anti-essentialists would be well served to appeal to any of these four options, in their attempt to formulate a consistent and plausible set of basic postulates governing our linguistic or inferential practices involving artifacts, than essentialists were found to be, in their endeavor to formulate a consistent and plausible account of artifact-essences. The second strategy proposed by Sullivan and Waters is also not especially well suited to the formulation of an anti-essentialist account of artifacts. For our engagement with the realm of artifacts is not primarily aimed at explanation, and the expressions and concepts we use to describe artifacts are not embedded in explanatory practices governed by objective norms in the same way, and to the same extent, as is the case with respect to the special science vocabulary which features prominently in Sullivan’s and Waters’ accounts. To illustrate, we would not expect, for example, that certain features of screwdrivers are antecedently ill-understood (e.g., that the shafts of screwdrivers retain their physical integrity even after repeated use), but become transparent only after the relevant population of experts (e.g., engineers) has uncovered certain other explanatorily more basic features of screwdrivers by empirical means (e.g., that their primary intended use is for tightening and loosening screws). Of course, some questions concerning screwdrivers are no doubt the entirely appropriate targets of empirical investigations carried out by engineers or other relevant experts, e.g., which metals are particularly suitable for the manufacturing of screwdriver shafts, given that these shafts are to retain their physical integrity even after repeated use. But such empirical questions only arise once we have taken as fixed that screwdrivers are primarily intended to be used by agents who wish to engage in certain kinds of actions, viz., to tighten and loosen screws. (A similar contrast between natural kinds and artifact kinds also emerged as a result of our discussion of unity in Chapter 7, especially Section 7.5.)

8.6 Conclusion We began this chapter with a brief discussion of the question of how to distinguish artifacts (and artifact kinds) from natural things (and natural kinds). Next, we considered existing essentialist accounts of artifact-essences which are formulated in terms of (i) maker’s intentions, (ii) creative acts, or (iii) primary intended functions. In the course of our discussion, we found that author-intention-based accounts of artifact-essences are open to a variety of serious challenges. Finally, we noted that

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

240  Artifacts existing anti-essentialist strategies are also not as readily applicable to the special case of artifacts as one might have expected them to be. Although I did not attempt to propose a positive theory of artifacts in the present context, several important lessons did emerge from the preceding discussion. For one thing, a successful alternative treatment of artifacts should reflect the conclusions we reached in Section 8.4.4, according to which human creative intentions are neither as powerful nor as discriminating as author-intention-based accounts make them out to be. Rather, our discussion there indicated that an artifact’s kind-relevant features should not in general be taken to be stipulatively determined by, or analytically tied to, the concept or intention guiding the artifact’s inventor during prototype production. For the question of how an artifact is best put to use is at least in part an empirical matter to which the artifact’s inventor does not always have privileged epistemic access. At the same time, however, an adequate treatment of artifacts must also be sensitive to the fact that the aim of our engagement with artifacts is primarily practical and actionbased, and hence not accurately captured by the emphasis placed on our semantic, inferential, or explanatory practices by existing anti-essentialist frameworks. For the time being, the special case of artifacts thus seems to elude an entirely satisfactory treatment and calls for the further development and refinement of existing essentialist and anti-essentialist approaches to the metaphysics of concrete particular objects.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Conclusion In this study, I have developed and defended a hylomorphic conception of concrete particular objects as compounds of matter (hulē) and form (morphē or eidos). Such an approach to the metaphysics of concrete particular objects, I have argued, is wellpositioned to compete with non-hylomorphic alternative conceptions of concrete particular objects when evaluated against a range of measures of success. The application of this doctrine to the specific case at hand, however, does not yield a single uniform hylomorphic conception of concrete particular objects. Rather, theorists who are drawn to the hylomorphic paradigm encounter a series of decision points and desiderata along the way, leading to a potential multiplicity of hylomorphic approaches to the metaphysics of concrete particular objects. These different hylomorphic conceptions of concrete particular objects, however, share a unique strength which, in my view, recommends these approaches over their non-hylomorphic competitors. Those who favor a hylomorphic approach have at their disposal the ability to attribute to concrete particular objects an internal structural complexity that is not recognized by non-hylomorphic alternatives. Hylomorphists associate with each concrete particular object two principles, matter and form, each performing its own range of characteristic explanatory tasks. Moreover, hylomorphists locate these two explanatory principles, in some sense, inside the objects with which they are associated (e.g., by counting them among the object’s proper parts or constituents). This apparatus, I have argued, especially when combined with a certain (non-modal or, more generally, non-reductive) conception of essence, makes hylomorphism a rich and flexible tool for the metaphysical analysis of concrete particular objects and provides hylomorphists with explanatory options that are not, or at least not as readily, available to those who do not endorse the hylomorphic outlook. In Part I of this study, I presented my own preferred interpretation of the doctrine of hylomorphism, as applied to the specific case of concrete particular objects. I argued, firstly, that hylomorphists should view the matter composing a concrete particular object not as prime matter or stuff, but rather as nothing more than its material parts which are themselves analyzed as matter–form compounds, provided that they are structured wholes. Thus, the matter composing a concrete particular object, on this conception, belongs to the same ontological type as the concrete particular object it composes. Secondly, hylomorphists, I argued, should reject the universal forms hypothesis (according to which forms are shareable, repeatable, or multiply located entities of some sort) and

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

242  Conclusion instead endorse a version of the individual forms hypothesis, according to which each concrete particular object has an individual form numerically distinct from, but possibly similar to, the individual forms associated with numerically distinct concrete particular objects. Thirdly, I recommended that hylomorphists should opt for a compositional approach to the relation between a matter–form compound and its form. As a result, continuing in the vein of the neo-Aristotelian structure-based mereology laid out in Koslicki (2008a), I advocate that matter and form count literally and strictly speaking as proper parts of the concrete particular object they compose, according to a single notion of parthood. Finally, I also urged, due to the very real challenges arising from the Grounding Problem, that hylomorphists should adopt a robust construal of form, according to which forms do not bear the very same relation both to the matter–form compound (essentially) and to the matter composing it (accidentally). In Part II of this study, I turned to the question of whether concrete particular objects, when construed as matter–form compounds, nevertheless deserve to be assigned a privileged status within a hylomorphic ontology, and, if so, according to what criterion or criteria of “ontological privilege.” This question is particularly pressing for those who find themselves drawn to the following three tenets: (i) a hylomorphic analysis of concrete particular objects; (ii) an independence criterion of substancehood; and (iii) the desire to assign substance status to at least some hylomorphic compounds. These three commitments, when taken together, appear to lead to the following conflict: given their metaphysical complexity, one wonders whether hylomorphic compounds will not turn out to be ontologically dependent on other entities numerically distinct from them, viz., their form and possibly their matter as well, thereby apparently jeopardizing their inclusion in the category of substances, when these are conceived of as ontologically independent entities. My main aim in the second part of this inquiry was to present my own proposal as to how Aristotelians should proceed if they wish to resolve the apparent tension just noted in a way that is maximally consistent with their other central commitments. I argued that in order for a criterion of substancehood to yield the desired results when applied to hylomorphic compounds, a unity criterion of substancehood for composite entities is more suitable to the task at hand than an independence criterion, despite a general preference among Aristotelians for the latter. I offered an account of unity designed to capture the idea that hylomorphic compounds are more unified than other types of composite entities (e.g., heaps or non-empty sets). I ended with some preliminary remarks concerning the status of artifacts within a hylomorphic ontology, given their apparent dependence on the mental states of intentional agents who invent, design, produce, or use them. Various questions, which require further investigation, were left open along the way. I mention a few important points here to indicate fruitful directions for future research. Firstly, the present study, which concerns itself purely with a particular application of the doctrine of hylomorphism to the specific case of concrete particular objects, does not settle how, and whether, the hylomorphic apparatus should be extended to other problem areas within or outside of metaphysics. Some illustrations of how the doctrine

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

conclusion  243 of hylomorphism has been usefully applied across a wide range of sub-disciplines of philosophy were mentioned in the Introduction. Secondly, according to the hylomorphic conception of matter proposed in Chapter 2, the matter of a concrete particular object is nothing other than its material parts and these, or so I argued, are themselves best conceived of as matter–form compounds. However, as noted in connection with this proposal, the hylomorphic conception of matter is subject to an important proviso: for should we ever encounter an empirically confirmed level in the compositional hierarchy at which we have reason to believe that the material parts composing a concrete particular object are no longer themselves structured wholes, then, so I have proposed, hylomorphists should ­concede that their apparatus is at that point no longer applicable. I take it to be an as-of-yet unanswered empirical question whether the compositional hierarchy in fact includes such an unstructured level. In addition, further study is required to determine how metaphysicians should best respond to such a scenario, if it in fact were to be empirically confirmed. Thirdly, I argued in Chapters 3 and 4 that forms are best construed as robust individual or particular (i.e., non-shareable, non-repeatable, non-multiply located) entities. And while this designation narrows down the range of available options somewhat, there are still further choices to be made by hylomorphists when it comes to the assignment of forms to their proper ontological category. For this version of the individual forms hypothesis is still compatible, for example, with the assignment of forms to the ontological category of objects, properties or relations, states, functions, powers, activities or processes, facts, actions or sui generis entities, as long as the occupants of the relevant category are construed as robust particulars. The observations brought to bear on this question in the preceding chapters (viz., the cross-world identification of concrete particular objects as well as the grounding of their modal profile) do not by themselves single out a particular choice from among these admissible options, and we must await the introduction of further considerations by means of which these questions can be adjudicated in a principled manner. Next, a fourth and rather ambitious research project, which requires further attention, concerns the question of how best to construe the relation between essence and necessity given a non-modal (or, more generally, non-reductive) conception of essence. I mentioned in Chapter 4 that hylomorphists will need an answer to this question in order to provide a full response to the Grounding Problem. (The Grounding Problem poses the question of how to explain the apparent differences between an object and its matter with respect to their modal and other properties.) I have argued that the relevant consequence relation which connects propositions stating what is necessary but non-essential to objects to propositions stating what belongs to their essence proper cannot in general be logical entailment (see Koslicki (2012b), and Section 5.3.2 in this work). To illustrate, a triangle’s being three-sided is not logically entailed by its being three-angled, unless additional premises are added which take the relationship in question for granted and therefore make the derivation in question

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

244  Conclusion trivial and hence philosophically uninteresting, e.g., that every closed geometrical figure with three angles also has three sides and that triangles are closed geometrical figures.1 The question of how best to account for the relevant non-logical asymmetric explanatory connection between basic non-modal facts about the essences of objects and derivative facts about their modal profile still awaits a philosophically satisfactory treatment.2 Finally, and relatedly, I noted in connection with the discussion of artifacts in Chapter  8 that the status of these apparently mind-dependent entities within a hylomorphic ontology cannot be fully resolved until existing essentialist and antiessentialist frameworks are further refined and developed to accommodate their special features. We saw that prominent author-intention-based theories of artifact essences on the one hand, are susceptible to the concern that they exaggerate the creative and discriminating power of human intentions. Existing anti-essentialist frameworks on the other hand, tend to trace the ascriptions of modal features to objects back to our semantic, inferential, or explanatory practices and are hence not particularly well suited to capture the primarily practical and action-based orientation of our engagement with the realm of artifacts. For the time being, the special case of artifacts thus eludes an entirely satisfactory treatment and a proper clarification of their status within a hylomorphic ontology still awaits further research on the part of essentialists and anti-essentialists alike. Overall, the conclusions reached in Parts I and II together allow us to classify ­matter–form compounds in a comparative way, as more deserving of substance status than other types of composite entities, due at least in part to their highly unified nature. As I have emphasized elsewhere (see Koslicki (2015a, 2016a)), such a comparative classification of entities is not to be confused with a categorization of these entities as absolutely fundamental or as substances simpliciter. At the same time, however, the importance of assigning matter–form compounds to their proper position within the hierarchy of the relatively fundamental should not be underestimated. As we have seen throughout this study, many of the most interesting metaphysical projects, especially those dear to Aristotelians, take place within the sphere of what, in absolute terms, would count as ontologically derivative. In particular, in the preceding chapters, we have encountered reasons for thinking that multiple explanatory factors are relevant to the classification of an entity or type of entity as more or less fundamental (or derivative) than another, e.g., its status as more or less unified, mind-independent, or constructionally complex than other comparable entities or types of entities. Given the multidimensional approach to relative fundamentality I favor, one and the same 1   Here, I take a different approach from that defended in Fine (1994a, 1995a, 1995b, 1995c) as well as Hale (2015), where it is assumed that the relevant consequence relation connecting what is essential proper to what is merely metaphysically necessary is that of logical entailment. 2   Lest it be thought that the notion of ground can be expected to take on this explanatory burden, see my arguments in Koslicki (2015a) to the effect that this notion is not sufficiently fine-grained to illuminate the nature of the explanatory connections it is intended to capture. See also Koslicki (2016b) for arguments against understanding the alleged analogy between grounding and causation in terms of the formalism utilized by structural equation models of causation (see Schaffer (2016)).

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

conclusion  245 entity can be classified, without giving rise to contradictions, as more fundamental than another in certain respects (e.g., its constructional complexity) but not in others (e.g., the extent to which it is unified or mind-independent). In its broadest statement, then, the aim of this inquiry has been, firstly, to argue for a hylomorphic conception of concrete particular objects; and, secondly, to illustrate and recommend a particular approach to meta-ontology or meta-metaphysics, viz., one which, while remaining neutral with respect to questions concerning absolute fundamentality, recognizes that multiple dimensions of relative fundamentality and derivativeness are needed to do justice to the full range of data which presents itself to those engaged in the study of being.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

Bibliography Adams, Robert (1979). “Primitive Thisness and Primitive Identity,” The Journal of Philosophy, Vol. 76, No. 1 (January 1979), pp. 5–26. Alanen, Lilli and Witt, Charlotte (eds) (2004). Feminist Reflection on the History of Philosophy. Kluwer Academic Publisher: Dordrecht. Anagnostopoulos, George (ed.) (2009). A Companion to Aristotle. Wiley-Blackwell: Chichester/ Malden, MA. Armstrong, David M. (1978). Universals and Scientific Realism, Vols. I and II. Cambridge University Press: Cambridge. Armstrong, David M. (1986). “In Defense of Structural Universals,” Australasian Journal of Philosophy, Vol. 64, No.1, pp. 85–96. Armstrong, David M. (1989). Universals: An Opinionated Introduction. Westview: Boulder, CO. Armstrong, David M. (1991). “Classes are States of Affairs,” Mind, Vol. 100, pp. 189–200. Armstrong, David M. (1997). A World of States of Affairs. Cambridge University Press: Cambridge. Armstrong, David M. (2004). Truth and Truthmakers. Cambridge University Press: Cambridge. Audi, Paul (2012a). “A Clarification and Defense of the Notion of Grounding,” in Fabrice Correia and Benjamin Schnieder (eds), Metaphysical Grounding: Understanding the Structure of Reality. Cambridge University Press: Cambridge, pp. 101–21. Audi, Paul (2012b). “Grounding: Toward a Theory of the In-Virtue-Of Relation,” Journal of Philosophy, Vol. 109, No. 12 (Dec. 2012), pp. 685–711. Bader, Ralf (2010). “Neo-Aristotelian Mereology,” Metascience, Vol. 19, pp. 413–15. Bailey, Andrew (2012). “No Bare Particulars,” Philosophical Studies, Vol. 158, pp. 31–41. Baker, Lynne Rudder (2000). Persons and Bodies: A Constitution View. Cambridge University Press: Cambridge. Baker, Lynne Rudder (2002). “On Making Things Up: Constitution and Its Critics,” Philosophical Topics, Vol. 30, No. 1, Identity and Individuation (Spring 2002), pp. 31–51. Baker, Lynne Rudder (2004). “The Ontology of Artifacts,” in Klaus Petrus (ed.), On Human Persons. Ontos Verlag: Frankfurt, pp. 23–39. Baker, Lynne Rudder (2007). The Metaphysics of Everyday Life: An Essay in Practical Realism. Cambridge University Press: New York. Baracchi, Claudia (ed.) (2014). The Bloomsbury Companion to Aristotle Bloomsbury: London. Barnes, Elizabeth (ed.) (2017). Current Controversies in Metaphysics. Routledge: New York. Barnes, Jonathan (1984). The Complete Works of Aristotle: The Revised Oxford Translation, Jonathan Barnes (ed.), Vols One and Two, Bollingen Series LXXI.2. Princeton University Press: Princeton, NJ. Barney, Rachel (2008). “Aristotle’s Argument for a Human Function,” Oxford Studies in Ancient Philosophy, Vol. 34, pp. 293–322. Bar On, Bat-Ami (ed.) (1994). Engendering Origins: Critical Feminist Readings in Plato and Aristotle. State University of New York Press: Albany, NY. Bennett, Karen (2004). “Spatio-temporal Coincidence and the Grounding Problem,” Philosophical Studies, Vol. 118, No. 3, pp. 339–71.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

248  bibliography Bennett, Karen (2011a). “Construction Area: No Hard Hat Required,” Philosophical Studies, Vol. 154, pp. 79–104. Bennett, Karen (2011b). “By Our Bootstraps,” Philosophical Perspectives, Vol. 25, No. 1, pp. 27–41. Bennett, Karen (2011c). “Koslicki on Formal Proper Parts,” Analysis, Vol. 71, No. 2 (April 2011), pp. 1–5. Bennett, Karen (2017). Making Things Up. Oxford University Press: Oxford. Bigaj, Tomasz and Wüthrich, Christian (eds) (2016). Metaphysics in Contemporary Physics, Pozna Studies in the Philosophy of the Sciences and the Huma (Book 104). Brill Rodopi: Boston, MA/Leiden. Bird, Alexander (2007). Nature’s Metaphysics: Laws and Properties. Oxford University Press: New York. Bird, Alexander and Tobin, Emma (2018). “Natural Kinds,” in The Stanford Encyclopedia of Philosophy (spring 2018 edn), Edward N. Zalta (ed.), , last accessed March 22, 2018. Black, Max (1952). “The Identity of Indiscernibles,” Mind, Vol. 61, No. 242 (April 1952), pp. 153–64. Blatti, Stephan and Lapointe, Sandra (eds) (2016). Ontology After Carnap. Oxford University Press: Oxford. Bostock, David (ed.) (1994). Aristotle, Metaphysics, Books Z and H, trans. with commentary by Bostock, Clarendon Press: Oxford. Bottani, Andrea, Carrara, Massimiliano, and Giaretta, Pierdaniele (eds) (2002). Individuals, Essence and Identity. Kluwer Academic Publishers: Dordrecht. Brading, Katherine and Castellani, Elena (eds) (2003). Symmetries in Physics: Philosophical Reflections. Cambridge University Press: Cambridge. Britton, Teresa (2012). “The Limits of Hylomorphism,” Metaphysica, Vol. 13, Issue 2 (October 2012), pp. 145–53. Brower, Jeffrey (2010). “Aristotelian Endurantism: A New Solution to the Problem of Temporary Intrinsics,” Mind, Vol. 119, No. 476, pp. 883–905. Brower, Jeffrey (2012). “Matter, Form and Individuation,” in Brian Davies and Eleonore Stump (eds), The Oxford Handbook to Aquinas. Oxford University Press: New York, pp. 85–103. Brower, Jeffrey (2014). Aquinas’s Ontology of the Material World: Change, Hylomorphism, and Material Objects. Oxford University Press: Oxford. Brower, Jeffrey (2016). “Aquinas on the Problem of Universals,” Philosophy and Phenomenological Research, Vol. 92, No. 3 (May 2016), pp. 715–35. Brower, Jeffrey (2017). “Aquinas on the Individuation of Substances,” Oxford Studies in Medieval Philosophy, Vol. 5, pp. 122–50. Burkhardt, Hans, Seibt, Johanna, and Imaguire, Guido (eds) (2017). The Handbook of Mereology. Philosophia Verlag: München. Burnyeat, Myles (2008). Aristotle’s Divine Intellect, The Aquinas Lecture. Marquette University Press: Milwaukee, WI. Campbell, Keith (1990). Abstract Particulars. Basil Blackwell: Oxford. Carnap, Rudolf (1950). “Empiricism, Semantics and Ontology,” Revue Internationale de Philosophie, Vol. 4, pp. 20–40. Carrara, Massimiliano and Vermaas, Pieter (2009). “The Fine-Grained Metaphysics of Artifactual and Biological Functional Kinds,” Synthese, Vol. 169, pp. 125–43.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

bibliography  249 Chalmers, David (2012). Constructing the World. Oxford University Press: Oxford. Chalmers, David, Manley, David, and Wasserman, Ryan (eds) (2009). Metametaphysics: New Essays on the Foundations of Ontology. Clarendon Press: Oxford. Charlton, W. (1970). Aristotle’s Physics, Books I and II, Clarendon Aristotle Series, trans. with introduction and notes. Clarendon Press: Oxford. Code, Alan (1984). “The Aporematic Approach to Primary Being in Metaphysics Z,” in Francis Jeffry Pelletier and John King-Farlow (eds), New Essays on Aristotle, Canadian Journal of Philosophy, Supplementary Volume X, pp. 1–20. Cohen, S. Marc (2009). “Substances,” in George Anagnostopoulos (ed.), A Companion to Aristotle. Wiley-Blackwell: Chichester/Malden, MA, pp. 197–212. Connell, Sophia M. (2016). Aristotle on Female Animals: A Study of the Generation of Animals. Cambridge University Press: Cambridge. Corcilius, Klaus and Gregoric, Pavel (2010). “Separability vs. Difference: Parts and Capacities of the Soul in Aristotle,” Oxford Studies in Ancient Philosophy, Vol. 39, pp. 81–119. Corkum, Phil (2008). “Aristotle on Ontological Dependence,” Phronesis, Vol. 53, pp. 65–92. Corkum, Phil (2013). “Substance and Independence in Aristotle,” in Miguel Hoeltje, Benjamin Schnieder, and Alex Steinberg (eds), Varieties of Dependence: Ontological Dependence, Grounding, Supervenience, Response-Dependence. Philosophia Verlag: München, pp. 65–95. Correia, Fabrice (2004). “Husserl on Foundation,” Dialectica, Vol. 58 (No. 3), pp. 349–67. Correia, Fabrice (2005). Existential Dependence and Cognate Notions. Philosophia Verlag: München. Correia, Fabrice (2008). “Ontological Dependence,” Philosophy Compass, Vol. 3, No. 5, pp. 1013–32. Correia, Fabrice (2013). “Metaphysical Grounds and Essence,” in Miguel Hoeltje, Benjamin Schnieder, and Alex Steinberg (eds), Varieties of Dependence: Ontological Dependence, Grounding, Supervenience, Response-Dependence. Philosophia Verlag: München, pp. 271–96. Correia, Fabrice and Schnieder, Benjamin (eds) (2012). Metaphysical Grounding: Understanding the Structure of Reality. Cambridge University Press: Cambridge. Cotnoir, Aaron J. and Baxter, Donald L. M. (eds) (2014). Composition as Identity. Oxford University Press: Oxford. Cross, Richard (2003). “Medieval Theories of Haecceity,” in The Stanford Encyclopedia of Philosophy (summer 2014 edn), Edward N. Zalta (ed.), , last accessed March 22, 2018. Davies, Brian and Stump, Eleonore (eds) (2012). The Oxford Handbook to Aquinas. Oxford University Press: New York. Deleuze, Gilles and Guattari, Félix (1987). A Thousand Plateaus: Capitalism and Schizophrenia, trans. Brian Massumi. University of Minnesota Press: Minneapolis. deRosset, Louis (2011). “What is the Grounding Problem?,” Philosophical Studies: An International Journal for Philosophy in the AnalyticTradition, Vol. 156, No. 2 (November 2011), pp. 173–97. Diekemper, Joseph (2009). “Thisness and Events,” Journal of Philosophy, Vol. 106, No. 5, pp. 255–76. Diekemper, Joseph (2015). “The Ontology of Thisness,” Philosophy and Phenomenological Research, Vol. 90, No. 1, pp. 49–71. Divers, John (2008). “Coincidence and Form,” Proceedings of the Aristotelian Society, Supple­ mentary Volume 82, pp. 119–37.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

250  bibliography Donnelly, Maureen (2011). “Using Mereological Principles to Support Metaphysics,” Philosophical Quarterly, Vol. 61, No. 243 (April 2011), pp. 225–46. Dumitru, Mircea (ed.) (2018). Metaphysics, Meaning and Modality: Themes from Kit Fine. Oxford University Press: Oxford. Elder, Crawford (2007). “On the Place of Artifacts in Ontology,” in Eric Margolis and Stephen Laurence (eds), Creations of the Mind: Theories of Artifacts and Their Representation. Oxford University Press: New York, pp. 33–51. Evnine, Simon (2011). “Constitution and Composition: Three Approaches to Their Relation,” Protosociology, Vol. 27 (Modernization in Times of Globalization II), pp. 212–35. Evnine, Simon (2013). “Ready-Mades: Ontology and Aesthetics,” British Journal of Aesthetics, Vol. 53, No. 4 (October 2013), pp. 407–23. Evnine, Simon (2016a). Making Objects and Events: A Hylomorphic Theory of Artifacts, Actions, and Organisms. Oxford University Press: Oxford. Evnine, Simon (2016b). “Much Ado about Something-from-Nothing: or, Problems for Ontological Minimalism,” in Stephan Blatti and Sandra Lapointe (eds), Ontology After Carnap. Oxford University Press: Oxford, pp. 145–64. Evnine, Simon (2018). “Mass Production,” in Bill Brewer and Javier Cumpa (eds), The Nature of Ordinary Objects. Cambridge University Press: Cambridge. Feser, Edward (ed.) (2013). Aristotle on Method and Metaphysics. Palgrave Macmillan: Basingstoke. Fine, Kit (1982). “Acts, Events and Things,” in Werner Leinfellner, Eric Kraemer, and Jeffrey Schank (eds), Language and Ontology, Proceedings of the Sixth International Wittgenstein Symposium, Volume 8 of Schriftenreihe der Wittgenstein-Gesellschaft. Hölder-PichlerTempsky: Vienna, pp. 97–105. Fine, Kit (1994a). “Essence and Modality,” Philosophical Perspectives, Vol. 8 (Logic and Language), pp. 1–16. Fine, Kit (1994b). “A Puzzle Concerning Matter and Form,” in Theodore Scaltsas, David Charles, and Mary Louise Gill (eds), Unity, Identity and Explanation in Aristotle’s Metaphysics. Clarendon Press: Oxford, pp. 13–40. Fine, Kit (1995a). “Ontological Dependence,” Proceedings of the Aristotelian Society, Vol. 95, pp. 269–90. Fine, Kit (1995b). “Senses of Essence,” in Walter Sinnott-Armstrong, Diana Raffman, and Nicholas Asher (eds), Modality, Morality, and Belief, Essays in Honor of Ruth Barcan Marcus. Cambridge University Press: New York, pp. 53–73. Fine, Kit (1995c). “The Logic of Essence,” Journal of Philosophical Logic, Vol. 24, pp. 241–73. Fine, Kit (1999). “Things and Their Parts,” Midwest Studies in Philosophy, Vol. 23, pp. 61–74. Fine, Kit (2000). “A Counter-example to Locke’s Thesis,” The Monist, Vol. 83, No. 3 (Temporal Parts) (July 2000), pp. 357–61. Fine, Kit (2001). “The Question of Realism,” Philosophers’ Imprint, Vol. 1, No. 1 (June 2001), , last accessed March 22, 2018, pp. 1–30. Fine, Kit (2008). “Coincidence and Form,” Proceedings of the Aristotelian Society, Supplementary Vol. 82, pp. 101–18. Fine, Kit (2010). “Towards a Theory of Part,” Journal of Philosophy, Vol. 107, No. 11 (November 2010), pp. 559–89. Fine, Kit (2012). “A Guide to Ground,” in Fabrice Correia and Benjamin Schnieder (eds), Metaphysical Grounding: Understanding the Structure of Reality. Cambridge University Press: Cambridge, pp. 37–80.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

bibliography  251 Forbes, Graeme (1985). The Metaphysics of Modality, Clarendon Library of Logic and Philosophy series. Oxford University Press: Oxford. Forbes, Graeme (1986). “In Defense of Absolute Essentialism,” Midwest Studies in Philosophy, Vol. 11 (Studies in Essentialism), P. A. French, T. E. Uehling, and H. K. Wettstein (eds). University of Minnesota Press: Minneapolis, MN, pp. 3–31. Forbes, Graeme (1997). “Essentialism,” in Bob Hale and Crispin Wright (eds), Companion to the Philosophy of Language. Blackwell: Malden, MA, pp. 515–33. Forbes, Graeme (2002). “Origins and Identities,” in Andrea Bottani, Massimiliano Carrara, and Pierdaniele Giaretta (eds), Individuals, Essence and Identity. Kluwer Academic Publishers: Dordrecht, pp. 319–40. Fraenkel, Abraham A., Bar-Hillel, Yehoshua, and Levy, Azriel (1973). Foundations of Set Theory, 2nd rev. edn, in collaboration with Dirk van Dalen. Noord-Hollandsche U.M.: Amsterdam. Frede, Michael (1985). “Substance in Aristotle’s Metaphysics,” repr. in Michael Frede (1987b), Essays in Ancient Philosophy. University of Minnesota Press: Minneapolis, MN, pp. 72–80. Frede, Michael (1987a). “Individuals in Aristotle,” in Michael Frede (1987b), Essays in Ancient Philosophy. University of Minnesota Press: Minneapolis, MN, pp. 49–71. Frede, Michael (1987b). Essays in Ancient Philosophy. University of Minnesota Press: Minneapolis, MN. Frede, Michael (2000). “Introduction,” in Michael Frede and David Charles (eds), Aristotle’s Metaphysics Lambda, Symposium Aristotelicum. Oxford University Press: New York, pp. 1–53. Frede, Michael (2008). “Aristotle on Thinking,” Rhizai, Vol. 2, pp. 287–301. Frede, Michael and Charles, David (eds) (2000). Aristotle’s Metaphysics Lambda, Symposium Aristotelicum. Oxford University Press: New York. Frede, Michael and Patzig, Günther (1988). Aristoteles Metaphysik Z. Text, Übersetzung und Kommentar, Vols 1 and 2. Verlag C.H. Beck: Munich. Freeland, Cynthia A. (ed.) (1998). Feminist Interpretations of Aristotle. Pennsylvania State University Press: University Park. Frege, Gottlob (1953). The Foundations of Arithmetic: A Logico-Mathematical Enquiry into the Concept of Number, trans. J. L. Austin, 2nd edn, Blackwell: Oxford. Garcia-Valdecasas, Miguel, Murillo, José Ignacio, and Barrett, Nathaniel (eds) (2016). Biology and Subjectivity: Philosophical Contributions to Non-Reductive Neuroscience. Springer: Cham. Gibbard, Allan (1975). “Contingent Identity,” Journal of Philosophical Logic, Vol. 4, pp. 187–221. Gill, Mary Louise (1989). Aristotle on Substance: The Paradox of Unity. Princeton University Press: Princeton, NJ. Gorman, Michael (2005). “The Essential and the Accidental,” Ratio (New Series), Vol. 18, No. 3 (September 2005), pp. 276–89. Gorman, Michael (2006a). “Substance and Identity-Dependence,” Philosophical Papers, Vol. 35, No. 1, pp. 103–18. Gorman, Michael (2006b). “Independence and Substance,” International Philosophical Quarterly, Vol. 46, No. 2, Issue 182 (June 2006), pp. 147–59. Gorman, Michael (2012). “On Substantial Independence: A Reply to Patrick Toner,” Philosophical Studies, Vol. 159, No. 2 (June 2012), pp. 293–7. Groff, Ruth and Greco, John (eds) (2013). Powers and Capacities in Philosophy: The New Aristotelianism. Routledge: New York. Hale, Bob (2015). Necessary Beings: An Essay on Ontology, Modality, and the Relations Between Them. Oxford University Press: New York.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

252  bibliography Hale, Bob and Hoffman, Aviv (2010). Modality: Metaphysics, Logic, and Epistemology. Oxford University Press: Oxford. Hale, Bob and Wright, Crispin (eds) (1997). Companion to the Philosophy of Language. Blackwell: Malden, MA. Harte, Verity (2002). Plato on Parts and Wholes: The Metaphysics of Structure. Clarendon Press: Oxford. Hashemi, Ataollah (2017). “How to Deal with the Puzzle of Coincident Objects,” unpublished Masters Thesis, University of Alberta, Edmonton, AB. Hawley, Katherine (2006). “Weak Discernibility,” Analysis, Vol. 66, No. 4 (October 2006), pp. 300–3. Hawley, Katherine (2010). Review of The Structure of Objects, International Studies in the Philosophy of Science, Vol. 24, No. 3, pp. 336–9. Hazen, Allen (1991). “Small Sets,” Philosophical Studies, Vol. 63, pp. 119–23. Hempel, Carl G. and Oppenheim, Paul (1948). “Studies in the Logic of Explanation,” Philosophy of Science, Vol. 15, No. 2. (April 1948), pp. 135–75. Henry, Devin (2006a). “Aristotle on the Mechanism of Inheritance,” Journal of the History of Biology, Vol. 39, No. 3, pp. 425–55. Henry, Devin (2006b). “Understanding Aristotle’s Reproductive Hylomorphism,” Apeiron, Vol. 39, No. 3, pp. 257–87. Henry, Devin and Nielsen, Karen Margrethe (eds) (2015). Bridging the Gap between Aristotle’s Science and Ethics. Cambridge University Press: Cambridge. Hershenov, David B. and Koch, Rose J. (2005). “How a Hylomorphic Metaphysics Constrains the Abortion Debate,” The National Catholic Bioethics Quarterly, Vol. 5, No. 4, pp. 751–64. Hilpinen, Risto (2011). “Artifact,” in The Stanford Encyclopedia of Philosophy (winter 2011 edn), Edward N. Zalta (ed), , last accessed March 22, 2018. Hirsch, Eli (2002). “Quantifier Variance and Realism,” Philosophical Issues, Vol. 12, pp. 51–73. Hoeltje, Miguel, Schnieder, Benjamin, and Steinberg, Alex (eds) (2013). Varieties of Dependence: Ontological Dependence, Grounding, Supervenience, Response-Dependence. Philosophia Verlag: München. Hoffman, Joshua and Rosenkrantz, Gary (1994). Substance Among Other Categories. Cambridge University Press: New York. Hoffman, Joshua and Rosenkrantz, Gary (1997). Substance: Its Nature and Existence. Routledge: London. Hoffman, Joshua and Rosenkrantz, Gary (1999). “On the Unity of Compound Things: Living and Non-Living,” in David Oderberg (ed.), Form and Matter: Themes in Contemporary Metaphysics. Blackwell: Oxford, pp. 76–102. Hovda, Paul (2009). Review of The Structure of Objects, Notre Dame Philosophical Reviews, April 26, 2009. Humberstone, I. Lloyd (1996). “Intrinsic/Extrinsic,” Synthese, Vol. 108, pp. 205–67. Hume, David (2007). A Treatise on Human Nature: A Critical Edition, David Fate Norton and Mary J. Norton (eds). Clarendon, Oxford University Press: Oxford. Husserl, Edmund (1900–1). Logische Untersuchungen, 1st edn. M. Niemeyer: Halle. Ingarden, Roman (1925). “Essentiale Fragen: Ein Beitrag zum Problem des Wesens,” Jahrbuch für Philosophie und Phänomenologische Forschung, Sonderausdruck, Vol. 7, Edmund Husserl (ed.). Max Niemeyer: Halle, pp. 125–304.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

bibliography  253 Ingarden, Roman (1960). Das Literarische Kunstwerk, mit einem Anhang (Von den Funktionen der Sprache im Theaterschauspiel), 2nd edn. Max Niemeyer Verlag: Tübingen. Ingarden, Roman (1965). Der Streit um die Existenz der Welt, Vols. I–II. Max Niemeyer Verlag: Tübingen. Ingarden, Roman (1974). Der Streit um die Existenz der Welt, Vol. III: Über die Kausale Struktur der Realen Welt. Max Niemeyer Verlag: Tübingen. Inman, Ross (2011). Review of The Structure of Objects, Philosophia Christi, Vol. 13, No. 1, pp. 219–23. Ismael, Jenann and Schaffer, Jonathan (2016). “Quantum Holism: Non-Separability as Common Ground,” Synthese, published online September 9, 2016, DOI 10.1007/s11229-016-1201-2. Jago, Mark (ed.) (2016a). Reality Making. Oxford University Press: Oxford. Jago, Mark (2016b). “Essence and the Grounding Problem,” in Mark Jago (ed.), Reality Making. Oxford University Press: Oxford, pp. 99–120. Jaworski, William (2014). “Hylomorphism and the Metaphysics of Structure,” Res Philosophica, Vol. 91, No. 2 (April 2014), pp. 179–201. Jaworski, William (2016a). “Hylomorphism: Emergent Properties without Emergentism,” in Miguel Garcia-Valdecasas, José Ignacio Murillo, and Nathaniel Barrett (eds), Biology and Subjectivity: Philosophical Contributions to Non-Reductive Neuroscience. Springer: Cham, pp. 41–61. Jaworski, William (2016b). Structure and the Metaphysics of Mind: How Hylomorphism Solves the Mind-Body Problem. Oxford University Press: New York. Jenkins, Carrie (2011). “Is Metaphysical Dependence Irreflexive?,” The Monist, Vol. 94, No. 2, pp. 267–76. Johansen, Thomas Kjeller (2012). The Powers of Aristotle’s Soul. Oxford University Press: Oxford. Johnson, Monte Ransome (2005). Aristotle on Teleology. Oxford University Press: New York. Johnston, Mark (2002). “Parts and Principles: False Axioms in Mereology,” Philosophical Topics, Vol. 30, No. 1, pp. 129–66. Johnston, Mark (2006). “Hylomorphism,” The Journal of Philosophy, Vol. 103, No. 12, pp. 652–98. Jubien, Michael (1993). Ontology, Modality, and the Fallacy of Reference. Cambridge University Press: New York. Kanzian, Christian (2009). Ding – Substanz – Person: Eine Alltagsontologie. Philosophische Analyse, Herbert Hochberg, Rafael Hüntelmann, Christian Kanzian, Richard Schantz, and Erwin Tegtmeier (eds), Vol. 33. Ontos Verlag: Frankfurt. Kanzian, Christian (2013). “Hylomorphism and Jonathan Lowe’s Ontology of Human Persons,” in Miroslaw Szatkowski (ed.), Dualistic Ontology of the Human Person. Philosophia: München, pp. 129–48. Kanzian, Christian (2015). “Existential Dependence and other Formal Relations,” in Miroslaw Szatkowski (ed.), God, Truth, and Other Enigmas. De Gruyter (Philosophical Analysis): Berlin, pp. 183–96. Kanzian, Christian (2016). Wie die Dinge Sind: Noch eine Alltagsontologie. Philosophische Analyse, Herbert Hochberg, Rafael Hüntelmann, Christian Kanzian, Richard Schantz, and Erwin Tegtmeier (eds), Vol. 66. De Gruyter: Berlin. Keller, John A. (ed.) (2016). Freedom, Metaphysics, and Method: Themes from van Inwagen. Oxford University Press: Oxford. Kim, Jaegwon (1984). “Concepts of Supervenience,” Philosophy and Phenomenological Research, Vol. 45, pp. 153–76.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

254  bibliography Kim, Jaegwon (1990). “Supervenience as a Philosophical Concept,” Metaphilosophy, Vol. 21, pp. 1–27, repr. in Jaegwon Kim (1993). Supervenience and Mind: Selected Philosophical Essays, Cambridge Studies in Philosophy. Cambridge University Press: New York, pp. 131–60. King, Hugh R. (1956). “Aristotle Without Prima Materia,” Journal for the History of Ideas, Vol. 17, pp. 370–89. Kleinschmidt, Shieva (2007). “Some Things About Stuff,” Philosophical Studies, Vol. 135, pp. 407–23. Koons, Robert (2014). “Staunch vs. Faint-Hearted Hylomorphism: Toward an Aristotelian Account of Composition,” Res Philosophica, Vol. 91, No. 2 (April 2014), pp. 151–77. Kornblith, Hilary (1980). “Referring to Artifacts,” Philosophical Review, Vol. 89, pp. 109–14. Kornblith, Hilary (2007). “How to Refer to Artifacts,” in Eric Margolis and Stephen Laurence (eds), Creations of the Mind: Theories of Artifacts and Their Representation. Oxford University Press: New York, pp. 138–49. Korsgaard, Christine M. (2008). The Constitution of Agency: Essays on Practical Reason and Moral Psychology. Oxford University Press: Oxford. Koslicki, Kathrin (1995). “Talk about Stuffs and Things: The Logic of Mass and Count Nouns,” unpublished PhD diss., Massachusetts Institute of Technology, Cambridge, MA. Koslicki, Kathrin (1997). “Isolation and Non-Arbitrary Division: Frege’s Two Criteria for Counting,” Synthese, Vol. 112, No. 3 (September 1997), pp. 403–30. Koslicki, Kathrin (1999a). “The Semantics of Mass-Predicates,” Nous, Vol. 33, No. 1 (March 1999), pp. 46–91; repr. in The Philosopher’s Annual, Vol. 22, pp. 101–54. Koslicki, Kathrin (1999b). “Genericity and Logical Form,” Mind and Language, Vol. 14, No. 4 (December 1999), pp. 441–67. Koslicki, Kathrin (2003). “The Crooked Path from Vagueness to Four-Dimensionalism,” Philosophical Studies, Vol. 114, No. 1–2 (May 2003), pp. 107–34. Koslicki, Kathrin (2005). “Almost Indiscernible Objects and the Suspect Strategy,” Journal of Philosophy, Vol. 102, No. 2, pp. 55–77. Koslicki, Kathrin (2006). “Aristotle’s Mereology and the Status of Form,” Journal of Philosophy, Special Issue: “Parts and Wholes,” Wolfgang Mann and Achille Varzi (eds), Vol. 103, No. 12 (December 2006), pp. 715–36. Koslicki, Kathrin (2007a). “Review of Henry Laycock, Words Without Objects: Semantics, Ontology, and Logic for Non-Singularity, Oxford University Press, 2006,” Australasian Journal of Philosophy, Vol. 85, No. 1 (2007), pp. 160–3. Koslicki, Kathrin (2007b). “Towards a Neo-Aristotelian Mereology,” Dialectica, Special Issue: “The Philosophy of Kit Fine,” Kevin Mulligan (guest ed.), Vol. 61, No. 1, pp. 127–59. Koslicki, Kathrin (2008a). The Structure of Objects. Oxford University Press: Oxford. Koslicki, Kathrin (2008b). “Natural Kinds and Natural Kind Terms,” Philosophy Compass, Vol. 3/4 (2008), pp. 789–802. Koslicki, Kathrin (2012a). “Varieties of Ontological Dependence,” in Fabrice Correia and Benjamin Schnieder (eds), Metaphysical Grounding: Understanding the Structure of Reality. Cambridge University Press: Cambridge, pp. 186–213. Koslicki, Kathrin (2012b). “Essence, Necessity and Explanation,” in Tuomas Tahko (ed.), Contemporary Aristotelian Metaphysics. Cambridge University Press: Cambridge, pp. 187–206. Koslicki, Kathrin (2013a). “Ontological Dependence: An Opinionated Survey,” in Miguel Hoeltje, Benjamin Schnieder, and Alex Steinberg (eds), Varieties of Dependence: Ontological

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

bibliography  255 Dependence, Grounding, Supervenience, Response-Dependence. Philosophia Verlag: München, pp. 31–64. Koslicki, Kathrin (2013b). “Substance, Independence and Unity,” in Edward Feser (ed.), Aristotle on Method and Metaphysics. Palgrave Macmillan: Basingstoke, pp. 169–95. Koslicki, Kathrin (2014). “The Causal Priority of Form in Aristotle,” in Riin Sirkel and Tuomas Tahko (eds), Studia Philosophica Estonica, Special Issue: “Aristotelian Metaphysics. Essence and Ground,” Vol. 7.2, pp. 113–41. Koslicki, Kathrin (2015a). “The Coarse-Grainedness of Grounding,” Oxford Studies in Metaphysics, Vol. 9 (March 2015), pp. 306–44. Koslicki, Kathrin (2015b). “In Defense of Substance,” in Sandra Lapointe (ed.), Grazer Philosophische Studien, Festschrift in Honor of Peter Simons, Vol. 91 (2015), pp. 59–80. Koslicki, Kathrin (2016a). “Questions of Ontology,” in Stephan Blatti and Sandra Lapointe (eds), Ontology After Carnap. Oxford University Press: Oxford, pp. 220–41. Koslicki, Kathrin (2016b). “Where Grounding and Causation Part Ways: Comments on Jonathan Schaffer,” Philosophical Studies, Vol. 173, No. 1 (January 2016), pp. 101–12. Koslicki, Kathrin (2018a). “Essence and Identity,” in Mircea Dumitru (ed.), Metaphysics, Meaning and Modality: Themes from Kit Fine. Oxford University Press: Oxford. Koslicki, Kathrin (2018b). “Towards a Hylomorphic Solution to the Grounding Problem,” forthcoming in Royal Institute of Philosophy Supplements to Philosophy, Vol. 82 (2018), Metaphysics. Kosman, Aryeh (2013). The Activity of Being. Harvard University Press: Cambridge, MA. Kripke, Saul (1971). “Identity and Necessity,” in Milton Munitz (ed.), Identity and Individuation. New York University Press: New York, pp. 135–64. Kripke, Saul (1980). Naming and Necessity. Harvard University Press: Cambridge, MA. Originally published in Donald Davidson and Gilbert Harman (eds) (1972), Semantics of Natural Language. D. Reidel Publishing: Dordrecht, pp. 253–355 and pp. 763–9. Laurence, Stephen and Macdonald, Cynthia (eds) (1998). Contemporary Readings in the Foundations of Metaphysics. Blackwell: Oxford. Laycock, Henry (2006). Words Without Objects: Semantics, Ontology, and Logic for NonSingularity. Oxford University Press: New York. Lennox, James G. (2008). “ ‘As If We Were Investigating Snubness’: Aristotle on the Prospects for a Single Science of Nature,” Oxford Studies in Ancient Philosophy, Vol. 35, pp. 149–86. Lennox, James G. (2009). “Form, Essence and Explanation in Aristotle’s Biology,” in George Anagnostopoulos (ed.), A Companion to Aristotle. Wiley-Blackwell: Chichester/Malden, MA, pp. 348–67. Leonard, Henry and Goodman, Nelson (1940). “The Calculus of Individuals and Its Uses,” Journal of Symbolic Logic, Vol. 5, pp. 45–55. LePoidevin, Robin (ed.) (2008). Being: Developments in Contemporary Metaphysics. Cambridge University Press: Cambridge. LePoidevin, Robin, Simons, Peter, McGonigal, Andrew, and Cameron, Ross (eds) (2009). Routledge Companion to Metaphysics. Routledge: London. Lewis, David (1986). On the Plurality of Worlds. Basil Blackwel: Oxford. Lewis, David (1991). Parts of Classes. Basil Blackwell: Oxford. Lewis, David and Langton, Rae (1996). “Defining Intrinsic,” Philosophy and Phenomenological Research, Vol. 58, pp. 333–45.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

256  bibliography Lewis, Frank A. (1984). “What is Aristotle’s Theory of Essence?,” in Francis Jeffry Pelletier and John King-Farlow, New Essays on Aristotle, Canadian Journal of Philosophy, Supplementary Volume 10, pp. 89–131. Locke, John (1975). Essay Concerning Human Understanding, Peter H. Nidditsch (ed. and foreword). Oxford University Press: New York. Loux, Michael J. (1991). Primary Ousia: An Essay on Aristotle’s Metaphysics Z and H. Cornell University Press: Ithaca, NY. Loux, Michael J. (2002). Metaphysics: A Contemporary Introduction, 2nd edn. Routledge: London. Loux, Michael J. (2005). “Aristotle on Matter, Form, and Ontological Strategy,” Ancient Philosophy, Vol. 25, pp. 81–123. Loux, Michael (2006a). Metaphysics: A Contemporary Introduction, 3rd edn. Routledge: New York. Loux, Michael J. (2006b). “Aristotle’s Constituent Ontology,” in Dean Zimmerman (ed.), Oxford Studies in Metaphysics, Vol. 2. Oxford University Press: New York, pp. 207–50. Loux, Michael J. (2014). “Aristotle’s Hylomorphism,” in Daniel D. Novotný and Lukáš Novák (eds), Neo-Aristotelian Perspectives in Metaphysics. Routledge: New York, pp. 138–63. Lowe, E. J. (1989). “What Is a Criterion of Identity?,” Philosophical Quarterly, Vol. 39, No. 154, pp. 1–21. Lowe, E. J. (1994a). “Ontological Dependency,” Philosophical Papers, Vol. 23, No. 1, pp. 31–48. Lowe, E. J. (1994b). “Primitive Substances,” Philosophy and Phenomenological Research, Vol. 54, No. 3 (September 1994), pp. 531–52. Lowe, E. J. (1997). “Objects and Criteria of Identity,” in Bob Hale and Crispin Wright (eds), Companion to the Philosophy of Language. Blackwell: Malden, MA, pp. 613–33. Lowe, E. J. (1998). The Possibility of Metaphysics: Substance, Identity, and Time. Oxford University Press: Oxford. Lowe, E. J. (1999). “Form Without Matter,” in David Oderberg (ed.), Form and Matter: Themes in Contemporary Metaphysics. Blackwell: Oxford, pp. 1–21. Lowe, E. J. (2006). The Four-Category Ontology: A Metaphysical Foundation for Natural Science. Clarendon Press: Oxford. Lowe, E. J. (2007). “A Problem for a Posteriori Essentialism Concerning Natural Kinds,” Analysis, Vol. 67, No. 4 (October 2007), pp. 286–92. Lowe, E. J. (2008). “Two Notions of Being: Entity and Essence,” in Robin LePoidevin (ed.), Being: Developments in Contemporary Metaphysics. Cambridge University Press: Cambridge, pp. 23–48. Lowe, E. J. (2009). More Kinds of Being: A Further Study of Individuation, Identity, and the Logic of Sortal Terms. Wiley-Blackwell: Malden, MA/Oxford. Lowe, E. J. (2012a). “Asymmetrical Dependence in Inviduation,” in Fabrice Correia and Benjamin Schnieder (eds), Metaphysical Grounding: Understanding the Structure of Reality. Cambridge University Press: Cambridge, pp. 214–33. Lowe, E. J. (2012b). “A Neo-Aristotelian Substance Ontology: Neither Relational Nor Constituent,” in Tuomas Tahko (ed.), Contemporary Aristotelian Metaphysics. Cambridge University Press: Cambridge, pp. 229–48. Lowe, E. J. (2013). “Some Varieties of Metaphysical Dependence,” in Miguel Hoeltje, Benjamin Schnieder, and Alex Steinberg (eds), Varieties of Dependence: Ontological Dependence, Grounding, Supervenience, Response-Dependence. Philosophia Verlag: München, pp. 193–210.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

bibliography  257 MacBride, Fraser (2005). “The Universal-Particular Distinction: A Dogma of Metaphysics?,” Mind, Vol. 114, pp. 565–614. Macdonald, Cynthia (2005). Varieties of Things: Foundations of Contemporary Metaphysics. Blackwell: Oxford. MacIntyre, Alasdair (2007). After Virtue. University of Notre Dame Press: Notre Dame, IN. Mackie, Penelope (2002). “Forbes on Origins and Identities,” in Andrea Bottani, Massimiliano Carrara, and Pierdaniele Giaretta (eds), Individuals, Essence and Identity. Kluwer Academic Publishers: Dordrecht, pp. 341–52. Mackie, Penelope (2009). How Things Might Have Been: Individuals, Kinds, and Essential Properties. Oxford University Press: New York. Mann, Wolfgang-Rainer (2011). “Elements, Causes and Principles: A Context for Metaphysics Z.17,” Oxford Studies in Ancient Philosophy, Vol. 40, pp. 29–61. Marenbon, John (2012). Oxford Handbook of Medieval Philosophy, Oxford University Press: Oxford. Margolis, Eric and Laurence, Stephen (eds) (2007). Creations of the Mind: Theories of Artifacts and Their Representation. Oxford University Press: New York. Markosian, Ned (1998). “Simples,” Australasian Journal of Philosophy, Vol. 76, No. 2 (June 1998), pp. 213–28. Markosian, Ned (2004a). “SoC It To Me? Reply to McDaniel on MaxCon Simples,” Australasian Journal of Philosophy, Vol. 82, No. 2 (June 2004), pp. 332–40. Markosian, Ned (2004b). “Simples, Stuff, and Simple People,” The Monist, Vol. 87, No. 3, pp. 405–28. Markosian, Ned (2015). “The Right Stuff,” Australasian Journal of Philosophy, Vol. 93, No. 4, pp. 665–87. Marmodoro, Anna (2013). “Aristotle’s Hylomorphism Without Reconditioning,” Philosophical Inquiry, Vol. 36, No. 1–2, pp. 5–22. McDaniel, Kris (2003). “Against MaxCon Simples,” Australasian Journal of Philosophy, Vol. 81, No. 2 (June 2003), pp. 265–75. McFarland, Andrew (2011). “Remarks on Koslicki’s Theory of Hylomorphism,” Protosociology, , last accessed March 22, 2018. McKay, T. J. (1986). “Against Constitutional Sufficiency Principles,” Midwest Studies in Philosophy, Vol. 11 (Studies in Essentialism), P. A. French, T. E. Uehling, and H. K. Wettstein (eds). University of Minnesota Press: Minneapolis, MN, pp. 295–304. McLaughlin, Brian (1995). “Varieties of Supervenience,” in Elias E. Savellos and Ümit D. Yalçin (eds), Supervenience: New Essays. Cambridge University Press: Cambridge, pp. 16–59. McLaughlin, Brian and Bennett, Karen (2005). “Supervenience,” in The Stanford Encyclopedia of Philosophy (spring 2018 edn), Edward N. Zalta (ed.), , last accessed March 22, 2018. Merricks, Trenton (2009). “Review of Kathrin Koslicki, The Structure of Objects,” Journal of Philosophy, Vol. 106, No. 5 (May 2009), pp. 301–8. Millikan, Ruth (1984). Language, Thought and Other Biological Categories. MIT Press: Cambridge, MA. Moltmann, Friederike (1990). “Semantic Selection and the Determination of Part Structures in The Use of Natural Language,” in Martin Stokhof and Leen Torenvliet (eds), Proceedings of the Seventh Amsterdam Colloquium (December 19–22, 1989), Institute for Language, Logic and Information (ITLI). Universiteit van Amsterdam: Amsterdam, pp. 19–22.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

258  bibliography Moltmann, Friederike (1997a). Parts and Wholes in Semantics, Oxford University Press: Oxford. Moltmann, Friederike (1997b). “Part Structure Modifiers,” in Mandy Simons and Teresa Galloway. Proceedings of SALT V. Cornell University Press: Ithaca, NY, pp. 255–72. Moltmann, Friederike (1998). “Part Structures, Integrity and the Mass-Count Distinction,” Synthese, Vol. 116, pp. 75–111. Moltmann, Friederike (2005). “Part Structures in Situations: The Semantics of Individual and Whole,” Linguistics and Philosophy, Vol. 28, pp. 599–641. Mouracade, John (2008). “Aristotelian Hylomorphism and Non-Reductive Materialism,” Apeiron, Vol. 41, No. 3, pp. 153–78. Mulligan, Kevin (1998). “Relations—Through Thick and Thin.” Erkenntnis, Vol. 48, pp. 325–53. Mulligan, Kevin, Simons, Peter, and Smith, Barry (1984). “Truth-Makers,” Philosophy and Phenomenological Research, Vol. 44, No. 3, pp. 287–321. Munitz, Milton (ed.) (1971). Identity and Individuation. New York University Press: New York. Nicolescu, Basarab (2000). “Hylemorphism, Quantum Physics and Levels of Reality,” in Demetra Sfendoni-Mentzou (ed.), Aristotle and Contemporary Science, Vol. I. Peter Lang: New York, pp. 173–84. Novaes, Catarina Dutilh (2012a). “Form and Matter in Later Latin Medieval Logic: The Cases of Suppositio and Consequentia,” Journal of the History of Philosophy, Vol. 50, No. 3, pp. 339–64. Novaes, Catarina Dutilh (2012b). “Reassessing Logical Hylomorphism and the Demarcation of Logical Constants,” Synthese, Vol. 185, No. 3, pp. 387–410. Novotný, Daniel D. and Novák, Lukáš (eds) (2014). Neo-Aristotelian Perspectives in Metaphysics. Routledge: New York. Nussbaum, Martha C. (2014). “Aristotle on Human Nature and the Foundations of Ethics, with an Addendum,” in Claudia Baracchi (ed.), The Bloomsbury Companion to Aristotle. Bloomsbury: London, pp. 191–226. Nussbaum, Martha C. and Rorty, Amélie Oksenberg (eds) (1992). Essays on Aristotle’s De Anima. Clarendon Press: Oxford. Oderberg, David (ed.) (1999). Form and Matter: Themes in Contemporary Metaphysics. Blackwell: Oxford. Oderberg, David (2007). Real Essentialism. Routledge: New York. Oderberg, David (2011). “Essence and Properties,” Erkenntnis, Vol. 75, No. 1 (July 2011), pp. 85–111. Oderberg, David (2012). “Survivalism, Corruptionism and Mereology,” European Journal for Philosophy of Religion, Vol. 4, No. 4 (Winter 2012), pp. 1–26. Oderberg, David (2014). “Is Form Structure?,” in Daniel D. Novotný and Lukáš Novák (eds), Neo-Aristotelian Perspectives in Metaphysics. Routledge: New York, pp. 164–80. O’Leary Hawthorne, John and Cover, J. A. (1998). “A World of Universals,” Philosophical Studies, Vol. 91, No. 3 (September 1998), pp. 205–19. Olson, Eric T. (2001). “Material Coincidence and the Indiscernibility Problem,” Philosophical Quarterly, Vol. 51, No. 204, pp. 337–55. O’Rourke, Fran (2004). “Aristotle and the Metaphysics of Evolution,” Review of Metaphysics, Vol. 58, No. 2, pp. 3–59. Pasnau, Robert (2001). Aquinas on Human Nature: A Philosophical Study of Summa Theologiae Ia 75–89. Cambridge University Press: Cambridge.

OUP CORRECTED PROOF – FINAL, 07/23/2018, SPi

bibliography  259 Pasnau, Robert (2011). Metaphysical Themes 1274–1671. Oxford University Press: New York. Pasnau, Robert (2012). “Mind and Hylomorphism,” in John Marenbon (ed.), Oxford Handbook of Medieval Philosophy. Oxford University Press: Oxford, pp. 485–504. Pasnau, Robert (2014). “Form and Matter,” in Robert Pasnau and Christina van Dyke (eds), The Cambridge History of Medieval Philosophy, 2nd edn. Cambridge University Press: Cambridge, pp. 635–46. Pasnau, Robert and van Dyke, Christina (eds) (2014). The Cambridge History of Medieval Philosophy, 2nd edn. Cambridge University Press: Cambridge. Paul, L. A. (2002). “Logical Parts,” Nous, Vol. 36, pp. 578–96. Paul, L. A. (2012). “Building the World from its Fundamental Constituents,” Philosophical Studies, Vol. 158, pp. 221–56. Paul, L. A. (2013). “Categorical Priority and Categorical Collapse,” Aristotelian Society Supplementary Volume, Vol. 87, pp. 89–113. Paul, L. A. (2016). “A One Category Ontology,” in John A. Keller (ed.), Freedom, Metaphysics, and Method: Themes from van Inwagen. Oxford University Press: Oxford, pp. 32–61. Paul, L. A. (2017). “Mereological Bundle Theory,” in Hans Burkhardt, Johanna Seibt, and Guido Imaguire (eds) (2017), The Handbook of Mereology. Philosophia Verlag: München. Pelletier, Francis Jeffry and King-Farlow, John (1984). New Essays on Aristotle, Canadian Journal of Philosophy, Supplementary Volume 10. Peramatzis, Michail (2008). “Aristotle’s Notion of Priority in Nature and Substance,” Oxford Studies in Ancient Philosophy, Vol. 35, pp. 187–247. Peramatzis, Michail (2011). Priority in Aristotle’s Metaphysics. Oxford University Press: New York. Peramatzis, Michail (2018). “Aristotle’s Hylomorphism: The Causal-Explanatory Model,” Metaphysics: Journal of the Canadian Metaphysics Collaborative, Vol. 1, No. 1, pp. 12–32.: “Hylomorphism,” Margaret Cameron, Kathrin Koslicki, and Michael Raven (eds), online at