Dispositions and Causal Powers 9781281104403, 128110440X

Citation preview

DISPOSITIONS AND CAUSAL POWERS

This page intentionally left blank

Dispositions and Causal Powers

Edited by MAX KISTLER University Paris X-Nanterre, Institut Jean Nicod, Paris, France BRUNO GNASSOUNOU University of Nantes, France

© Max Kistler and Bruno Gnassounou 2007 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, electronic, mechanical, photocopying, recording or otherwise without the prior permission of the publisher. Max Kistler and Bruno Gnassounou have asserted their moral right under the Copyright, Designs and Patents Act, 1988, to be identified as the editors of this work. Published by Ashgate Publishing Limited Gower House Croft Road Aldershot Hampshire GU11 3HR England

Ashgate Publishing Company Suite 420 101 Cherry Street Burlington, VT 05401-4405 USA

Ashgate website: http://www.ashgate.com British Library Cataloguing in Publication Data Dispositions and causal powers 1.Dispositions (Philosophy) 2.Causation 3.Science – Philosophy I.Kistler, Max II.Gnassounou, Bruno 111.1

Library of Congress Cataloging-in-Publication Data Dispostions and causal powers / edited by Max Kistler and Bruno Gnassounou. p. cm. Includes bibliographical references and index. ISBN–13: 978–0–7546–5425–4 (hardcover : alk. paper) 1. Disposition (Philosophy) 2. Causation. I. Kistler, Max. II. Gnassounou, Bruno. BD374.D57 2006 111–dc22 2006018032 ISBN–13: 978–0–7546–5425–4

Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire.

Contents Notes on Contributors Acknowledgements Introduction Bruno Gnassounou and Max Kistler Part 1

vii xi 1

The Metaphysics of Dispositions and Causal Powers

1 Dispositions and Counterfactuals. From Carnap to Goodman’s Children and Grandchildren François Schmitz

43

2 Filled In Space Stephen Mumford

67

3 Dispositions and Essences Claudine Tiercelin

81

4 The Causal Efficacy of Macroscopic Dispositional Properties Max Kistler

103

5 Opium’s Virtus Dormitiva Cyrille Michon

133

6 Conditional Possibility Bruno Gnassounou

151

7 On Ascribing Dispositions Ludger Jansen

161

Part 2

Dispositions and Causal Powers in Science

8 An Extended Semantic Field of Dispositions and the Grounding Role of Causal Powers Rom Harré 9 What Makes a Capacity a Disposition? Nancy Cartwright

181

195

vi

Dispositions and Causal Powers

10 Causation, Laws and Dispositions Andreas Hüttemann

207

11 Can Capacities Rescue us from Ceteris Paribus Laws? Markus Schrenk

221

12 Dispositions, Relational Properties and the Quantum World Mauro Dorato

249

13 Are Specific Heats Dispositions? Anouk Barberousse

271

Bibliography Index

283 299

Notes on Contributors Anouk Barberousse is a research fellow at the Institut d’Histoire et de Philosophie des Sciences et des Techniques (CNRS, Paris). Her main topics of research are the epistemology of complex systems and computer simulation, and the cognitive foundations of scientific pictures. Among her publications are: «The Diversity of Models in Statistical Mechanics. Views about the Structure of Scientific Theories», Foundations of the Formal Sciences II, Applications of Mathematical Logic in Philosophy and Linguistics, B. Löwe, W. Malzkorn and T. Räsch (eds), Kluwer, 2003, pp. 1-24: La physique face à la probabilité, 2000, Vrin, coll. Mathesis. Nancy Cartwright is Professor of Philosophy at the Department of Philosophy, Logic and Scientific Method at the London School of Economics and at the University of California, San Diego. She also is the Chair of the Centre for Philosophy of Natural and Social Science. Her principal interests are philosophy and history of science (especially physics and economics), causal inference and objectivity in science. Her book publications include How the Laws of Physics Lie (1983), Nature’s Capacities and their Measurement (1989), Otto Neurath: Philosophy between Science and Politics [co-author] (1995), The Dappled World: A Study of the Boundaries of Science (1999) and Hunting Causes and Using Them: Approaches in Philosophy and Economics (forthcoming). Mauro Dorato is Professor of Philosophy of Science in the Department of Philosophy of the University of Rome 3. His areas of interest include the philosophy of time, laws of nature and the relationship between physics and metaphysics. Among his books, Time and Reality, Clueb, Bologna, 1995; The Software of the Universe, Ashgate, 2005. Among recent papers: ‘Absolute Becoming, Relational Becoming and the Arrow of Time’, in Studies in History and Philosophy of Modern Physics 37, 3 (2006): pp. 559-576. Address: via Ostiense 234, 00146, Rome, Italy. [email protected] Bruno Gnassounou is Lecturer (Maître de Conférences) in Philosophy at the University of Nantes. His main topic of research concerns modality and he has written several papers on questions of action, causality, meaning and mind. Rom Harré began his academic career teaching mathematics and physics. Turning to philosophy of science he was much involved in setting up the school of physics and philosophy in Oxford. In recent years he has worked at Georgetown and American Universities in Washington DC. His most recent books include One Thousand Years

viii

Dispositions and Causal Powers

of Philosophy (2000), Cognitive Science (2002) and Modeling: Gateway to Nature (2004). Andreas Hüttemann is Professor of Philosophy at the University of Münster. His areas of specialization are philosophy of science and early modern philosophy. He is particularly interested in the relation of science and metaphysics. Among his recent publications: What’s wrong with Microphysicalism, London (Routledge) 2004; ‘Explanation, Emergence and Entanglement’, Philosophy of Science 72 (2005), pp. 114–127; ‘Physicalism Decomposed’ (with David Papineau), Analysis 65 (2005) pp. 33–39. Ludger Jansen teaches philosophy at the University of Rostock. His current topics of research are ontology (in particular social and biomedical ontology) and Ancient and Medieval Philosophy (in particular Aristotle and Aquinas). He is the author of Tun und Können (Frankfurt 2002), an analysis and defence of Aristotle’s theory of dispositions and of numerous articles. He is co-editor of Anthropologie in der antiken Philosophie, (forthcoming) and of the book series Topics in Ancient Philosophy. Institut für Philosophie, Universität Rostock, 18051 Rostock, Germany. [email protected] Max Kistler is Lecturer (Maître de Conférences) in Philosophy at Université Paris X, Nanterre, and a member of Institut Jean Nicod, Paris. His main topics of research are causation, laws of nature, reduction and emergence in general philosophy of science and philosophy of mind. He is the author of Causalité et lois de la nature, Vrin, 1999, English version Causation and Laws of Nature, Routledge, 2006, and coauthor (with Anouk Barberousse and Pascal Ludwig) of La philosophie des sciences au XXe siècle, Flammarion, 2000. [email protected] Cyrille Michon is Professor of Philosophy at the University of Nantes. He was formerly Professor of Medieval Philosophy and has published books, papers and translations on Ockham and Aquinas mainly. He is now more involved in topics of analytic philosophy, especially philosophy of action, metaphysics and philosophy of religion. His last publications include Prescience et liberté. Essai de théologie philosophique (PUF, 2004) and Thomas d’Aquin et la controverse sur l’éternité du monde (GF-Flammarion, 2004). [email protected] Stephen Mumford is Professor of Metaphysics at the University of Nottingham. He is the author of Dispositions (OUP 1993) and Laws in Nature (Routledge 2004) and he is editor of Russell on Metaphysics (Routledge 2003) and Powers, by George Molnar (OUP 2003). His current research interest is the philosophy of causation, in particular how to get a theory of causation from the causal powers ontology. He is

Notes on Contributors

ix

also at present writing a book on the philosophy of David Armstrong for Acumen. [email protected] François Schmitz is Professor of Logic and Philosophy at the University of Nantes (France). A large part of his work has been devoted to the study of Wittgenstein’s philosophy, in its relation to Russell and Frege (Wittgenstein, la philosophie et les mathématiques, Paris, PUF, 1988, inter alia). He has also published some critical papers on modal logic and is now working on the Vienna Circle, especially in its relation to physics. Markus Schrenk is a postdoctoral research fellow within the AHRC Metaphysics of Science Project in Nottingham. He also holds a Junior Research Fellowship at Worcester College, University of Oxford. Markus is mainly interested in metaphysics, especially in causation, laws of nature, and dispositions. He is the author of The Metaphysics of Ceteris Paribus Laws, Frankfurt (Ontos), forthcoming 2007; Einführung in die Sprachphilosophie (An Introduction to Philosophy of Language) with Albert Newen, Darmstadt (Wissenschaftliche Buchgesellschaft (WBG)), forthcoming 2007; and ‘Galileo vs. Aristotle on Free Falling Bodies’ in Logical Analysis and History of Philosophy, Volume 7, 2004. [email protected]

Claudine Tiercelin is Professor of Philosophy at the University of Paris XII and a researcher at the Institut Jean Nicod. She has mainly worked and written on C.S. Peirce and pragmatism and is currently working on issues related to the logic, epistemology and metaphysics of vagueness. Her main published books are: La pensée-signe, éditions J. Chambon, Nîmes, 1993; Peirce et le pragmatisme, Presses universitaires de France, 1993; Hilary Putnam, l’héritage pragmastiste, Presses universitaires de France, 2002; Le doute en question, éditions de l’éclat, Paris 2005. And among her articles related to metaphysics: «La métaphysique et l’analyse conceptuelle», Revue de Métaphysique et de Morale, n. 4, oct-dec. 2002, 559-584; «Sur la réalité des propriétés dispositionnelles», Cahiers de Philosophie de l’université de Caen, «Le réalisme des universaux», no. 38–39, 2002, 127-157; «Le problème des universaux: aspects historiques et perspectives contemporaines», in J.-M. Monnoyer (ed.), La structure du monde: objets, propriétés, états de choses: le renouveau de la métaphysique australienne, Paris,Vrin, 2004. [email protected]

This page intentionally left blank

Acknowledgements Most of the papers collected in this volume have been presented at a conference on ‘Dispositions and Causal Powers’ in Paris, 19-21 September, 2002. We would like to express our gratitude to the University of Paris X-Nanterre and the Ecole Normale Supérieure of Paris, rue d’Ulm, who hosted the conference, and the following institutions for their generous support: Université Paris X-Nanterre, Centre d’Histoire et de Philosophie des Sciences, Département de philosophie, and Service des relations internationales, Université de Nantes, Département de philosophie, Institut Jean Nicod (CNRS), Institut Universitaire de France, Université Paris IV, Ecole des Hautes Etudes en Sciences Sociales. We owe a large debt to Sophie Bilardello at Institut Jean Nicod who has edited the whole manuscript, and patiently spotted and corrected many errors.

This page intentionally left blank

Introduction Bruno Gnassounou and Max Kistler

Do we need the notions of causal power and disposition to account for change? If there are powers, are they causes? Or do we rather have reason to deny such powers any ontological reality? Although questions of this sort occupy centre stage in a considerable part of contemporary work on powers and dispositions,1 they are not new. However, received opinion has it that modern philosophy and the new science of mechanics of the seventeenth-century were constituted in opposition to those all too famous ‘virtues’ – a term used during that epoch to designate powers – upon which the scholastic doctors – the medieval philosophers and their seventeenth-century heirs – were reputed to ground all causal explanations. Suddenly, or so tradition has it, it became evident how ridiculous it was to explain, as Molière’s2 bachelor candidate, that opium makes one sleepy because opium possesses a ‘dormitive virtue’ (virtus dormitiva), in other words, a power to make people sleep. The final blow against the notion of power would have been struck by Hume whose analysis of causation eliminates all recourse to the notion of power in the explanation of change. Since then, causation has been considered as a certain type of relation between two actual events that must satisfy a certain number of conditions (temporal succession, contiguity, repetition).3 The aim of this introduction is to improve on the traditional way of summing up the history of the notions of power and disposition, by uncovering some of the complexities that remain hidden behind such an oversimplification. We shall proceed in two steps. First, we shall have a look at the way in which the notion of power, handed over from the Aristotelian and scholastic tradition, was interpreted by some of the most representative authors of the modern epoch, in the seventeenth-century. Second, we shall examine why, once the notion of a law of nature had become available, many philosophers, from that epoch to the present, judged the notion of disposition superfluous, and why that notion has nevertheless returned to occupy centre stage in the work of many contemporary thinkers.

1 In this introduction, we shall use the expressions ‘power’, ‘disposition’ and ‘capacity’ interchangeably. 2 The famous scene in Molière’s Le malade imaginaire, Third Interlude, in G. Couton (ed.), Molière, Oeuvres complètes (Paris, 1971), p. 1173; trans. A.R. Waller (Edinburgh, 1926), vol. 8, p. 308. 3 We shall come back to Hume later in this introduction.

2

Dispositions and Causal Powers

Powers and Dispositions in Early Modern Philosophy It goes without saying that we do not pretend to provide, in these pages, the complete history of philosophical thought about powers and dispositions. However, it is quite safe to say that the seventeenth-century constitutes a turning point in that history. Descartes, in particular, plays a crucial role. Indeed, arguing against a scholastic tradition that grounded the explanation of change on the action of ‘substantial forms’ or ‘real accidents’ (the precise nature of which we shall return to later), Descartes replaces this kind of explanation with a conception according to which matter has an essentially spatial nature. The wealth of qualities and changes of matter, as well as the different phenomena of generation and corruption, must and can be explained by the mere configuration and movement of the parts of material objects. Thus, weight and heat, which give bodies the power of falling and fire the power of burning them, must not be explained by the presence in the rock of a real accident that ‘pushes’ the body to fall, or by a substantial form of fire that explains the body’s containing heat as well as its capacity to burn other bodies. Weight and heat should rather be conceived of as purely mechanical effects of different parts of extended matter that the new science of matter would explain. Before returning to this opposition, crucial for our purposes, between the scholastic doctrine and Cartesian philosophy, we should pause to note that Molière’s candidate’s answer is not as empty as it appears. We can see why, by formulating more charitably the question he might have faced: that is, not why opium makes people sleep, but rather why people fall asleep after having absorbed opium. In replying to the latter question, it is not trivial to answer that opium has the power to make people sleep, in other words that it is a soporific. Criticism of the answer, thus understood, raises substantial issues rather than boiling down to a mere accusation of ‘tautology’. On this basis, we can distinguish three types of criticisms that this reasoning elicits in the seventeenth century.4 First, the explanation relying on dormitive virtue can be criticized on empirical grounds: it is simply wrong that opium has such a capacity, just as inquiry has shown it to be wrong that garlic has any anti-magnetic power. The very idea of a power inhering in things is not absurd; it is just that experience shows that certain attributions of faculties to certain substances turn out to be unfounded. Second, without eliminating the dormitive power of opium, it may be argued that it reduces to another type of power. This reductionist reply aims at reducing the number of powers. Maybe opium has only a virtus fatigativa, a power to bring about weariness that it shares with bad novels, and it is a person’s weariness that eventually induces sleep. The complaint that explanation by powers tends to arbitrarily multiply the number of virtues and thereby makes them lose their explanatory strength because a new specific power is hastily introduced for each new effect to be explained, was often voiced in the seventeenth century. Nevertheless, it does not by itself discredit 4 We draw here on the excellent analysis in K. Hutchison, ‘Dormitive Virtues, Scholastic Qualities, and the New Philosophies’, History of Science, 29, no. 85 (1991): 245-250.

Introduction

3

the notion of power as such. Newton occasionally accuses those who postulate specific powers behind each effect, not to explain anything at all.5 However, all he wants to say in this context is that it is not sufficient, in order to explain the elliptic form of planetary orbits, to attribute a virtus elliptofactanda6 to the sun, bypassing in this way authentic scientific work. Newton’s criticism does not exclude the hypothesis that the sun has a gravitational power whose mode of operation it took Newton’s genius to reveal. A good scientific explanation must end up reducing the number of powers that must be postulated, but not necessarily eliminate the concept of power itself. Finally, it is possible to deny the very existence of powers in natural beings, for a very particular reason: between ingesting opium and falling asleep one observes the existence of a relation of succession; nothing forces us to postulate anything like a real causal relation and hence an active power in opium to ground that relation. In fact, it is not logically absurd to suppose that an external agent arranges things in such a way that, on the occasion of a person ingesting opium, the person falls asleep. The only power at work in this bringing about the person’s falling asleep belongs to the external agent. Once again, the coherence of the concept of cause itself is not at issue. It is only its extension that is reduced to a minimum, because the capacity to act is attributed to one single being alone. Malebranche’s writings illustrate well this ‘occasionalist’ position, according to which all apparent operation in nature must be attributed to a supernatural agent, God. In other words, all causal relations are grounded in a ‘general order of nature’ governed by divine operation, rather than by the particular faculties inherent in natural substances, the latter being deprived of any capacity of affecting other substances.7 As we shall see, it suffices to abstract from the divine hypothesis to open the path for Hume’s new analysis of causation. Change and Power in Aristotle In the seventeenth century, it is common to attribute the abhorred doctrine to the ‘School’ or to the ‘peripateticians’, in other words to the medieval heirs of Aristotle and hence to Aristotle himself. However, Aristotle’s own writings do not contain any 5 For Newton, ‘to tell us that every Species of Things is endowed with an occult specifick Quality by which it acts and produces […] effects, is to tell us nothing’. Newton, E.T. Whittaker (ed.), Opticks, or Treatise of the Reflections, Refractions, Inflections and Colours of Light (New York, 1931), p. 401; quoted in K. Hutchison ‘Dormitive Virtues, Scholastic Qualities, and the New Philosophies’, p. 250. 6 See K. Hutchison ‘Dormitive Virtues, Scholastic Qualities, and the New Philosophies’, p. 251. 7 See, e.g. Malebranche: ‘I hold, as I have said elsewhere, that bodies, for example, do not have the force to move themselves and that therefore their motor force is but the action of God’. N. Malebranche, ‘Eclaircissements sur la recherche de la vérité’, in G. Rodis-Lewis (ed.), Oeuvres (Paris, 1979), vol. I, p. 1003; trans. T.M. Lennon and P.J. Olscamp, ‘Elucidations of The Search After Truth’, in The Search After Truth, Elucidations of The Search After Truth (Columbus, 1980), p. 678.

4

Dispositions and Causal Powers

trace of the notions of real accident and substantial form as entities distinct from the substance of which they are supposed to be the accident or the form. Let us briefly recall Aristotle’s theory of change. Where there is change, there are causes of change. Aristotle distinguishes four types of causes (or types of explanations, the Greek term aitia allows for both readings):8 that of which a thing is made, or material cause (the clay is the cause of the statue in this sense, as are the letters of a syllable), the form of the thing or the definition providing the essence of the thing (this piece of bronze is a statue, this flowing water is a river), that which produces the thing or the ‘agent that produces the effect, changing it from what it was to what it is to be’9 (the father is the cause of the child in this sense, as is the architect of the house) and the end or purpose (health is the reason, or cause, of the walk). There are three important things to note about this classification. First, Aristotle’s causes are by definition that what is given by the explanation of what happens, of what persists and in general of what must be accounted for. Thus, the definition or form of a thing is a cause because it ‘accounts for’ what the thing is by giving its essence. Hence, causes are not in the first place, nor essentially, events (or descriptions of events), as they are in modern theories of causation. In Hume’s theory in particular, goals, agents, form and matter could not be taken to be causes because they are not events or ‘happenings’. Second, the formal cause or simply ‘form’ of a thing constitutes what Aristotle calls its nature. Aristotle’s physics is structured around the concept of beings that ‘have within themselves a principle of movement (or change) and rest’,10 contrasted with artificial objects, such as a bed or a tripod. In particular, each of the four elementary substances, water, fire, air and earth, behaves differently in circumstances of a given type (for example of temperature). These ‘simple bodies’,11 as Aristotle calls them, behave in a specific manner according to their nature, in the sense that their nature determines the types of changes these simple bodies may undergo. Thus, the nature or form of a thing accounts for the type of change it undergoes, and can therefore be counted among its causes. This brings us to our third point. For Aristotle, no change can come about without an agent (the moving or, in the literal use of the term, efficient cause) and his conception of an agent is essentially that of a substance operating on another substance, the latter suffering the process of change. This is why it is named ‘patient’.12 Even the behaviour of living beings must be conceived of according to this model, even if in their case, unlike the case of nonliving natural substances, agent and patient are

8 Aristotle, trans. P.H. Wicksteed and F.M. Cornford, Physics (Cambridge, 1929 and 1934), book II, ch. 3, 194b24-195a3 and ch. 7, 198a14-21. 9 Aristotle, Physics, book II, ch. 3, 194b31. 10 Aristotle, Physics, book II, ch. 1, 192b14. 11 Aristotle, Physics, book II, ch. 1, 192b10 and Aristotle, trans. E.S. Forster, On Comingto-be and Passing-away (Cambridge, 1955), book II, ch. 3, 330b30. 12 Aristotle says that the change is in the patient; see Physics, book III, ch. 3, 202a15.

Introduction

5

identical. The acting agent who affects the patient therefore has the capacity or the power to act; and the patient has the capacity or power to undergo the agent’s action. From this it follows, at least in principle, that one would commit a category mistake (in the modern sense of the expression) if one confused the formal and the efficient cause. In the strict sense, the form does not act.13 Opium certainly possesses a form that accounts for the type of action that it can perform on certain patients. However, this does not at all imply that the form is an internal entity that pushes the opium to have its cataleptic effects. It is the opium, not its form that operates to bring about the patient’s falling asleep, where the form merely determines the types of action that its bearer is capable of performing and thereby the types of capacities it possesses. This conception of change leads Aristotle to distinguish between potentiality or power (dunamis) and effectivity (energia), which is the realization of the power.14 Any entity that is potentially something realizes this potentiality by being the subject of a process, taking it from a state in which it is not that something to a state in which it is effectively that something. In this sense, the acorn is potentially the oak tree, and the oak tree is the realization of that power. That process of change (métabolé) can take four forms: substantial (coming into being, ceasing to be, for substances in general; birth and death, for organisms in particular), quantitative (growth, shrinking), qualitative (gain or loss of this or that quality, for example becoming white) and local (change of place, locomotion).15 A change in this sense is a process (by opposition to a simple activity, such as seeing or walking) characterized by the fact that 1) it manifests a progression from an initial state to a final state, from a state where that which undergoes the transformation, the patient, has something potentially to a state where the same patient has this something effectively, 2) it takes time, which makes it possible to say that it is quick or slow and 3) it is constituted by differentiated steps (when one builds a house, one starts with the foundations, then builds the walls and ends with the roof).16 Aristotle sums up this doctrine in his famous definition of change as ‘the progress of the realizing of a potentiality, qua potentiality’.17 This definition has been accused of being circular, because it defines change by characterizing it as a process going from the power to the realized power. However, effectivity does not mean ‘movement of realization’ or actualization, but rather actuality or reality. Aristotle simply intends to categorize movement among the effective realities, and not as something ‘under way’ between a power and its realization. Indeed, the pieces of wood before me have two kinds of power: to be a table and to get transformed into a table (to be changed into a table). The table that exists is the effectivity of the first power, whereas the 13 On this difficult problem, see A. Mansion, Introduction à la physique aristotélicienne (Louvain-la-Neuve, 1987), ch. 7. 14 Aristotle offers an analysis of the notion of power in Aristotle, trans. H. Tredennick, The Metaphysics (Cambridge, 1933), book IX (θ). 15 See Aristotle, Physics, book II, ch.1, 192b15, On Coming-to-be and Passing-away, book I, 319b32-320a2. 16 See Aristotle, The Metaphysics, book IX. 17 Aristotle, Physics, III, ch. 1, 201a10-11.

6

Dispositions and Causal Powers

change affecting the pieces (the action of constructing the table) is the effectivity of the second power. My action of constructing the table is the effectivity of the power inherent in the pieces of wood, the effectivity of the constructible as constructible (Aristotle says, of the constructible ‘as such’), a power the pieces possess in virtue of being capable of undergoing that change, and not, e.g. by virtue of being owned by the agent. The distinction between power and effectivity can be drawn both with respect to a substance and with respect to an accident (for example a quality). The wine in this bottle can acquire new aromas by ageing: if this power gets realized, the substance, i.e. the wine, persists but acquires new qualities. The change is accidental. However, the same wine can also get transformed into vinegar. That change is substantial because the wine and the vinegar are two different substances: they are not the same sort of thing. Aristotle calls matter what is capable of substantial change, i.e. what remains identical while the substance changes. The matter appears first in one form then in another. It is for this reason that Aristotle’s followers, without at this point betraying the spirit of the master, will say that the matter had the substantial form of wine and then acquired the substantial form of vinegar. The substantial form is what makes it the case that this wine is a substance of this kind. This ‘making the case’ is very different from the active ‘bringing about’ that produces the substance. Not to pay attention to this difference would amount to confusing the formal cause with the efficient cause. The substantial form provides what contemporary philosophers18 often call a thing’s criterion of identity (which is what allows justifying that this wine is different from that beer and, in a different manner, that this wine yesterday is the same thing as that wine today). Matter provides a principle of individuation: it is what makes this beer distinct from that other beer although both drinks otherwise satisfy the same criterion of identity. One question that arises is whether there must always be some substantial predicate that is true of matter that remains identical throughout a substantial change. In many cases it is clear that there is such a predicate: when for example a piece of copper is transformed into bronze, the matter undergoing a substantial change remains a metal throughout. However, Aristotle19 and later Aquinas held that nothing prevents matter from undergoing a substantial change such that no substantial predicate is true of it before, during and after the change. Such matter, called ‘prime matter’, cannot exist without form, even if it is not tied to a particular form. Similarly, no substantial form can exist without being constituted of matter. Aristotle’s doctrine does not attribute any ontological autonomy either to substantial forms or to prime

18 See D. Wiggins, Sameness and Substance (Oxford, 1980) and E.J. Lowe, A Survey of Metaphysics (Oxford, 2002). 19 Whether Aristotle himself accepted the existence of prime matter is controversial. See On Coming-to-be and Passing-away, book I, ch. 4, 320a2 and book II, ch. 1, 329a32.

Introduction

7

matter; only concrete substances exist, and these substances cannot be conceived as compositions of substantial form and matter.20 In this general framework, Aristotle raises the problem as to whether it is legitimate to conceive of a power that never gets realized.21 If, for one reason or another, it is established in advance that the coat gets worn out before it is torn up, can it be said that it has ever had the power of being torn up? Aristotle made considerable efforts to justify the attribution of powers to things even if they are not exercised, against those who, like the Megarians, held that a power that is not exercised is a contradictory notion, and who concluded from this that a possibility should be identified with its realization. Possibilities exist only if they are realized and, in the end, something is possible now only if it is effective now. On this view, if Socrates is now sitting, it must be denied that he now has the power to stand later. Although Aristotle criticized this thesis, he was not able to provide a clear answer as to whether the possibility today implies the realization one day or another.22 Clearly, the whole debate revolves around the issue of the ontological status of powers: if a power precedes its exercise, if it constitutes an actual property of a thing, it is natural to wonder about the mode of being of such a mere potentiality. How does one conceive the mode of existence of a mere possibility? The Reification of Powers, Real Accidents and Substantial Forms at the Beginning of the Seventeenth Century One solution to the problem is to lend quasi-substantial reality to powers. A power becomes something that acts. From the point of view of modern philosophy, this reification of power is characteristic of the late scholastic school. What exactly is this reification of which the moderns accuse the schoolmen? Certain texts of the Spanish Jesuit Suarez, author of a gigantic synthesis, both systematic and original, of scholastic knowledge, provide some essential keys for understanding the debate. Written at the beginning of the seventeenth century, Suarez’ ontology is based on the distinction between substance and accident and his analysis of the notion of power and effectivity faithfully follows the Aristotelian tradition.23 However, Suarez’ analysis, accomplishing an evolution whose source can be found in medieval scholasticism, gives an important new twist to the understanding of the very notions of substantial form and accident, and particularly to that of quality. 20 This is not completely true for Aquinas who, mostly for theological reasons, holds that the substantial forms of humans, which he calls souls, can persist even while they are separated from the body. 21 See Aristotle, On Interpretation, in Aristotle, trans. H.P. Cooke, Categories, On Interpretation, Prior Analytics (Cambridge, 1938), ch. 9. pp. 111-179. 22 On this thesis, often called ‘the principle of plenitude’, see J. Hintikka, Time and Necessity (Oxford, 1973), ch. 5; S. Knuutila, Modalities in Medieval Philosophy (London, 1993), ch. 1 and J. Vuillemin, Nécessité ou contingence (Paris, 1984) chs. 1-3. 23 F. Suarez (1597), ‘Disputationes Metaphysicae’, in C. Berton (ed.), Opera omnia. Editio nova (Hildesheim, 1998), Disputatio 43, XXVI, pp. 633-663.

8

Dispositions and Causal Powers

Let us first consider the latter. A blade of a knife having a quality such as sharpness accounts for its operations of cutting. To a very general inquiry about the ontological status of powers, it is natural to reply that very often they are qualities. However, once this first reply has been given, a second question immediately arises: are qualities ontologically independent of the substances of which they are qualities? William of Ockham who had already raised this question in the fourteenth century, replied that certain qualities are ontologically independent but others are not. He has a simple criterion to ground this distinction: qualities are distinct things, res distincta, in case it is impossible to account for their acquisition or loss by a mere local movement. Thus, the geometrical forms of objects, e.g. their being straight or bent, are not distinct res because it is possible to explain their acquisition by the mere spatial shifting of the object’s parts.24 The same is true of the beauty of a thing, which depends on the relation between its parts. However, dispositions (in the sense in which one can say that someone is well disposed), habits (habitus) which are acquired powers, and natural capacities (potentiae naturales) sometimes need to be considered as distinct res: this is the case for the habits of will whose acquisition cannot be explained by any rearrangement of its parts (because the will has no parts); however, it is not the case of the disposition of health which can be explained by the simple proportion of humours.25 When Suarez takes up this issue, he maintains Ockham’s distinction, among accidents in general and qualities in particular, between those that can and those that cannot be considered as entities truly distinct from the substance they inhere in. Thus, the quantity of a thing ‘is not only a mode, but something distinct from the substance’.26 Suarez calls such properties that can be separated from their substance, real accidents. By contrast, those properties of substances that cannot be separated from them are mere modes of those substances. A mode exists only by inhering in a substance. Take away the inherence and you have taken away the mode. The important point is that the separability of two ‘entities’ is taken to be sufficient for judging that there is a real distinction between them, not only a modal27 distinction,

24 See William of Ockham, Summa logicae, in P. Boehner, G. Gal and S. Brown (eds), Opera philosophica (New York, 1980), I, 180 (Pars I, Cap. 55); and William of Ockham, trans. A.J. Freddoso, Quodlibetal Questions, 2 vols (New Haven: 1980, vol. II, 597 (Seventh Quodlibet, Question 2, vol. IX, pp. 707-708, opera theologica, New York, 1980) 25 See Ockham, IX, p. 94 (First Quodlibet, Question 18); trans. Ockham (1991), I, 80. 26 ‘Quantity is therefore not so much a mode, as a thing distinct from the substance (ergo quantitas non est tantum modus, sed res distincta a substantia)’, F. Suarez, Disputation 40, XXVI, §8, p. 535. The context is the Miracle of the Multiplication of the Loaves and the Fishes. To account for it, one must suppose, says Suarez, that God can separate the quantity of the substances bread and wine from those substances themselves. 27 One of Suarez’ great achievements is to have provided a systematic and original presentation of what may be called the theory of distinctions. See Suarez, Disputation 7, XXV, 250-274.

Introduction

9 28

and at the same time for saying that a mode is not a res because it is not separable. Suarez takes it to be certain that the powers of a thing belong to the category of real accidents, not to that of modes, which means that powers are really distinct from substances.29 According to Descartes, substantial forms ‘were introduced by philosophers solely to account for the proper actions of natural things, of which they were supposed to be the principles and bases (principium et radix)’.30 On this point, Aristotle and the scholastics would agree. It is indeed in this form, that, in Suarez’ words, ‘all that variety of accidents and powers has its root (radix) and a certain unity’.31 Therefore, a substance must be thought to possess a substantial form that is ‘distinct from its accidents’32 (and its powers) and allows to account for them. Here, however, scholastic doctrine adds an important new twist to Aristotle’s analysis. Indeed, Suarez takes matter and substantial form to be both entities that can exist in separation one from the other,33 although they call for each other to compose a complete and concrete substance. Furthermore, the substantial form surreptitiously changes its status as a criterion of identity of the composed substance and the actions it is capable of accomplishing, in other words its status as merely formal cause, to become the efficient cause of that composed substance: it becomes, in Suarez’ own words, its ‘principle of operation’.34 The substantial form that explains the operations of the complete substance is now considered as an entity distinct from it. To say that opium has the power to make people sleep, means that it contains a separate entity making it what it is: it is the opium’s substantial form that acts on the patient after it has been ingested. To use one of Suarez’ own examples, the fact that heated water cools down when the heat source has been removed, can be explained by the existence of a substantial form that is the active source of the water’s tendency to cool down instead of remaining at the temperature it has acquired.35 The ultimate

28 ‘Hence there exist, in created things, certain modes affecting them whose nature seems to consist in this, that they are not sufficient by themselves (per se) to constitute a being (ens) or an entity (entitatem) in the nature of things, but intrinsically require an entity that they affect and without which they can in no way (nullo modo) exist.’ (Suarez, Disputation 7, 1, §18, p. 256.) 29 Suarez says that the ‘proximate power of acting and operating (proximam virtutem agendi et operandi) is distinct from its substance and is consequently an accident’, (Suarez, Disputation, 18, 3, § 17, trans. p. 103). 30 Descartes, Letter to Regius, January 1642, in Descartes, Oeuvres complètes, ed. C. Adam and J. Tannery, vol. III, p. 506; trans. Descartes (1984), III, p. 208. 31 Suarez, Disputation 15, 1, § 7, p. 499. 32 Ibid. 33 Suarez, Disputation 15, 5, §§ 1-2, pp. 517-518. 34 ‘In material things, it appears that form is a principle of operation (principium operandi), which is a formal and active principle’. (Suarez , Disputation 15, 11, § 4, p. 559, our italics.) 35 Suarez, Disputation 15, I, 8, p. 500.

10

Dispositions and Causal Powers

source of a substance’s powers lies in its substantial form whose status is first that of a substance though an incomplete one, and second and above all that of an agent.36 Descartes and Boyle Robert Boyle, Descartes’ and Bacon’s heir and Locke’s direct source of inspiration, was the most eloquent and in many respects most fascinating advocate of the new anti-scholastic philosophy. The attack launched both by Descartes and Robert Boyle is mostly directed at the ‘quasi-substantialization’ of qualities and at the authentic substantialization of substantial forms. If there is one key element in the wide-reaching conceptual change introduced by Descartes, it is the distinction between the mental and the physical, in other words dualism of thought and extension replacing the distinction between form and matter constitutive of Aristotle’s hylomorphism. This new conception of the relation between mind and body is at the source of Descartes’ relentless criticism of the scholastic theory of the action of physical substances. In retrospect, Descartes appears to have been right in rejecting the idea that physical substances are literally animated by an inner soul, and that it is this soul which gives them the capacity to act. So let us look a little closer at the way in which Descartes and Boyle criticize the notions of real accident and substantial form. Descartes puts forward two fundamental arguments against the notion of real accident.37 The first and doubtless most famous argument consists in accusing Aristotle of ‘metaphysical paganism’:38 by attributing to each substance an internal capacity of change, he populates the universe with so many little souls. Let us take the case of heaviness, which is the only quality Descartes really analyses.39 This quality belongs to the essence of a body and is independent of its extension. It is somehow united to that extension though it does not depend on it. This is shown by the fact that it can produce its effects by being localized at one point of the body: if the body is suspended on a thread, it seems as if its heaviness acts entirely at the point at which the body is suspended, a point that may ideally be reduced to a ‘mathematical point’. For Descartes, this conception amounts to considering heaviness as a little soul: just as a soul, it appears to be both present ‘in each part’ of the body at the same time and 36 To explain how, in the concrete actions of a complete substance, the substantial form and the real qualities are associated in accomplishing certain operations is a complex issue, and not necessary to pursue here. 37 On this point, as well as about substantial forms, see the excellent work by M. Rozemond, Descartes’s Dualism (Cambridge, 1998), ch. 4. See also E. Gilson, Etudes sur le rôle de la pensée médiévale dans la formation du système cartésien (Paris, 1967), p. 141ff. 38 The expression is ours, not Descartes’. However, Malebranche can be found accusing the peripateticians and more generally the ancients, of ‘paganism’. See Malebranche, ‘Eclaircissements sur la recherche de la vérité’, book VI, 2nd part, ch. 3, p. 643ff., trans. p. 446ff. 39 See Replies to the Sixth Set of Objections, in Descartes, Oeuvres, vol. VII, pp. 441442 (Latin) and IX, pp. 240-241 (French); trans. Descartes (1984), II, pp. 297-298.

Introduction

11

capable of producing its entire action in each of these parts, just as a human agent does when she moves one of her members. True, heaviness also appears to us as a quantity that is measurable, divisible and therefore quantifiable, just as extension. But for Descartes this only shows that the scholastic real qualities are the result of a regrettable confusion between mental properties and purely bodily properties. If we are inclined to attribute a real quality to the heavy body, this is because we confusedly attribute to it something equivalent to the faculty of knowledge, which really belongs only to mankind. This faculty is thought to allow the body to ‘know’ where it should aim its movement, namely to the centre of Earth. This conception results from illegitimately projecting on the relations between simple physical bodies, a model of movement that fits perfectly well to human action where the soul acts on the body (thereby inducing movement): the relations between simple physical bodies can be given a fundamental explanation by the laws of collision.40 In sum, by attributing to the body something that is equivalent both to a capacity of representation and to a faculty of volition, we fall prey to a spontaneous and naive form of anthropomorphism. No body, as a body, has any internal power of moving itself or of acting on other bodies. Descartes’ second argument is at least as strong as the first, which suffers from being applicable only to the quite specific case of heaviness. It consists of rejecting the very concept of a real quality. Recall that according to late scholastic doctrine, separability was considered as sufficient condition for the existence of two distinct res. However, being a res did not imply being a substance. This is precisely the logical point of the doctrine of real accidents or qualities that Descartes rejects. From his point of view, the distinction between two res makes them necessarily two substances, for they can clearly and distinctly be conceived as independent. If indeed we ‘imagined […] to be real’ the ‘various qualities of bodies such as heaviness, heat, etc.’, we thereby imagined them ‘to have an existence distinct from that of bodies, and so to be substances’.41 Now, ‘it is completely contradictory that there should be real accidents, since whatever is real can exist separately from any other subject; yet anything that can exist separately in this way is a substance, not an accident.’42 The very notion of a real accident is absurd because it attributes two contradictory properties to one entity: to be an accident and therefore to be ontologically dependent, and to be a substance, and therefore to be ontologically independent. Boyle criticizes the notion of real accident in a similar way. Boyle defends a ‘corpuscular philosophy’: all so-called ‘secondary’ or ‘non-mechanical’ qualities (manifest qualities, such as cold and hot, occult qualities such as magnetism and

40 ‘So I think that we have hitherto confused the notion of the soul’s power to act on the body with the power one body has to act on another.’ (Letter to Elizabeth, 21 May 1643, Descartes, Oeuvres, vol. III, p. 667, trans. Descartes (1984), III, 219). 41 Letter to Elizabeth, 21 May 1643, Descartes, Oeuvres, vol. III, p. 667; trans. Descartes (1984), III, 219; italics ours. 42 Replies to the Sixth Set of Objections, in Descartes, Oeuvres, vol. VII, 434-435 (Latin) and vol. IX, 234-235 (French); trans. Descartes (1984), II, 293.

12

Dispositions and Causal Powers

electricity, or sensible qualities such as colours) attributed to things must ultimately be explained by the fundamental ‘primary’ or ‘mechanical’ qualities of the basic material units of composition, such as form, size, movement and texture. Boyle refrains from saying whether corpuscles are in principle divisible, which makes his position weaker than atomism, which holds that those basic units are indivisible. Texture plays a particularly important role because it is the particles’ disposition,43 taken in the structural sense of geometrical configuration, which is in many cases supposed to account for the operations of the agent. The issue that interests Boyle is not whether these properties are physical or not (they certainly are), but rather in which ontological category they belong: are they substances or accidents? The doctrine of real accidents, for Boyle, is a contradiction in terms: it construes qualities as substances that ‘may exist separate from all matter whatsoever’,44 as entities that do not need other entities to exist, hence really as substances. Note that Boyle’s objection against the scholastic doctrine is, in his own view, grounded on a well established Aristotelian distinction. It is of no fundamental importance whether these entities are spiritual or physical: even a real physical accident is an absurdity. The important point is that qualities are modes (or moods) of matter rather than entities, be it material entities, that are superadded to this matter. What about substantial forms? The very notion of real quality is rejected because it contains a straightforward logical contradiction, as we have just seen. Such is not the case with the notion of substantial form whose incoherence is established in a more indirect fashion. Even Descartes, anxious not to overly offend the authority of the School, occasionally uses the expression ‘substantial form’ to qualify the thinking substance. Nevertheless, Descartes does contest that the notion it expresses is well grounded, and as he himself says, his criticism is both metaphysical and theological.45 The scholastic tradition generally distinguishes the substantial form of humans, whose soul is created by God, and the substantial forms of natural substances, that come into existence through natural process; to this difference in origin corresponds a difference in destiny, insofar as the human soul survives the disappearance of the body to which it gives form, whereas all other substantial forms disappear together with the bodies to which they are attached. This difference in origin and destiny does not prevent the scholastics from considering that all substantial forms are separable from matter: this separability is thought to be natural in the case of the human soul, but requires divine intervention in the case of other substantial forms. However, as we have already seen, in the scholastic doctrine, separability does not imply 43 On this meaning of the word ‘disposition’, which has nothing to do with the dispositio as capacity, see Descartes, Les Principes de la philosophie, IV, §198, in Descartes, Oeuvres, vol. IX, p. 317; trans. Descartes (1984), I, 285. Descartes, Les Principes de la philosophie, trans. E.S. Haldane and G.R.T. Ross, The Principles of Philosophy, in The Philosophical Works of Descartes, vol. 1 (Cambridge, 1912), pp. 201-302. 44 R. Boyle (1772, p.16). The Origin of Forms and Qualities, in The Works, III, ed. Thomas Birch, (Hildesheim, 1966), p. 16. 45 We rely here essentially on the long letter to Regius from January 1642, in Descartes, Oeuvres, vol. III, pp.491-510; partial trans. Descartes (1984), III, 205-209.

Introduction

13

substantiality, such that it is possible to say both that substantial forms are separable and that they are nevertheless not themselves substances. Nevertheless, for reasons arising from his transformation of formal cause into a principle of operation and of unity of the substance, Suarez ends up considering separability not only as the sign of a real distinction, but also of substantiality. This is precisely the idea that Descartes rejects, just as he does in the case of real accidents. For him, as we have seen, separability is a condition sufficient for substantiality; hence, every substantial form must be considered as a separate substance. Therefore, there is no logical ground on which to justify the distinction between natural substantial forms and souls, and the distinction between natural separability and divine separability is simply not relevant. Now, for Descartes who attributes real separability only to the thinking substance, all non material substances must be considered to be really separate from matter. This means that the schoolmen attribute to substantial forms in general a property that belongs only to thought, so that the idea of non human substantial forms is the result of a regrettable confusion between thought and matter (and therefore, according to Descartes, extension). In this way, Descartes again justifies the idea that attributing a substantial form to natural, hence physical, things, is ipso facto giving material bodies a soul that alone accounts for their behaviour, their capacity to act and to undergo action in the way they do, and this argument once more accuses the scholastic doctrine of anthropomorphism. This accusation can also be found in Boyle’s writings;46 however, he makes the objection on purely conceptual grounds: if real accidents are considered to be detached from substances, they have been made substances and therefore substantial forms. Substantial forms are treated exactly like real accidents. For Boyle, contrary to what we have seen in Descartes’ argument, it does not matter whether a substantial form is physical or not. However, this consideration is supplemented by a nominalist thesis. Indeed, a substance in the ordinary sense of the term (hence with the exception of the fundamental elements of reality, which are the particles) is nothing else but an association of qualities.47 This does not prevent Boyle from thinking that, at a deeper level of analysis, these qualities themselves are grounded on primary qualities of the particles that really constitute a substance. Every substantial change can be explained by a local displacement of the material parts of the substance.48

46 ‘Now when any body is referred to any particular species (as of a metal, a stone, or the like) […] most of the writers of physicks have been apt to think, that besides the common matter of all bodies, there is but one thing that discriminates it from other kinds, and makes it what it is, and this, for brevity’s sake, they call a form: which, because all the qualities and other accidents of the body must depend on it, they also imagine to be a very substance, and indeed a kind of soul […]’, R. Boyle, The Origin of Forms and Qualities, p. 27. 47 Boyle, ibid., pp. 27-29. 48 Boyle, ibid., pp. 37-38.

14

Dispositions and Causal Powers

Elimination and/or Reduction of Powers Does all this mean that the notion of power has no longer any role to play? Insofar as Descartes and the tradition following him is concerned, it is undeniable that the notion of power is simply eliminated. This is because a power is a power to act, whereas the traditional category of action does not have any place in Descartes’ physics. First, every change is reduced to locomotion.49 Second and most important, Descartes holds that ‘motion is the transfer of one piece of matter […] to the vicinity of other bodies […]. I say “the transfer” as opposed to the force or action which brings about the transfer, to show that motion is always in the moving body as opposed to the body which brings about the movement’.50 No material body possesses the internal power of motion or rest, as the critique of substantial forms and real qualities has shown.51 Therefore, no body can act, and in particular act on other bodies. Motion is only a mode or state of the mobile body, which cannot be attributed to any other substance, neither to an internal substance (which would make it a soul), nor to an external material substance. As Martial Gueroult notes, Descartes leaves ‘force as a faculty of a subject, as a power to bring about a given effect, entirely aside, and takes into account only the acting force that is identified with its geometrically expressible effect’.52 However, by saying that force is always active or that power is always effective, one identifies power with its exercise and therefore makes nonsense of the distinction between power and exercise, which is necessary to give sense to the notion of power itself. The writings of Boyle or Locke who takes inspiration from him, show however that this elimination of powers is not at all straightforward. From the fact that the qualities which account for the powers of substances to act are not real accidents, it does by no means follow that they are fictions; one might just as well infer from the fact that Socrates’ size is not a substance, that it is not real. Here Boyle and Locke hit upon a problem that many contemporary philosophers try to solve (the present volume being intended to bear witness of some such efforts). In their texts, a crucial distinction is alluded to but never explicitly stated, between the power to accomplish a certain operation on one hand, such as the power of water to dissolve sugar, and on the other hand the property or quality (whose nature to reveal is the task of chemistry) that explains or accounts for that power, that capacity of water to dissolve sugar. That quality is permanent and exists clearly before the substance exercises its

49 See Principles of Philosophy, II, § 23, in Descartes, Oeuvres, vol. IX, p.75; trans. Descartes (1984), I, 232-233. 50 Principles of Philosophy, II, § 25, in Descartes, Oeuvres, vol. IX, p.76, trans. Descartes (1984), 1, 233, emphasis Descartes’. 51 This is also Descartes reason for rejecting the Aristotelian and scholastic distinction between natural and violent movement, i.e. between movements in agreement with and movements in opposition to the internal principle of change. See Descartes, Principles of Philosophy, II, §37, in Descartes, Oeuvres, vol. IX, p.85; trans. Descartes (1984), I, 241. 52 M. Guéroult, Leibniz. Dynamique et métaphysique (Paris, 1967), p. 67.

Introduction

15 53

power. It constitutes what is now called the categorical basis of the power. Now, the question arises whether a power is identical with its categorical basis. There certainly is a conceptual distinction (in contemporary vocabulary, one may also call it a semantical distinction) between a power and its categorical basis (knowing that water has the power to dissolve sugar is not itself sufficient for knowing the chemical composition of water). However, is this enough to show that there also exists an ontological distinction between them, a difference in the things themselves? Locke’s texts are far from clear on this point. Let us take a paradigmatic example of what one may call a dispositional property, i.e. a property that can be described in terms of its power to produce certain characteristic effects, which does not yet prejudge the issue of whether the property is identical with that power: colour. Any analysis of colour perception inevitably encounters the following classical problem: do they exist independently of being perceived? (This is the question of esse est percipi). Locke gives an affirmative answer,54 and takes, as does Boyle, the example of the porphyry that is white and red in daylight and that is taken to a dark cave. It nevertheless remains coloured, says Locke, because the colour of the porphyry is merely its disposition to produce in a perceiving subject the visual experience of red (the idea of red), and this disposition exists even when it happens not to be exercised. Locke does not ask explicitly whether one can reduce and more particularly identify this disposition with its categorical basis which he takes to be a ‘configuration of particles’. However, it seems that he makes this reduction because he proceeds by writing: ‘whiteness or redness are not in it [the Porphyre] at any time, but such a texture, that hath the power to produce such a sensation in us’.55 White is (identical with) a disposition, but also (identical with) a texture. Therefore, the disposition is (identical with) a texture. In fact, things are a good deal more complicated because, as the quoted text shows, it is not the porphyry itself that is said to exercise the power to produce certain visual sensations, a power that could be identified with a certain quality of the porphyry, i.e. its texture, but rather the quality of the porphyry… One can find the same hesitation in Boyle’s writings.56 However, the texts of the English scientist and philosopher contain traces of a thesis, never explicitly stated, that will soon get lost and is only rarely found in contemporary analyses of causal powers, namely the thesis that power is a relational notion. Boyle offers the example of the opening power or ‘virtus aperitiva’ possessed by a key, or more precisely, the 53 On this issue, see among others, the contributions by Mumford, Kistler and Tiercelin to this volume. 54 J. Locke, P.H. Nidditch (ed.), An Essay concerning Human Understanding, (Oxford, 1975), II, 8, § 19. 55 Ibid. 56 On Boyle, see in particular P. Alexander, Ideas, Qualities and Corpuscules (Cambridge, 1985) and P.R. Anstey, The Philosophy of Robert Boyle (London, 2000). More recently, S. Shoemaker and G. Molnar have discussed Boyle’s key-lock problem. See S. Shoemaker, ‘Causality and Properties’, in D.H. Mellor and A. Oliver (eds), Properties (Oxford, 1997), p. 243 and G. Molnar, S. Mumford (ed.), Powers. A Study in Metaphysics, (Oxford, 2003), p.102.

16

Dispositions and Causal Powers

capacity of a lock to be opened by a key.57 Think of Tugal Cain, the first blacksmith according to the biblical tradition. He makes a lock, the first one ever to exist. The lock can be described in a purely structural way, by the relative position of its parts. The key does not yet exist. However, let us imagine that he now makes a key that opens the lock. Once the key has been made, the lock acquires the power to be opened by the key (which therefore itself has the power to open the lock). Boyle argues that there is no power of the lock that gets added to the mutual position of its parts (the categorical basis of the power), because for the lock to acquire that power, no one has in any way acted on the lock (therefore it has not acquired any additional property in this sense). Boyle seems to rediscover at this point a notion that was familiar in the ancient and scholastic tradition: there is no relational change. Socrates is taller than Theaitetus is as a child. However, Socrates is smaller than Theaitetus as an adult. However, Socrates has not acquired any additional property, because it is precisely by not changing in size that he becomes smaller than Theaitetus. Only Theaitetus undergoes a change. But Boyle, more or less explicitly, adds an idealist touch to this consideration: the property to be smaller than Theaitetus is relational, which implies that it is not real at all, but rather the result of a comparison the mind operates between Socrates and Theaitetus. The same is true of the power of the lock: because it is only by virtue of its relation to the key that one attributes this property to it, one should conclude that the only ground of the power lies within the mind contemplating it and associating it to the key. However, one might also draw a different conclusion, which is suggested by some passages of Boyle’s own text. What the case of the first lock and key shows is that we cannot have the power of the lock to get opened by the key without the related power of the key to open the door. In short, it is logically impossible that there be only one power in the world. A power exists only in relation to another power. But if power is relational in nature, it cannot be simply identified with its categorical basis, for that basis clearly belongs only to the thing itself that possesses the power and can be attributed to it independently of whether any other thing exists: the lock itself has a certain physical conformation that cannot be altered by the mere existence of the key in addition to it. Locke himself recognizes the existence of both active and passive powers,58 although it is not clear whether he considers them to depend on each other. Only a few decades later, Thomas Reid finds the very idea of a passive power unintelligible.59 Indeed, relative beings have always been a major challenge for philosophical analysis.60

57 Boyle, The Origin of Forms and Qualities, pp. 18-19. 58 See, e.g. Locke, An Essay concerning Human Understanding, II, ch.21, §2. 59 ‘I conceive passive power is no power at all.’ T. Reid, Essays on the Active Powers of the Human Mind (Cambridge, 1969), p. 23. 60 On the history of the analysis of this notion in the Middle Ages, see M.G. Henninger, Relations (Oxford, 1989). The notion of relation has recently been analysed by V. Descombes, Les Institutions du sens (Paris, 1996), chs 16 and 17.

Introduction

17

Leibniz, a critical reader of Locke, never conceived of a disposition without the context of its ‘actuation’, to use one of his favourite terms for designating the effectiveness of a power. As Hide Ishiguro61 has noted, if a disposition such as tenderness reduced itself to the internal state of a particular, it would follow that it is possible that someone manifest tenderness independently of the circumstances making this manifestation intelligible. However, would we understand a person describing a sudden whiff of tenderness towards a glass of milk? A disposition is always attributed on a background of circumstances appropriate to its manifestation. Therefore, it cannot be reduced to a strictly internal state of its bearer. The Emergence of Dynamism There is another respect in which Leibniz plays a key role. As certain historians of science have noted, the materialist, corpuscular and mechanist analysis of Descartes and Boyle never succeeded in entirely extinguishing another tradition, that of a ‘dynamical’ conception of matter, of which Gilbert, with his De Magnete,62 was a somewhat naïve ‘precursor’ and which Leibniz has contributed to making respectable. Indeed, Leibniz can be seen to take up faithfully Aristotle’s analysis of change, and to state in particular that ‘power (puissance) […] in general’ is ‘the possibility of change’, that change ‘is action in one subject and passion in another’, and that therefore there are ‘two powers, one active and one passive’.63 However, beyond the fact that these traditional elements become integrated into a general metaphysical framework (that of pre-established harmony) that partially alters their meaning, the problems raised by the new physics lead Leibniz to elaborate on the distinction between active and passive power, in order to account for the notion of force. For matter in itself (which for Leibniz is equivalent to what is elsewhere called mass) is purely passive and could not by itself be capable of acting, in particular of moving. It only accounts for bodies’ impenetrability and inertia. To make sense of the idea that this prime matter has the force of moving, it must therefore be attributed to something analogous to a soul or a substantial form. It thereby becomes secondary matter.64 However, this theory need not necessarily be interpreted as a return to the separate substantial forms of the late scholastic tradition. Manifestly, Leibniz aims at a conceptual distinction between a simple (purely passive) capacity and what he calls a tendency,65 which is not yet the act itself, but something followed by the act if all 61 See H. Ishiguro, Leibniz’s Philosophy of Language (Cambridge, 1990), p. 111. Leibniz uses the example of tenderness in New Essays on Human Understanding, book II, ch. 22, §10, GP, V, 199-200; trans. Leibniz (1981), 215. 62 On W. Gilbert, see, e.g. the presentation by E.J. Dijksterhuis, The Mechanization of the World (Princeton, 1985), p.39ff. 63 Leibniz, New Essays on Human Understanding, II, ch. 21, §1, GP V, 155; trans. Leibniz (1981), 169. E.J. Dijksterhuis, The Mechanization of the World. 64 See Nature Itself, GP IV, p. 510 and p. 512; trans. Leibniz (1998), pp. 216-218. 65 See, e.g. Nature Itself, G. IV, p. 512; trans. Leibniz (1998), p. 218; New Essays on Human Understanding, II, ch. 21, §1, trans. Leibniz (1981), p. 169, and Theodicy, § 87, G IV, p. 149.

18

Dispositions and Causal Powers

obstacles that might hinder its coming about are taken away. The idea of a tendency or force is also one of the crucial points of Leibniz’ critique of Descartes’ account of the nature of movement. To the extent that one conceives of movement, as does Descartes, in a purely geometrical way that reduces it to the successive positions occupied at each instant by the moving object, it is impossible to distinguish, at a given moment, between Zeno’s arrow being at rest and Zeno’s arrow moving.66 In fact, the arrow bears at each instant a tendency to move towards a given direction later on.67 This critique of Cartesian geometrism leads Leibniz in the end to reject the primacy of extension in the account of matter insofar as it is in motion. Ultimately, extension itself gets assigned a derivative status with respect to force. What is fundamental in a corporeal substance is not its ‘extension’, but rather its capacity to act on other substances and to resist the action of other substances. According to some,68 Leibniz anticipates, at a metaphysical level, a line of thought that will find its physico-mathematical expression in the notion of (electromagnetic) field, by Faraday and above all Maxwell69 in the nineteenth-century. In their analysis of the intellectual evolution culminating in the notion of field, certain commentators have noted that Newton had clearly anticipated the necessity of such a notion, whereas from a philosophical perspective, Kant70 (after Boscovich) had followed the radical implications of this point of view to their end, by conceiving physics as grounded, rather than on an extension sprinkled with bodies, on two fundamental forces, attraction and repulsion.71 However, this perspective on the philosophy of science should not hide the fact that Leibniz, Newton and Maxwell raised a general ontological problem: is it

66 Leibniz, Nature Itself, §13, G IV, p. 513; trans. Leibniz (1998), pp. 218-219. 67 All this is related to the intricate issues of the relativity of movement and the nature of the relations between the quantity of movement (mv) and the ‘force vive’ (mv2). See, e.g. Leibniz, Discourse on Metaphysics, § 18, GP IV, 444; trans. Leibniz (1998), pp. 71-72. On the history of the problem raised by the possible relativity of movement, see J.B. Barbour, Absolute or Relative Motion (Cambridge, 1989), vol. 1 and on the problem of the status of the ‘force vive’, see the excellent M. Spector, ‘Leibniz vs. the Cartesians on Motion and Force’, Studia Leibniziana, 7, no. 1 (1975): 135-144, as well as R. Dugas, La Mécanique au 17ème siècle (Paris, 1954), p. 477ff. 68 E.g. R. Harré and E.H. Madden, Causal Powers (Oxford, 1975), p. 167. 69 Faraday was a great physicist and experimenter of the nineteenth century; Maxwell, a physicist sometimes considered to be on a par in importance with Newton, unified the theories of electricity and magnetism, taking his inspiration from Faraday’s discoveries, by constructing a powerful mathematical formalism that allowed him to master the notion of field. However, it should be noted that he tried to provide a mechanical model of the field. 70 See Kant, chapter II, pp. 40-94. 71 On all this, see L. Bloch, Les Origines de l’Ether et la physique de Newton (Paris, 1908) and J. Vuillemin, Physique et métaphysique kantiennes (Paris, 1955), in particular p.148ff. On the nineteenth-century and the beginning of the twentieth, see O. Darrigol, Electrodynamics from Ampère to Einstein (Oxford, 2000).

Introduction

19

possible to accept the existence of causal powers without any categorical basis?72 There are many reasons to ask this question, independent of the problem of the nature of the notion of field in physics. But it is safe to say that one of the historical sources of the problem lies in the metaphysical interpretation of the notion of force, as it has developed in the scientific context.73 Although Leibniz obviously does not possess the notion of field, his thought plays a central role for a different reason. Although he takes up, as we have already seen, Aristotelian language, his metaphysics does not allow a substance to act on any other substance, which excludes every form of transitive causation. Nevertheless, contrary to Descartes or Malebranche, he does not conclude from this that substances have no power to act, but rather that their action is immanent to the substance. Thus, every state of the substance is a consequence of its internal activity.74 The ‘connection (liaison)’75 and the correspondence between the apparent action of the agent on the patient (the opium’s action on man) and the apparent passion the patient undergoes on behalf of the agent (the man’s falling asleep that is brought about by the opium) reflect a general order of nature whose institution is of divine origin (this is Leibniz’ ‘pre-established harmony’). Here we have the beginning of the idea that causation between substances is first of all grounded on the existence of a law, an idea that will progressively be detached from its religious origin. The Appearance of the Notion of Law We hope that our brief presentation has shown to what extent it is an oversimplification of the history of the concepts of power and disposition to only say that they have fallen prey to Descartes’ attack against the application of the concept of substantial form to substances other than the human soul. Boyle and Locke propose to account for the powers of macroscopic bodies by attributing properties (‘primary qualities’) to their microscopic constituents, whereas the ‘dynamical’ conception of matter, such as that of Leibniz which led to the scientific notion of field, conceives all substances in terms of their powers and dispositions. Nevertheless, since the seventeenth century the use of these concepts has become marginal both in metaphysics and in 72 This is one of the issues that is the object of contemporary research. We shall come back to it later in this introduction. 73 R. Harré is one of the first philosophers to have brought together the general metaphysical issue of the nature of powers with the history of science and contemporary scientific practice, in particular in quantum physics. See, e.g. R. Harré, ‘Is There a Basic Ontology for the Physical Sciences?’, Dialectica, 51 (1997): 17-34. 74 ‘It can therefore be said, in a way, and in a good sense […] that one particular substance never acts on another particular substance any more than it is acted on by it. For consider: what happens to each one is only a consequence of its idea or complete notion and nothing else […]’ Leibniz, Discourse on metaphysics, § 14; GP iv, p. 440; trans. Leibniz (1998), p. 67. 75 Leibniz, Discourse on metaphysics, § 14; GP IV, p. 440; trans. Leibniz (1998), p. 67.

20

Dispositions and Causal Powers

philosophy of science. This is partially due to the fact that, in addition to Descartes’ critique, the use of the notion of cause in scientific explanation has also been questioned. The new experimental conception of science has given rise to a new conception of scientific explanation, in which the search for causes is considered to go beyond the legitimate reach of scientific inquiry. Since Galileo, the Aristotelian paradigm of explaining natural phenomena by the knowledge of their causes has been replaced by the nomological76 paradigm of explanation. Without yet using the word ‘law’, Galileo assigns the task of searching for causes to metaphysics, where it is condemned to remain victim of the arbitrariness of the different doctrines. According to the modern conception of science, inaugurated by Galileo, scientific explanations must be based on what is accessible to observation and experiment: this includes the regularities found among natural phenomena, but not their causes. ‘The present does not seem to be the proper time to investigate the cause of the acceleration of natural motion concerning which various opinions have been expressed by various philosophers [...]. At present it is the purpose of our Author merely to investigate and to demonstrate some of the properties of accelerated motion (whatever the cause of this acceleration may be) […].’77 Henceforth, the concept will be associated with the postulate of an obscure entity that escapes observation and therefore experimental control. The new conception of explanation can be found expressed in Berkeley: ‘It is not, however, in fact the business of physics or mechanics to establish efficient causes, but only the rules of impulsions or attractions, and, in a word, the laws of motions, and from the established laws to assign the solution, not the efficient cause, of particular phenomena’.78 Rather than as a search for its causes, the explanation of an event or phenomenon is now conceived of as a demonstration that it belongs to a regular sequence of events or phenomena of the same type. By conceiving natural regularities as the expression of underlying laws, the fact that a singular phenomenon abides by a certain law, suffices to explain it: one has accounted for the phenomenon as soon as one has shown that it had to come about by virtue of a law of nature. Auguste Comte expresses a similar idea when he opposes mysterious causes to laws that are transparent to the mind because they are accessible to experience: ‘The fundamental revolution that characterizes the virility of our intelligence consists essentially in substituting everywhere, for the inaccessible determination of causes properly speaking, the simple search for laws, i.e. of the constant relations that exist between observed phenomena. Whether we are concerned with the smallest or with the most sublime effects, those of shock and heaviness or those of thought and 76 The word ‘nomological’, from the Greek ‘nomos’ – the law – and ‘logos’ – the doctrine, is the adjective corresponding to the noun phrase ‘law of nature’. A nomological statement is a statement (i.e. a sentence uttered on a given occasion) expressing a law of nature. 77 G. Galilei (1638), ‘Discorsi e dimostrazioni matematiche intorno a due nuove scienze’, in Le Opere di Galileo Galilei (Firenze, 1898), p. 202; and trans. H. Crew and A. de Salvio, Dialogues Concerning Two New Sciences (New York, 1963), p. 160. On Galileo’s conception of causation and explanation, see E. Yakira, La causalité de Galilée à Kant (Paris, 1994). 78 G. Berkeley, ‘Of Motion (De Motu)’, in A.A. Luce and T.E. Jessop (eds), The Works of George Berkeley, Bishop of Cloyne, (London, 1951), § 35, p. 40.

Introduction

21

morality, we can really know only the different mutual relations characterizing their accomplishment, without ever penetrating the mystery of their production’.79 Hume’s Analysis of Causation and the Disappearance of the Notion of Power Hume attacks the traditional concept of cause and the obscure notion of a power necessarily generating effects from a different angle: instead of banishing the concept of cause itself, Hume proposes to purify it by isolating its rational core. Hume does not say that there are no causes or that causation defies rational understanding. Rather, Hume’s project is to conceive of causation in a new way, as a relation satisfying certain constraints. According to his famous definition, a cause is ‘an object precedent and contiguous to another, and where all the objects resembling the former are placed in like relations of precedency and contiguity to those objects, that resemble the latter’.80 The cause is not any more conceived according to the model of a substance that acts: from Hume’s perspective, something is a cause only relative to a given effect.81 Hume’s substitution of the concept of cause as the term of a relation to its conception as an active principle has exercised an enormous influence on the subsequent history of philosophical thinking about causation. In the tradition begun by Hume, causes are considered as events whose status as causes is only determined by their relation to other events, their effects. True, Hume calls causes ‘objects’ rather then ‘events’. However, it has been argued82 that the conception of causation as a relation leads naturally to conceiving the terms of this relation as events, i.e. as entities whose identity is essentially determined by their time of occurrence. One can reach that conclusion by the following reasoning. A causal process goes on at a particular time: if a stone breaks a window, it is true that, at first look, one will 79 A. Comte, Discours sur l’esprit positif (Paris, 1983), pp. 19-20. ‘La révolution fondamentale qui caractérise la virilité de notre intelligence consiste essentiellement à substituer partout, à l’inaccessible détermination des causes proprement dites, la simple recherche des lois, c’est-à-dire des relations constantes qui existent entre les phénomènes observés. Qu’il s’agisse des moindres ou des plus sublimes effets, de choc et de pesanteur comme de pensée et de moralité, nous n’y pouvons vraiment connaître que les diverses liaisons mutuelles propres à leur accomplissement, sans jamais pénétrer le mystère de leur production.’ 80 Hume (1739), L.A. Selby-Bigge and P.H. Nidditch (eds), A Treatise of Human Nature (Oxford, 1955), p. 170; similarly p. 172. See also Hume (1748), ‘An Enquiry Concerning Human Understanding’, in L.A. Selby-Bigge and P.H. Nidditch (eds), Enquiries Concerning Human Understanding and Concerning the Principles of Morals (Oxford, 1975), p. 76. 81 One could say that Hume takes up Boyle’s above-mentioned thesis of the essentially relational character of cause and effect – each being what it is only by virtue of its relation to the other – but gives it a new, non dispositional meaning. For Hume, the cause is not any more an active power and the effect is not any more a passive capacity; rather, both are events manifesting themselves as phenomena and bringing about ideas of sensation in subjects observing them. 82 On the contemporary development of the conception of causes and effects as events, see D. Davidson, Essays on Actions and Events (Oxford, 1980) and T. Parsons, Events in the Semantics of English (Cambridge, 1990).

22

Dispositions and Causal Powers

identify an object, the stone, as the cause of the window’s breaking. But the logic of the relational conception of causation makes us look more closely: the effect is not the window, an object, but rather an event: the window’s breaking. It is essential for a window’s breaking that it occurs at a given moment; more generally, the conditions of identity of events necessarily contain a moment (or a stretch) of time. Once one looks at the situation from this perspective, it becomes clear that the stone is only in a direct causal relation to the breaking of the window at a precise moment: that which just precedes the breaking. It seems therefore inevitable to consider that strictly speaking it is not the stone as such that is the cause, but rather an event happening to the stone: the stone’s flying with a given speed in a given direction at the moment preceding its impact on the window. Logical Empiricism and the Deductive-Nomological Model Logical empiricism (also called ‘neopositivism’) developed primarily by the philosophers of the Vienna Circle in the first decades of the twentieth-century is heir to both of these traditions: following Galileo and Comte, the logical empiricists’ model of scientific explanation requires that such explanations be grounded on laws of nature. According to their ‘deductive-nomological’ model, which belongs to the heart of the doctrine of logical empiricism and whose classical formulation is due to Hempel and Oppenheim,83 an explanation is a deductive argument whose premises (the explanans – that which explains) contain first ‘statements of antecedent conditions’ (also called ‘initial conditions’) and second ‘general laws’,84 whereas the conclusion (the explanandum – that which is to be explained) contains a ‘description of the empirical phenomenon to be explained’ (ibid.). To explain scientifically is to show that a phenomenon or event is a logical consequence of the situation (described in the initial conditions) and the laws.85 For Hume and his followers, any succession of events that happens regularly and satisfies the constraints of contiguity and succession is ipso facto causal, so that causation does not contain any ‘thickness’ (to use Simon Blackburn’s expression) or obscurity going beyond the transparency of observable regularities. All explanations making use of laws (or ‘nomological’ explanations) are therefore causal explanations, without any implication that causes are something over and above the terms of observable regularities. Abstracting away from the regular relation that makes it a cause, a cause is simply an observable phenomenon. Leaving aside Hume’s restrictions of succession and contiguity, the 83 C.G. Hempel and P. Oppenheim, ‘Studies in the Logic of Explanation’, in C.G. Hempel, Aspects of Scientific Explanation and other Essays in the Philosophy of Science (New York Free Press, 1965). 84 Ibid., p. 249. 85 Strictly speaking, one should say that a certain description of the event to be explained is a logical consequence of a linguistic expression of the initial conditions and of the laws. However, insofar as there is no risk of misunderstanding, we shall nevertheless use the ‘material mode’ of speaking, instead of the ‘formal mode’, to use Carnap’s terminology, for the sake of its greater simplicity and naturalness.

Introduction

23

logical empiricists end up simply identifying nomological explanation with causal explanation. Carnap expresses this doctrine at a semantic level, as an analysis of the meaning of the word ‘cause’: ‘What is meant when it is said that event B is caused by event A? It is that there are certain laws in nature from which event B can be logically deduced when they are combined with the full description of event A’.86 Popper, though in many respects critical of logical empiricism, also identifies causal explanation with deductive-nomological explanation as such: ‘To give a causal explanation of an event means to deduce a statement which describes it, using as premises of the deduction one or more universal laws, together with certain singular statements, the initial conditions’.87 The Logical Empiricist’s Suspicion about Dispositional Terms Logical empiricists insist above all on the importance of clarity in philosophy. In their writings, the traditional suspicion of obscurity with respect to dispositions takes the form of doubt as to whether the meaning of dispositional predicates is well determined. What exactly does it mean to attribute a disposition to an object? The background of such doubts about the completeness of the meaning of dispositional predicates is the verificationist theory of meaning. According to verificationism, the meaning of a predicate consists in its conditions of application, to be spelled out in observable vocabulary. A predicate is categorical if those application conditions do not have a conditional form. ‘This cube is of sugar’ attributes to the sugar cube the categorical predicate ‘is of sugar’. ‘Is a square’ is another categorical predicate:88 a surface is square if all of its sides are of equal length, where ‘… and … are of equal length’ is itself a categorical predicate (with two places). However, ‘is soluble’ is not a categorical but rather a dispositional predicate, because the criterion of solubility has a conditional form: to say of something that it is soluble (in water) means that if it was put into water (in certain circumstances) it would dissolve. That this conditional character of the predicate, and thereby its dispositionality, belongs to its meaning is shown by

86 R. Carnap, Philosophical Foundations of Physics (New York, 1966), p. 194. 87 K. Popper, Logik der Forschung (Tübingen, 1969), p. 31, trans. p. 59; italics Popper’s. This doctrine exists in many variations. Davidson argues for the equivalence between a particular causal relation and the existence of a general law allowing us to deduce a description of the effect from an appropriate description of the cause, in the context of our reasons to believe in the truth of particular causal statements: ‘It does not follow that we must be able to dredge up a law if we know a singular causal statement to be true; all that follows is that we know there must be a covering law. And very often, I think, our justification for accepting a singular causal statement is that we have reason to believe an appropriate causal law exists, though we do not know what it is.’ D. Davidson, ‘Causal Relations’, in D. Davidson, Essays on Actions and Events (Oxford, 1980), p. 160. 88 As we shall see later, not all philosophers agree that geometrical properties are exclusively categorical.

24

Dispositions and Causal Powers

the fact that there is no categorical predicate having the same meaning as ‘is watersoluble’. What exactly is the alleged obscurity of such dispositional predicates, that justifies the logical empiricists’ judgement that they should not be used in scientific explanations? Imagine hesitating to attribute the dispositional predicate ‘is watersoluble’ to a given object. If the object undergoes the appropriate test, in this case, if it is put into water, all is well, because in the circumstances of the test, the application conditions of the predicate are not conditional: while it undergoes the test, the object is water-soluble if it dissolves and not water-soluble if it does not dissolve. The problem arises in circumstances in which the object under consideration does not undergo any appropriate test. What determines in this case whether the object is water-soluble? In his seminal paper, Testability and Meaning,89 Carnap replies that nothing does and that therefore the meaning of dispositional predicates is only partially determined. The logical empiricists of the beginning of the twentieth-century thus seem to put forward a new argument against dispositions. To the traditional accusation of epistemic obscurity, they add that of semantic indeterminacy: the meaning of predicates attributing dispositions is only partially determined, and the indeterminacy concerns precisely those circumstances in which the need for dispositions is the most pressing: where they do not manifest themselves. As a result, the conception of explanation in terms of nomic regularities promoted by the logical empiricists leaves no place for dispositions and causal powers. The predicates expressing lawful regularities must be categorical, for the simple reason that the meaning of the terms used in an explanation must at least be as clear as the meaning of the terms describing the phenomena to be explained. But descriptions of phenomena are always in categorical terms. Can Dispositional Predicates be Dispensed with? Goodman’s Answer There are also new reasons to justify the use of dispositional terms. The most important is the observation that dispositional predicates are omnipresent not only in our everyday way of talking but also in scientific discourse: ‘Besides the observable properties it exhibits and the actual processes it undergoes, a thing is full of threats and promises. The dispositions or capacities of a thing – its flexibility, its inflammability, its solubility – are no less important to us than its overt behaviour’.90 To justify this thesis, Goodman explains that the dispositional character of a predicate is not 89 R. Carnap, ‘Testability and Meaning’, I, Philosophy of Science, 3 (1936): 419-471. 90 N. Goodman, Fact, Fiction, and Forecast (Indianapolis 1973), p. 40. More recently, Rom Harré (R. Harré, ‘Is There a Basic Ontology for the Physical Sciences?’) and Nancy Cartwright have insisted on the central and ineliminable role of the reference to dispositions, capacities and powers in natural science. Here are the first sentences of Cartwright’s book Nature’s Capacities and their Measurement: ‘Science is measurement; capacities can be measured; and science cannot be understood without them. These are the major three theses of this book.’ (N. Cartwright, Nature’s Capacities and their Measurement (Oxford, 1989),

Introduction

25

always manifest in its morphology: many predicates whose morphology contains no explicit sign of dispositionality such as a suffix in ‘-able’ or ‘-ible’, have nevertheless a dispositional meaning. Goodman even holds that almost all predicates really have a dispositional meaning, the exception being predicates describing events, such as ‘bends’, ‘breaks’, or ‘burns’. However, most predicates attributing lasting properties are dispositional. ‘If a flexible object is one capable of bending under appropriate pressure, a hard object is one capable of resisting abrasion by most other objects. And for that matter, a red object is likewise one capable of certain color-appearances under certain lights.’91 Even geometrical figure is attributed by a dispositional predicate, to the extent that the application condition of predicates expressing figure can be expressed in conditional form: such a predicate describes what one would find if an appropriate act of measurement was executed on the object: ‘a cubical object is one capable of fitting try squares and measuring instruments in certain ways’.92 Once the omnipresence of dispositional terms has been recognized, both in scientific and everyday discourse, one may take one of two attitudes, adopted respectively by Goodman and Quine. The point of departure for both is the conviction that Carnap has conclusively shown that no complete analysis of the meaning of such predicates can be given in the framework of first order predicate logic.93 To the extent that one restricts the use of propositional connectors to negation, conjunction and disjunction, and material implication (‘⊃’), it is impossible to analyse the p. 1. She develops this thesis in her most recent book: The Dappled World, A Study of the Boundaries of Science (Cambridge, 1999). 91 Goodman, Fact, Fiction, and Forecast, p. 40. 92 Ibid., Goodman’s thesis is controversial. Popper generalizes it even more than Goodman: ‘We can say that all physical (and psychological) properties are dispositional. That a surface is coloured red means that is has the disposition to reflect light of a certain wavelength. That a beam of light has a certain wavelength means that it is disposed to behave in a certain manner if surfaces of various colours, or prisms, or spectrographs, or slotted screens, etc., are put in its way.’ (K. Popper, ‘The Propensity Interpretation of the Calculus of Probability, and the Quantum Theory’, in S. Körner (ed.), Observation and Interpretation (London, 1957), p. 70). Shoemaker (1980), Mumford (1998) and Mellor (2000) show that the paradoxical conclusion that all properties are dispositional can be avoided by conceiving of this distinction as of a semantic distinction between predicates, rather than as an ontological distinction between properties: those predicates are dispositional whose meaning implies analytically (i.e. merely by virtue of the predicate’s meaning) that certain counterfactual conditionals are true of the objects to which the predicate applies. ‘X is fragile’ implies analytically that, if x was put into water, x would dissolve. Attributions of categorical predicates also imply such counterfactual conditionals, but in their case the implication is always a posteriori: its truth is not guaranteed to be the mere meaning of the predicate, but rather by laws (known only a posteriori) applying to the property. S. Shoemaker (1980), ‘Causality and Properties’, repr. in S. Shoemaker, Identity, Cause and Mind (Cambridge, 1984), pp. 206-233. S. Mumford, Dispositions (Oxford, 1998). D.H. Mellor, ‘The Semantics and Ontology of Dispositions’, Mind, 109 (2000): 757-780. See the articles by Kistler and Mumford in this volume. 93 First order predicate logic governs the usage of the quantifying expressions, or ‘quantifyiers’, such as ‘all’, ‘some’or ‘no’.

26

Dispositions and Causal Powers

application conditions of a dispositional predicate in purely categorical terms, as far as objects not subject to any test for the disposition are concerned. In other words, one can only hope to find an analysis of the meaning of dispositional predicates if one uses a logical language stronger than standard first order logic. This is precisely the research tradition initiated by Goodman: his starting point is the observation that dispositions ‘strike us by comparison [with observable properties] as rather ethereal. And so we are moved to inquire whether we can bring them down to earth; whether, that is, we can explain disposition-terms without any reference to occult powers.’94 According to Goodman’s hypothesis, which constitutes until today the framework for this debate, the meaning of dispositional predicates can be analysed in categorical terms with the help of counterfactual conditionals: ‘Replacement of a statement like k was flexible at time t by a statement like If k had been under suitable pressure at time t, then k would have bent, has obvious promise as a step towards clarification.’95 However, the semantics of counterfactual conditionals, which cannot be reduced to that of formulas of standard first order logic, itself raises some hard problems. The truth value of statements of counterfactual conditional form is not simply a function of the truth values of their antecedent and consequent propositions: they are not ‘truth-functional’ propositions. Since Goodman’s seminal essay, considerable progress has been made in the analysis of the semantics of counterfactual conditionals, and these advances have fuelled the research about the analysis of the meaning of dispositional predicates in those terms. However, the standard way to carry out this analysis uses the concept of possible worlds, the interpretation of which, notwithstanding its formal clarity, raises problems difficult enough for some to judge the whole project a failure.96 Counterfactual analysis in terms of possible worlds proceeds in this manner. An object that is not subject to any test for flexibility such as applying a force on it, is flexible in the actual world w0, if and only if it bends without breaking in the nearest possible world in which it is subject to that test. In the context of the elucidation of the meaning of dispositional predicates, the most important problem this approach encounters is that of circularity. How do we know how a given object behaves in other possible worlds? Instead of having any independent grasp on the fact that the objects bends in a given non-actual possible world w1, in which it is subject to the testing force, the only ground we can have for judging that it bends in w1 is our judgement that it is flexible in the actual world w0. It may turn out to be true in general that attributing to objects in other possible worlds manifest properties they do not possess in the actual world, is just another way of attributing it the property in 94 N. Goodman, Fact, Fiction, and Forecast, p. 40. 95 Ibid., pp. 34-35. 96 See Schmitz’ contribution to this volume.

Introduction

27

a dispositional way, in the actual world. Goodman at any rate is convinced that the ‘problems of dispositions, counterfactuals, and possibles’97 are inextricably bound up with each other, so that any analysis of one of them necessarily makes use of the others. This does not however prevent him from thinking that even if such analyses are inevitably circular because the analysis of each of these concepts uses the others, they may nevertheless be useful for clarifying the meaning of each. … and Quine’s Quine on the contrary holds that the only way to give a clear sense to dispositional predicates is to break out of the circle made up by the concepts of disposition, counterfactual, and the modal notions of possibility and necessity. Quine’s thesis is that the attribution of a dispositional predicate to an object reflects our incomplete knowledge of that object: when we call the sugar ‘soluble’, we in fact attribute a categorical property to it, though one that we know only incompletely. The real properties making true attributions of dispositions are categorical properties of ‘microstructure’.98 By using a dispositional predicate, says Quine, ‘we can refer to a hypothetical state or mechanism that we do not yet understand’.99 From his point of view, to attribute a disposition to an object is just a provisional way to characterize it faute de mieux, awaiting a more direct way. There are really no dispositional properties, the dispositional being only an incomplete and provisional way to characterize microstructural categorical properties. A disposition is but a ‘partially discerned physical property that will be more fully identified, we hope, as science progresses’.100 This is also the position David Armstrong defends under the name ‘categoricalism’: according to this doctrine, what makes true the attribution to an object x of a dispositional predicate D is the possession by x of a categorical property. In Armstrong’s own words, ‘all true properties […] are non-dispositional’.101 The Contemporary Ontological Debate At this point, the debate takes an ontological turn. For Quine’s suspicion towards dispositional properties is really grounded in his scepticism about the idea of pure potentiality. Several authors have defended the thesis that only causally efficacious properties can be said to be real, and that causal efficacy of a property is equivalent to its being non-dispositional or categorical.102 Against Shoemaker, who holds on 97 N. Goodman, Fact, Fiction, and Forecast, p. 33. 98 W.V. Quine, The Ways of Paradox and Other Essays (Cambridge, 1976), p. 73. 99 W.V. Quine, The Roots of Reference (LaSalle, 1971), p. 10. 100 Ibid., p. 13. 101 D.M. Armstrong, A World of States of Affairs (Cambridge, 1997), p. 80; see also D.M. Armstrong, A Materialist Theory of the Mind (London, 1968), pp. 85-88. 102 Cf. P.J. Holt, ‘Causality and Our Conception of Matter’, Analysis, 37 (1976): 20-29; H. Robinson, Matter and Sense (Cambridge, 1982); R. Swinburne, ‘Reply to Shoemaker’, in L.J. Cohen and M. Hesse (eds), Aspects of Inductive Logic (Oxford, 1983), pp. 313-320; D.M.

28

Dispositions and Causal Powers

the contrary that there is nothing in a property beyond the dispositions it lends its bearer, Armstrong argues in the following way: ‘Every causal transaction, according to Shoemaker, is a matter of things with certain causal potentialities bringing about that these or other things have further potentialities, because properties are analyzed as nothing but potentialities. In Scholastic language, we never get beyond potency to act. Act, so far as it occurs, is just a shifting around of potencies. And is this a believable theory?’103 In the opposite camp, Shoemaker,104 Molnar,105 Lierse and Ellis106 argue that the use of the causal criterion of reality according to which what is causally efficacious is real and what is real is causally efficacious, leads to the doctrine opposite to Quine’s and Armstrong’s categoricalism: according to their ‘dispositionalism’, to conceive of a property as real means to consider that it is capable of acting causally and bringing about effects. This dispositionalism does not deny that, e.g. the dispositional property of being soluble is grounded in the microscopic structure of the soluble substance. It only rejects the metaphysical thesis that those grounding properties are necessarily categorical. On the contrary, to the extent that they are real, and therefore causally efficacious, those grounding properties are themselves essentially dispositional: their identity is essentially determined by the effects they are capable of bringing about in different circumstances. Mellor has offered a different defence of dispositions against the argument, put forward in particular by Quine, that the fact that their identity is determined by reference to possibilities sheds doubt on their reality. Mellor observes that such a link to the possible does not distinguish dispositional from categorical properties, because attributions of the latter imply just as well counterfactual conditionals, expressing how their bearers would behave in possible but non actual circumstances. Even a paradigmatic categorical property like that of being triangular appears as dispositional according to that criterion, because its attribution to an object implies the counterfactual conditional that ‘if the corners were (correctly) counted the result

Armstrong, ‘The Causal Theory of Properties: Properties according to Shoemaker, Ellis and others’, Metaphysica, 1 (2000): 5-20. See also Mumford’s paper in this volume. 103 D.M. Armstrong, ‘The Causal Theory of Properties: Properties according to Shoemaker, Ellis and others’, p. 14. 104 S. Shoemaker, ‘Causality and Properties’. 105 G. Molnar, ‘Are Dispositions Reducible?’, The Philosophical Quarterly, 49 (1999): 1-17 and G. Molnar, S. Mumford (ed.), Powers. A Study in Metaphysics (Oxford, 2003) 106 B.D. Ellis, and C. Lierse, ‘Dispositional Essentialism’, Australasian Journal of Philosophy, 72 (1994): 27-45; B.D. Ellis, ‘Causal Powers and Laws of Nature’, in H. Sankey (ed.), Causation and Laws of Nature (Dordrecht, 1999), pp. 19-34; B.D. Ellis, ‘The New Essentialism and the Scientific Image of Mankind’, Epistemologia, 23 (2000): 189-210; B.D. Ellis, Scientific Essentialism (Cambridge, 2001). See also M. Kistler, ‘The Causal Criterion of Reality and the Necessity of Laws of Nature’, Metaphysica, 3 (2002): 57-86. In this volume, Mumford and Kistler defend the thesis that there exist at least some dispositional properties lacking a categorical basis.

Introduction 107

29

would be three’. Their link to non actual possibilities cannot be used to deny the reality of dispositional properties, on pains of paradoxically denying reality to all properties. One might have expected that dispositions would have been rehabilitated automatically once it was recognized that verificationism, that extreme form of an empiricist theory of meaning, imposed too strong restrictions on the meaningfulness of expressions, in particular of theoretical terms. The predicate ‘has a spin of ½’ cannot be defined by means of a direct and unique test; but nevertheless physical theory, indirectly controlled by experimental tests, gives it a precise meaning. The conditions of attribution of dispositions appear to be neither more nor less problematic than the conditions of attribution of theoretical properties in general: the attribution of the property of having a spin of ½, not being directly observable, necessarily presupposes accepting a theory giving it a well determined meaning and specifying the conditions (ultimately but indirectly observational) under which it is legitimate to attribute it to a given object. Attributing a disposition to an object that does not manifest itself seems to obey essentially the same logic: the property of being water-soluble cannot be attributed to a given dry sugar cube on the basis of a direct observational criterion. However, this does not imply at all that its attribution is arbitrary; it is simply based on indirect criteria, exactly in the same way as the attribution of theoretical properties. If one already knows that a given object is an electron, one may attribute it the property of having a spin of ½, on the basis of the law that all electrons have this property. In the same way, it is enough to know that the cube in the sugar bowl is made of sugar to have a ground for attributing it the property of being soluble, by virtue of the lawful relation between the property of being of sugar and the property of being water-soluble. However, dispositions remained suspect even after Quine and others had convinced the philosophical community that the distinction between theoretical and observational vocabulary is not sharp but rather a matter of degree, and that therefore there was no reason to think that the theoretical and the observational were two types of terms with radically different meaning and application conditions. The dismissal of this corollary of verificationism took away the major obstacle standing in the way of seeing that theoretical terms can have a definite meaning without being directly and explicitly definable in observational vocabulary. However, the conditional nature of the meaning of dispositional predicates continued being a reason for doubting that that meaning was completely determined. To say of a substance that it is water-soluble, is not equivalent to saying what it is or what it does, but rather what it would be or what it would do were it subjected to certain tests. It cannot be denied that the conditional element in the meaning of dispositional predicates makes their attribution more problematic than the attribution of categorical predicates. However, the controversy about the status and origin of this special nature divides the accounts of dispositions presently defended in analytical metaphysics in two opposite camps. For those that may be called ‘realists’ about 107 D.H. Mellor, ‘In Defence of Dispositions’, Philosophical Review, 83 (1974): 171.

30

Dispositions and Causal Powers

dispositions, the difference is only epistemological: dispositions themselves are real and complete, in the sense of having causal efficacy; only our knowledge of them is incomplete. For the ‘anti-realists’ in the opposite camp however, the incompleteness of the meaning of dispositional predicates is a sign of the fact that dispositional properties are not real properties at all. In this sense, Prior, Pargetter and Jackson, in a famous paper on dispositions,108 argue that dispositions have no causal efficacy whatsoever and should therefore not be considered as real properties. What executes in their view the ‘causal work’ of making the sugar dissolve in the water, is not its solubility but rather a microscopic property underlying the disposition, called, as we have already seen above, the ‘categorical basis’ of the disposition. In the case of sugar, the basis may consist in a physico-chemical property of the sugar crystals. It is essential for the anti-realist position to conceive of dispositions as properties distinct from their basis. The major drawback of anti-realism is that it makes dispositions causally inert. Thereby it goes against what many think to be central to our conception of ourselves as agents: according to the so-called causal theory of action, as defended in particular by Davidson,109 our actions are not only motivated, justified, and ‘rationalized’ by our beliefs, desires and other mental states, but also causally brought about by them. This conviction is incompatible with anti-realism if those mental states are dispositional in nature. To get around this difficulty, Armstrong has proposed to make the hypothesis that mental dispositions are identical with their categorical bases. As the latter are no doubt causally efficacious, so are the mental dispositions identical with them. In the case of my mental disposition of knowing that Paris is the capital of France, Armstrong’s hypothesis is that it is identical with a categorical property of my brain. On this count, he agrees with Quine that it is only our knowledge of dispositional properties that is provisional and incomplete, whereas the property itself is complete and causally efficacious. Armstrong’s ‘identity theory’ of dispositions can be made even stronger by supplementing it with the above-mentioned thesis held among others by Shoemaker,110 Mumford111 and Mellor,112 that the distinction between the dispositional and the categorical is first of all a semantic distinction between two types of predicates, rather than an ontological distinction between types of properties. Just as individuals can be named with different proper names, such as ‘Venus’ and ‘Phosphorus’, so causally efficacious properties can be designated in two ways: either indirectly and conditionally by a dispositional predicate, or directly by a categorical predicate.

108 E. Prior, R. Pargetter and F. Jackson, ‘Three Theses about Dispositions’, American Philosophical Quarterly, 19 (1982): 251-257. 109 D. Davidson, ‘Actions, Reasons, and Causes’, in D. Davidson, Essays on Actions and Events. 110 S. Shoemaker, ‘Causality and Properties’. 111 S. Mumford, Dispositions (Oxford, 1998). 112 D.H. Mellor, ‘The Semantics and Ontology of Dispositions’, Mind, 109 (2000): 757780.

Introduction

31

The papers collected in the first part of this volume bear testimony of the vivid debate about the nature of dispositions and causal powers in contemporary metaphysics. The main thesis common to all of them is that dispositions and causal powers are an essential and indispensable part of our conceptual scheme. In his paper, Schmitz (Chapter 1) shows in detail that Goodman’s strategy of analysing attributions of dispositional predicates in terms of counterfactual conditionals founders on essentially the same problems that Carnap has discovered for their analysis in terms of material conditionals. Evaluating the relevant counterfactual in cases in which an object is not subject to any test in the actual world already presupposes knowledge of whether it possesses the disposition; if the object is of a type whose members we know to possess the disposition because other objects of that type have been subjected to tests, we can attribute the disposition to it; however, as Carnap has pointed out, the attribution of dispositions to objects belonging to kinds none of whose members have been subjected to any tests is indeterminate; the contemplation of non-actual possible worlds does not provide any information allowing us to overcome this indetermination. What is worse, the analysis of dispositions in terms of counterfactual conditionals raises new problems: Schmitz shows that, under certain plausible conditions, dispositional predicates do not obey an essential condition of empirical meaningfulness, that the principle of Excluded Middle be an analytical truth. On the counterfactual analysis, it does not come out as analytically true that, for a given dispositional predicate P, for all objects x, either x satisfies P or it does not. After these semantic arguments showing that dispositional predicates cannot be reduced to categorical predicates, even with the help of counterfactual conditionals, Mumford and Tiercelin (Chapters 2 and 3) argue on ontological grounds for considering dispositions as real properties in their own right. Mumford (Chapter 2) begins by setting out his ‘ungrounded argument’ for the thesis that there are ungrounded dispositions, i.e. dispositions that are not reducible to any categorical, and more basic, properties of their bearers. His main premises are that there are absolutely simple sub-atomic particles that have dispositional properties, and that the dispositions of an object can only be grounded on the properties of its constituents. The main part of his paper is dedicated to an analysis, and rebuttal, of the reasons for which the idea of an ungrounded disposition has been considered to be unacceptable within the Humean tradition. In this tradition, only categorical properties are considered to be able to provide truthmakers for attributions of predicates; in the absence of any such categorical properties, nothing ‘fills the space’ at which the possessor of the disposition is located, which is just a way of saying that ungrounded dispositions lack being, or reality. Mumford shows that neither Hume nor the prominent contemporary Humean, David Lewis, provide any strong reasons for denying that dispositions can ground other properties, nor for denying more generally that dispositions can be properties in their own right. On the contrary, Plato’s ‘causal criterion of reality’ according to which that which can affect other things causally is real (Mumford calls it ‘the Stranger’s test of reality’), pleads for considering dispositions as real, insofar as they are causal powers.

32

Dispositions and Causal Powers

Tiercelin (Chapter 3) begins by rehearsing the main tenets of ‘dispositional essentialism’ (or ‘scientific essentialism’) as recently defended by Brian Ellis and others. First, natural kinds of events and processes constitute a metaphysical category that is as fundamental as natural kinds of objects or substances. Second, it is fundamentally its powers, i.e. what a thing is disposed to do, that makes it the kind of thing it is. Third, a thing’s causal powers, showing themselves in laws of interaction, are essential to it, which means that those laws are necessary. In a second step, Tiercelin compares Ellis’s version of dispositional essentialism with Mumford’s. Mumford’s contribution is first of all to have clarified the semantic link between dispositional predicates and counterfactual conditionals: ascriptions of both categorical and dispositional predicates imply attributions of functional roles which in turn imply certain counterfactual conditionals. The difference is that the link between the predicate and the functional role is analytic in the case of dispositional predicates but a posteriori in the case of categorical predicates: in the case of the former but not the latter it is part of the predicate’s meaning that it corresponds to a functional role. Mumford shows that this implies that laws of nature have a derivative status with respect to dispositions, instead of being imposed ‘from outside’ on otherwise unconnected events, as the Humean tradition has it. Tiercelin argues that grounding laws on dispositions may solve the problem of exceptions (a thesis contested later in the volume by Schrenk): if ‘ravens are black’ attributes first of all a disposition to a natural kind, it may both explain why the disposition gives rise to a regularity and why it nevertheless allows for exceptions in which the disposition does not manifest itself, as it happens with albino ravens. But she also points out some difficulties for dispositional essentialism: once it is established that dispositional predicates cannot be reduced to conditionals, what is it that determines the objective essence of the disposition? She also shows the tension between Mumford’s dispositional essentialism and his thesis that the laws of nature are nevertheless contingent. Kistler’s paper (Chapter 4) is concerned with dispositional properties of macroscopic objects, such as fragility and solubility. He argues that such properties can be real in the sense of being causally efficacious without being identical to their microscopic base properties. This challenges a common presupposition of realists with respect to dispositions, such as Armstrong, for whom dispositional properties can contribute causally to their manifestations (which implies that they are real), and anti-realists, such as Jackson, Prior and Pargetter, according to whom dispositions are not themselves real (and thus not causally efficacious) because they are functional or second-order properties, whose first-order realizers are playing the causal role attributed by a dispositional predicate. Against the common idea that only microscopic base properties can be causally efficacious, Kistler argues that the criterion of causal efficacy is the existence of laws by virtue of which the property brings about other properties. Important laws of this type exist, e.g. for such cognitive dispositional properties as the representation of colour differences, which is why they can causally influence a subject’s behaviour. On the way to this conclusion, Kistler refutes a number of traditional and influential arguments against the efficacy of dispositions and shows that the issue of the micro-reduction of a macroscopic

Introduction

33

disposition is independent of the question of its efficacy. Kistler’s main thesis can be seen as complementary to Mumford’s: Whereas Mumford argues for the existence of ungrounded dispositions, Kistler argues that macroscopic dispositions that may turn out to be reducible can be real and efficacious without being identical to their reduction basis. The aim of Michon’s paper (Chapter 5) is to defend causal powers against the traditional objection of vacuity. Undeniably, it is unilluminating to explain why smoking opium makes the smoker sleepy by making reference to opium’s dormitive virtue. However, this does not show that the reference to opium’s dormitive virtue is never explanatory. In some contexts it is, as when one asks why smoking opium before driving a car enhances one’s probability to have an accident. Michon analyses the network of ontological, epistemological and logical presuppositions underlying the covering law account of causation, with its event ontology, and shows how these presuppositions make explanations in terms of causal powers inconceivable. He rehearses some arguments against the reduction of causality to lawful regularity: some causal relations appear to be unique and not to recur; not all laws are causal, whence the causal status of instances of the law cannot derive from their lawful status; the fact some event falls under a regularity does not by itself explain why it happens although it allows us to predict that it will happen. A genuine explanation requires reference to a mechanism underlying and responsible for the regularity. Michon develops a notion of natural action in which natural substances deploy their powers: such actions are the building blocks of the mechanisms whose discovery allows us to construct scientific explanations. Without powers, causal relations would appear capricious and remain incomprehensible. Michon joins Kistler in denying that powers are identical to the properties of what he calls (following Kenny) their ‘vehicles’ and Mumford in thinking that nothing stands in the way of the possibility of ungrounded powers. Michon introduces an important distinction between active tendencies and passive potentialities, the former having their ground within the (natural) agent, the latter being relative to powers external to it. Potentialities alone are incapable of making anything happen; however, tendencies or active powers can. This distinction provides the means for refuting an important objection against the dispositionalist ontology (put forward by Armstrong113): confusing tendencies with potentialities, the objection says that if all basic properties were dispositional, nothing would ever happen actually because all changes would remain forever potential. A natural agent acts necessarily in accordance with its active powers. Gnassounou’s paper (Chapter 6) is concerned with what he calls ‘conditional possibilities’, for which the assertion or denial of a modality is dependent on a specific condition, such as: ‘if the water is heated, it can boil’ and ‘if the water is not heated, it cannot boil’. He evaluates the pros and cons of an argument put forward by Lehrer and polished by Anscombe, relying on the use of such conditional possibilities, against any attempt to reduce the ascription of a possibility to a 113 D. Armstrong, ‘The Causal Theory of Properties: Properties according to Shoemaker, Ellis and others’.

34

Dispositions and Causal Powers

conditional in which no modal word occurs. Taking up the distinction established by Michon between two kinds of modalities, the general power of an agent (a disposition) and the circumstantial possibility of the power’s exercise, he confirms that the conditional analysis of the disposition is not vindicated nor invalidated by Lehrer’s argument, whereas the circumstantial possibility turns indeed out not to be reducible to a conditional. Jansen (Chapter 7) argues that, although dispositions can be attributed to entities of all four fundamental Aristotelian ontological categories, their attribution to individual substances is most fundamental, on three counts: metaphysically, logically and epistemically. He establishes metaphysical priority by showing that attributions of dispositions to universal kinds of substances, to property tokens (tropes) and to universal property types cannot be true without the corresponding attribution of the disposition to an individual substance being true. He establishes logical priority by analysing the logical form of disposition attributions to entities of the former three categories, showing that these forms all contain attributions to individual substances as an element. Finally, as to epistemic priority, he shows that we can know a lot about the dispositions of individuals without knowing anything about the dispositions of the kinds they belong to or of their properties, but that we can only obtain knowledge about the latter by gaining knowledge about the former. Thus, while agreeing with Gnassounou that individual substances are the fundamental bearers of dispositions, Jansen disagrees with him in arguing that they can also meaningfully be attributed to entities of the other three categories. Jansen shows that the attribution of dispositions (at least of what he calls ‘sure-fire’ dispositions and contrasts with probabilistic dispositions) presupposes a ‘principle of manifestation’ and a ‘principle of enabling’. According to the former, the presence of a disposition to M-in-circumstances-C, in a situation of type C, leads to manifestation M. According to the latter, if an object displays M in circumstances C, it must have had the disposition to M-incircumstances-C. The basic logical form of a disposition attribution is ‘x has the disposition to M-in-circumstances-C’, or ‘(dyn (M.C)) (x)’. But Jansen argues that we usually ascribe whole families of such dispositions, attributing to an object specific manifestations for a whole range of circumstances. This is compatible with all these attributions having a single property as their truthmaker. Dispositions in Contemporary Philosophy of Science Metaphysics is not the only philosophical discipline in which the interest for dispositions has grown. In fact we have already seen that philosophy of science has discovered its own new reasons for thinking that the notions of disposition and causal power are essential for scientific discourse. This is the main theme of the papers in the second part of this volume. One important topic in philosophy of science that has profited from the rediscovery of dispositions is the analysis of the role of idealization, itself indispensable for the discovery of laws of nature. As we have said before, the modern conception of

Introduction

35

scientific explanation tries to avoid postulating causes in the sense of substantial powers, mainly because such powers are thought to be inaccessible to scientific experiments. Modern science has habitually been thought to have substituted the search for regularities among phenomena to the search for causes. Scientific explanation is thought to consist in the integration of the phenomena to be explained within a regularity. However, those regularities that are lawful are in general not directly accessible to observation. To take a simple example, according to Galileo’s law of free fall, all bodies undergo the same acceleration if they are in free fall near the surface of the earth. However, the observation of real bodies falling in real circumstances never shows such a uniform acceleration. Common sense is surprised to learn this law, because what observation rather teaches us is that feathers fall more slowly than stones. In fact, the law of free fall only applies to bodies that are not subject to any other force than the gravitational attraction exercised by the earth, which is never the case in real circumstances. One can therefore only abstract that law from observation by abstracting away all other factors influencing the speed of a real body’s fall, such as the force exercised on the body by friction with the air. The law can only be obtained by ‘idealization’, by conceiving a phenomenon taking place in hypothetical circumstances in which it is determined by a single factor (or a small number of factors), instead of being determined by the large number of factors present in every real situation. Without constructing such simplified hypothetical situations, the law could not possibly have been discovered. (This point is taken up in particular by Hüttemann’s paper (Chapter 10) in this volume.) This simplification is fully justified by the possibility of getting back to real and complex situations in a second step: the properties of real falls are explained by conceiving them as the result of the superposition of the effects of all factors present in the situation, each obeying to its own laws. Taken in isolation, each of these laws does not correspond to any real regularity but to hypothetical regularities existing in hypothetical ideal circumstances. At this point, several authors have suggested that laws of nature do not directly bear on concrete objects, i.e. objects having many properties subject to any influences, but rather bear on their dispositions.114 One can fruitfully use the law of free fall to predict the properties of the falls of real objects although none of these falls directly obey that law, by making the hypothesis that real bodies nevertheless have the disposition to fall freely in accordance with that law. The concept of disposition also provides a justification for our usage of the concept of causation in Hume’s sense of an instance of a regularity. In a famous paper, Russell115 has observed that everyday events that we explain causally are in

114 See N. Cartwright, How the Laws of Physics Lie (Oxford, 1983). N. Cartwright, Nature’s Capacities and their Measurement (Oxford, 1989). N. Cartwright, The Dappled World, A Study of the Boundaries of Science (Cambridge, 1999). P. Lipton, ‘All Else Being Equal’, Philosophy, 74 (1999): 155-168. M. Kistler, ‘Laws of Nature, Exceptions and Tropes’, Philosophia scientiae, 7 (2003): 189-219. 115 B. Russell, ‘On the Notion of a Cause’, repr. in B. Russell, Mysticism and Logic (London, 1986).

36

Dispositions and Causal Powers

general so complex that they cannot give rise to any regularity. If, following Hume, one assumes that one event causes another event because regularly, events of the first type are followed by (contiguous) events of the second type, one must face the fact that everyday events never get repeated: no two car accidents exactly resemble each other in every respect. As Russell explains, the notion of exact resemblance only applies to the measurable quantities described by the quantitative laws of physics. However, he points out, it is conceptually impossible to consider the relata of these laws as causes and effects: physical laws relate simple properties not complex events. Furthermore, they do not bear on events taking place at different places and times, whereas causes always have spatiotemporal locations different from their effects. The differential equations of physics116 rather express functional dependencies of different properties an object possesses at a given moment. From this, Russell concludes that everyday causal judgements cannot be justified by laws of nature, which is the central hypothesis of the nomological account of causation mentioned above. The form of the laws of contemporary physics makes it impossible to use Hume’s concept of causation in terms of (lawful) regularity to characterize the relations between everyday macroscopic events, such as car accidents. However, once the properties of a complex object are conceived of in terms of the causal powers they give their bearers, it becomes possible to understand the evolution of the object as resulting from the superposition of the effects of many causal powers exercised by each of its properties in the circumstances. In the first part of the volume, several authors (in particular Mumford, Kistler, Michon and Gnassounou), show how realism about powers and dispositions provides the starting point for a non-Humean conception of causation. In his paper, Harré (Chapter 8) develops his thesis that the aim of science is to provide explanations of observed regularities by postulating powers that are not directly observable. He argues that powers allow us to account for the existence of exceptions to laws of nature, whereas no sense can be made of the notion of an exception within the Humean analysis of laws as universal regularities. (Schrenk, later in the volume, (Chapter 11) challenges the thesis, which Harré shares with Tiercelin, that powers allow us to make progress with the problem of exceptions.) According to Harré, natural kinds are just clusters of dispositions and powers. But he shows that the attribution of powers to particulars cannot be analysed without making reference to the ‘intrinsic nature’ of the particular, without making use of the ‘ceteris paribus clause’, and without making use of a conditional expression that can be satisfactorily analysed neither in standard first-order logic nor modal logic. (This was the main conclusion of Schmitz’ paper.) Taking up Michon’s and Gnassounou’s distinction between active powers and passive ‘liabilities’, Harré argues that there is a whole family of dispositional concepts, resembling each other on various counts, 116 Laws that have the form of differential equations express the dependency of the state of a physical system on the state preceding it by an infinitesimal time interval. The concept of such laws has been made accessible by the invention of the differential and integral calculus by Newton and Leibniz, which allows their expression in the form of differential equations.

Introduction

37

but not being susceptible of any common analysis. In line with the analysis put forward earlier in this introduction, Harré distinguishes between two traditions of physical explanation. A conservative tradition postulates, as the unobservable ground of observable phenomena, other and just smaller entities of the same kind. Such ‘iconic model building’ restricts its explanatory resources to concepts taken from the domain of observable phenomena. For the other more ‘radical’ tradition, represented by Gilbert, Boscovich, Kant, Faraday and Bohr, the scientific explanation of observable phenomena must proceed by postulating powers, in particular fields, which are entities of a radically heterogeneous type with respect to observables. Harré joins Mumford and Michon in pointing out that the postulate of irreducible (or ‘original’, as Harré calls them) powers is the price to pay for genuine scientific explanations. Cartwright (Chapter 9) agrees with Harré that there is a whole family of disposition concepts admitting of no common analysis. The main contribution of her paper is a careful distinction between two concepts of this family: dispositions and powers on one hand and capacities on the other. Here are some distinguishing features on her list: Powers or dispositions are not always displayed, whereas some capacities are. Powers but not capacities are basically ascribed to substances (cf. Jansen, Michon, and Gnassounou). Dispositions are characterized by a limited number of conditionals, linking them to characteristic manifestation in certain test conditions. By contrast, the list of manifestations of a capacity is open-ended; hence there is no set of conditionals defining the essence of a capacity. What Cartwright argues is the most important criterion for distinguishing dispositions from capacities is what she calls ‘malleability’: Dispositions can be interfered with (think of what we call ‘antidotes’, which rob dispositions of their typical manifestations), triggered and enhanced or diminished. Concerning triggering, enhancing and diminishing, she distinguishes fundamental physical capacities, such as the gravitational capacity, that need no triggering and cannot be enhanced or diminished, and derivative capacities, e.g. of economic systems, that can be triggered, as well as enhanced and diminished. This makes economic capacities more resemble dispositions than fundamental physical capacities do. Behind the possibility to be interfered with stands the most common feature of dispositions and capacities, their ‘two-sidedness’: Both dispositions and capacities are conceived in relation and opposition to their manifestations. This is what groups them together into one family distinct from the family of categorical properties. Finally, Cartwright joins Harré in concluding that both capacities and dispositions are an essential part of the conceptual apparatus of science. Hüttemann (Chapter 10) starts from observing that many physical laws describe the behaviour of objects in circumstances that may never be actually realized. How we can come to know such laws, and how laws bearing on ideal circumstances can nevertheless contribute to scientific explanations of what happens in non-ideal situations is a much discussed problem in philosophy of science. One traditional worry is that such laws can only be expressed by statements with ceteris paribus clauses, and that the essential indeterminacy of such clauses makes it impossible to confirm or refute ceteris paribus-law statements, which means that they do not have

38

Dispositions and Causal Powers

any definite meaning at all. Hüttemann argues that this problem can be solved by supposing that laws bear on dispositions, which are something that carries over from ideal to real (non-ideal) circumstances. However, as he points out, the epistemic transition between the ideal and the non-ideal presupposes that the disposition exhibits some kind of continuity between these kinds of circumstances. More precisely, only ‘continuously manifestable dispositions’ (or ‘CMDs’) can play a role in scientific explanations. It is only to the extent that a disposition is ‘continuously manifestable’ that we can infer a system’s real behaviour from its behaviour in ideal circumstances; conversely, only the supposition that a disposition is a CMD gives us epistemic access to the behaviour of systems in ideal circumstances, on the basis of the observation of their manifest evolution in non-ideal situations. Hüttemann joins the other authors of this volume in holding that the only way to make sense of modern science since Galileo is the hypothesis that dispositions are the fundamental object of scientific inquiry. In the second part of his paper, Hüttemann develops a new account of the relevance of dispositions for the analysis of causation. He agrees with other authors of this volume (in particular, Michon and Gnassounou) that dispositions are not to be construed as causes of their manifestations. He rather suggests that laws establishing connections between dispositions can serve as truthmakers for counterfactual conditionals, which can then be used, following David Lewis, to analyse the causal relation between events. Schrenk’s paper (Chapter 11) challenges the thesis shared by several authors of the volume (in particular, Tiercelin, Michon, Harré and Hüttemann) that the hypothesis that laws bear on dispositions rather than manifest or ‘categorical’ properties allows us to make sense of exceptions, and to give a definite meaning to law statements containing ceteris paribus clauses. He first reviews some of the wellknown difficulties for ceteris paribus law statements: they seem to be empirically vacuous or tautologous, they cannot be confirmed or refuted and they cannot support counterfactual conditionals. Schrenk then analyses carefully Cartwright’s version of the thesis that laws determine capacities. He discovers an important ambiguity in her writings: taking the example of the Coulomb force law, sometimes Cartwright considers the electrostatic force as a disposition, which is always present between electrically charged bodies at a given distance; but sometimes she treats the force as the manifestation of a disposition, which means that the force is not always present: whenever it is interfered with, it is not. More importantly, Schrenk argues that, even when the former reading is chosen, the thesis that the law attributes a capacity to electrically charged bodies does nothing to solve the problems raised by ceteris paribus clauses. His reason for this claim is that the relation between dispositions and their manifestations is plagued by exactly the same problems as the relation between laws bearing on the behaviour of systems in ideal circumstances and their consequences on real situations. As extensively shown in this introduction and other contributions to this volume, interferences and ‘antidotes’ make it impossible to state explicitly and generally how a system with a given disposition will behave in arbitrary circumstances. Hence, the thesis that laws ascribe dispositions does nothing

Introduction

39

to help us understand why we are able to test them in non-ideal circumstances and why we can use them in counterfactual reasoning. Dorato and Barberousse’s contributions (Chapters 12 and 13) make the case for the fundamental and irreducible status of dispositions in science stronger by looking closely at two important types of dispositional properties, the superposed states of quantum systems (Dorato) and specific heats (Barberousse). Dorato (Chapter 12) shows that, on the most important interpretations of quantum mechanics, the properties of quantum systems corresponding to observables for which the system is not in an ‘eigenstate’ (in other words, with respect to which it occupies a superposed state) are dispositional. Furthermore, on almost all interpretations (the exception being Bohm’s interpretation in terms of ‘hidden variables’) such dispositional properties of quantum systems before their measurement are irreducible, i.e. cannot be reduced to categorical properties of the system. This conclusion is mainly justified by the fact that the properties that can be attributed to quantum systems which are in superposed states with respect to a given observable, are contextual in a strong sense of the word: it is only through the interaction with a measuring device that they give rise to a categorical, measured value of the observable, where the obtained value depends also on the measuring device. Hence, before measurement, definite values of the observable can be ascribed to the system only as the set of dispositions to give rise to those definite values as a result of certain measuring interactions. In Bohr’s (so called ‘Copenhagen’) interpretation, these dispositions are considered as fundamental, i.e. not reducible to any intrinsic categorical properties of the system. Dorato shows that the same is true of almost all other important interpretations of quantum mechanics: Rovelli’s ‘non-collapse’ view, many-worlds and many-minds interpretations, and ‘spontaneous collapse’ theories, as proposed by Ghirardi, Rimini and Weber. Only David Bohm’s interpretation leads to a different result in this respect: in Bohm’s ‘hidden variable’ interpretation, a quantum particle’s position is considered to a have a definite (though ‘hidden’) value independent of measurements. However, position is the only observable whose value for a given system is not strongly contextual and can therefore be considered as categorical. All other variables, such as that of the spin of a particle in a given direction, are strongly contextual and therefore dispositional in the same sense as in the Copenhagen interpretation, the difference being that these dispositions are considered to be reducible to position and the context of measurement. If Dorato’s analysis of the properties of quantum mechanical systems shows that science supports Mumford’s thesis that there are ungrounded irreducible dispositions, Barberousse’s analysis (Chapter 13) of specific heats shows, in agreement with Kistler’s thesis about macroscopic dispositions, that science gives credit to the idea that macroscopic dispositional properties are real and causally powerful. Specific heats are measurable properties of macroscopic systems whose identity conditions are determined by the theoretical context of macroscopic thermodynamics, linking specific heats with other thermodynamic properties, such as entropy, heat, or (a gas’s) compressibility. Historically, the reductive explanation of the specific heats of gases had to wait for the development of the quantum mechanical description of the

40

Dispositions and Causal Powers

gas molecules. In the last part of her paper, Barberousse develops a subtle analysis of the relation of the heat capacity of a gas to its microscopic basis, i.e. the categorical properties of its constituent molecules. She shows that some influential arguments (due to D.H. Mellor) against the identification of the macroscopic heat capacity of a gas with microscopic properties of its constituents are not conclusive. Her own thesis is that the reduction of the specific heat of a gas to the microscopic properties of its constituent molecules rests upon irreducibly statistical properties, expressed in terms of infinite ‘ensembles’. It follows that such a thermodynamical dispositional property as the specific heat of a gas cannot be identified with any categorical (or ‘occurrent’) property of its microscopic constituents.

PART 1 The Metaphysics of Dispositions and Causal Powers

This page intentionally left blank

Chapter 1

Dispositions and Counterfactuals. From Carnap to Goodman’s Children and Grandchildren François Schmitz

In good Fregean orthodoxy, a concept or a relation is scientifically acceptable only if it has sharp boundaries, in other words, if it can be decided, for any object name filling the empty place of the conceptual word, whether the resulting proposition is true or false. It is well known that this requirement of sharp boundaries is the consequence of a twofold thesis: first, a concept does not subsist outside a propositional context, and second, a ‘scientific’ proposition is true or false, any third possibility being precluded. In good positivist-empiricist orthodoxy, a descriptive concept is scientifically acceptable only if it can be reduced to a ‘phenomenal basis’ or to ‘observational predicates’, for it must ultimately be possible to establish the truth or falsity of every proposition on the basis of the data of ‘experience’ (sense-data or observable characteristics). In good Russellian orthodoxy, ‘inferred’ entities are definitely suspect and it is desirable to be able to replace them by ‘logical fictions’, on the model of the construction of mathematical entities on the basis of a small number of ‘logical’ entities provided by Principia Mathematica, which means to construct them as classes. Carnap was one of the few who admired Frege; when he read Our Knowledge of the External World in 1921 he was enthusiastic, and he did not feel any antipathy for Mach’s ‘phenomenalist’ theses. The result was The Logical Construction of the World,1 written in the years around 1925 and published in 1928. For what will concern us here, Frege’s requirement of ‘sharp boundaries’ entails that the construction/definition of objects or quasi-objects makes use only of explicit definitions: definitions in the strict sense of the term (‘4=df 3+1’, so that 4 belongs to the same ‘object sphere’ as 3 and 1), or definitions in use (‘n is a whole number

1 R. Carnap (1928), Der Logische Aufbau der Welt (Hamburg, 1998), trans. Rolf A. George, The Logical Structure of The World and Pseudoproblems in Philosophy (Chicago, 2003). I shall refer to this work as Aufbau.

44

Dispositions and Causal Powers

that is the square of a whole number =df n is a whole number and there exists m such that m is a whole number and n = m × m’, ‘whole number that is the square of a whole number’ designates a quasi-object relatively to integers). Such definitions being considered as bi-conditionals between definiens and definiendum, it is always possible to eliminate those ‘constructed’ objects and to put in place of them the objects on the ground of which they have been constructed. So, every statement including expressions for ‘constructed’ concepts can be translated into a statement not including them. On the other hand, so called ‘implicit’ definitions (introducing concepts through axiomatic systems) are definitively ruled out because it is not possible to decide, for each object, whether, or not, it falls under so ‘defined’ concepts, as Frege had so many times emphasized in his articles on Hilbert’s Foundations of Geometry and as Carnap himself had settled anew in his 1927 paper, ‘Eigentliche und Uneigentliche Begriffe’.2 Applied to the descriptive concepts which appear in empirical sciences, Frege’s requirement has effects which seem puzzling: a scientific sentence in which appear concepts constructed in the constitutional system of Carnap sketched in the Aufbau, is, after substitution of the definientia to the definienda, reduced to a statement (very complicated) which expresses only a state of affairs concerning the basic relation; and this state of affairs is made of a finite number of basic elements (so-called ‘elementary experiences’) connected by the relation of ‘recollection of similarity’.3 In this way, we are sure that it is always possible to decide whether or not the state of affairs in question is the case. More accurately, this means that a scientific statement speaks only of a finite number of past experiences even if it seems to state the future occurrence of an event or a general law. As E. Kaila said in his critical study of the Aufbau: ‘… as regards its logical meaning – but not also as regards its “content” – no concept which does not directly refer to what is given can denote anything else than classes of basic elements, classes of classes or classes of relations of them, etc. […] Just as little as pure mathematics, if natural numbers constitute its basis, can contain anything other than statements about classes, classes of classes, classes of relations,… of natural numbers, just so little can empirical science contain anything other than statements about classes, classes of classes,… of “my” past elementary experiences’.4 Carnap was not saying anything else in ‘Testability and Meaning’5 when he wrote: ‘The laws of physics as well as all predictions were interpreted [in the Aufbau] as

2 R. Carnap, ‘Eigentliche und Uneigentliche Begriffe’, Symposion: Philosophische Zeitschrift für Forschung und Aussprache, Bd. 1, H. 4, Berlin, 1927. 3 Cf. R. Carnap, Der logische Aufbau der Welt, § 180. 4 E. Kaila, ‘Der logistische Neupositivismus’ (1930), trans. in E. Kaila, Reality and Experience (Dordrecht, 1979), pp. 15-16. 5 R. Carnap, ‘Testability and Meaning’, §23, p. 179, in R.R. Ammerman, Classics of Analytic Philosophy (New York, 1965), pp. 130-195.

Dispositions and Counterfactuals

45

record of present and (remembered) past experiences, namely those experiences from which the law and the prediction is usually said to be inferred by induction.’ Yet, one year before the publication of the Aufbau, the physicist P.W. Bridgman, following the well known formula of Einstein’s (‘this notion [sc. the simultaneity of two events] exists, for the physicist, only if he has found the possibility to verify in the concrete case whether or not it is accurate’), exposed his operationist conception of physical concepts: ‘In general, we mean by any concept nothing more than a set of operations; the concept is synonymous with the corresponding set of operations’.6 The immediate consequence of this thesis is that to each set of operations corresponds a different concept, even if, ‘intuitively’, it would seem that we deal with the same concept (for example, the concept of ‘length’ explodes into so many different concepts as there are different procedures of measuring length): ‘If we have more than one set of operations, we have more than one concept, and strictly there should be a separate name to correspond to each different set of operations’.7 Another consequence of this point of view was that a physical concept has a determinate meaning only in the domain or region in which the operations can be performed; where it is not possible to perform them, the corresponding concept lacks any meaning: ‘… the concepts can be defined only in the range of actual experiment, and are undefined and meaningless in regions as yet untouched by experiment’.8 In Frege’s words, all this amounts to say that a physical concept has no sharp boundaries: it is possible to decide for an object whether, or not, it falls under the concept only if the object in question belongs to the domain where the corresponding operations can be performed; for objects outside this domain, the question whether or not it falls under the concept, is merely meaningless. In other words, Bridgman’s theory results in giving to the definition of a concept the form of a conditional definition, whose antecedent includes experimental data.9 The Vienna Circle took notice of Bridgman’s book in 1929 and quickly recognized in its ‘operationism’ a point of view in keeping with its own empiricism. After having admitted, at the beginning of the 1930s, that the translation requirement championed in the Aufbau was too strong, Carnap brought up the weaker requirement of reducibility (and confirmability) which was a sort of formalization of a weak form of Bridgman’s operationism.

6 P.W. Bridgman, The Logic of Modern Physics (New York, 1927), p. 5 (Bridgman’s italics). 7 Ibid., p. 10. 8 Ibid., p. 7. 9 As it is known, this kind of definition was severely criticized by Frege (cf. Grundgesetze II, § 65, ‘Über die Grundlagen der Geometrie’, Jahresbericht der Deutschen MathematikerVereinigung, vol. 15 (1906): 379-380, Letter to Peano, 29 Sept. 1896), but also by Russell (cf. ‘Les paradoxes de la Logique’, Revue de Métaphysique et de Morale, tome 14 (1906): 641). The so-called ‘universality of logic’ follows from this denial of conditional definitions.

46

Dispositions and Causal Powers

This ‘liberalisation of empiricism’, which will take new forms after the Second World War, recognized that it could be interesting to introduce new concepts on the basis of ascertained regularities, that is to say, if one is ‘regularist’, on the basis of physical laws: for example (a simple one!) if I remark that some bodies, every time they are subject to a pressure and change their shape, recover their initial shape when the pressure is released (while others do not), I could consider advisable to subsume them under the concept of ‘elastic body’. Afterwards, on the demand of the Royal Society, I investigate the behaviour of two elastic bodies when they collide, and, after some experiments, I establish the ‘law’ according to which, when ‘two bodies running with equal and opposite velocities collide, they part company with the same velocities as before’. From this, it follows that if an elastic body collides with another elastic one at rest, it transmits to it all its velocity and stays at rest after the collision, etc. Then, I come perhaps to the idea that an elastic body is such that every change of its shape corresponds to a change in the system of pressure to which it is subjected, and reciprocally, etc. A property like that of ‘elasticity’ appears only when I remark that in some circumstances, or when subjected to some tests, some objects (here, bodies) behave regularly in a determinate way. This experimentally ascertained regularity suggests introducing a new concept as far as it seems reasonable to expect the objects falling under this concept to present other common characteristics in other circumstances (here, to obey particular laws of shock); it is clearly a bet on the future which could prove wrong, only future experiments being able to decide on this. As one knows, the so-called disposition-concepts, those which in everyday language end in ‘able’ (‘adaptable’), ‘ible’ (‘visible’) or ‘uble’ (‘soluble’) furnishes the prototype of such concepts. The aim of Carnap in ‘Testability and Meaning’ was to legitimate the introduction of such concepts, once it is admitted that all ‘theoretical’ concepts can be handled as disposition concepts. What does ‘to legitimate’ mean here? This means: 1. that it must be possible to continue distinguishing concepts of this kind – which, while they are not eliminable to the benefit of observational concepts or terms alone, are not without support in ‘experience’ – from ‘metaphysical’ concepts (such as the ‘entelechy’ of Driesch, for example); 2. that the distinction between analytic statements and synthetic ones can be made sharply. The historical situation of this question is rather curious. Carnap had accepted getting out of the frame of strict Fregean orthodoxy and had endeavoured to preserve, as much as possible, the empiricist orthodoxy. Later on, he eventually recognized that the solution, in terms of ‘disposition’, of the epistemological problem posed by the ‘theoretical terms’ was not fully satisfactory; this led him, by steps, to re-elaborate the solution sketched by Ramsey in 1929 and to appeal to the so-called ‘Ramsey sentences’.

Dispositions and Counterfactuals

47

On the other hand, the introduction of disposition predicates by means of the ‘reduction sentences’ put forward by Carnap in ‘Testability and Meaning’, gave rise, in the 1940s/50s to an abundant literature resting on the idea that one had to get off the extensional logic for the treatment of the disposition predicates because definitions of such predicates ought to be in terms of ‘counterfactual (or subjunctive) conditionals’, something like: ‘b is soluble =df if b were put into water, b would dissolve’. As it is well known, later on, the developments of ‘possible worlds’ semantics for modal logic led some authors in the 1960s (Stalnaker and D. Lewis are the best known) to treat again the question of a ‘logic’ of counterfactuals in this new frame. The idea was that the seemingly insuperable difficulties, which have been encountered in this question in the 1940s, could perhaps be solved by using this semantic approach. So, a new topic got in and grew in an autonomous fashion, but only loosely related with Carnap’s initial concerns. Yet, as the logics of counterfactuals have their remote origin in the way Carnap was treating disposition predicates in ‘Testability and Meaning’, it may be amusing, or at least instructive, to come back and to consider the following question: is the handling of the counterfactuals from the point of view of possible worlds semantics, more precisely, from that of the ‘ptolemaic’ semantics of D. Lewis, able to cope with the objections raised by Chisholm and Goodman against the introduction of disposition predicates by means of reduction sentences? Carnap, in 1963, expressed his open-mindedness on this topic and declared in his answer to the criticisms of A. Pap: ‘Only in the future, when the logic of modalities has been investigated much more thoroughly, will it be possible to judge whether an extensional or a modal language has the greater overall advantage’.10 Perhaps it is the moment to pass judgement on this matter. Vices and Virtues of Reduction Sentences A disposition property is ascribed to an object only under the condition that the object in question responds in a determinate way to a change in the surrounding circumstances: a piece of wax, when heated, may be said to be flexible if it changes its shape when subjected to a pressure. Strictly speaking, the piece of wax cannot be said to be flexible before it has been verified that it changes its shape when subjected to a pressure. So long as it has not undergone such a test, the assertion that it has this property is not supported. With the aid of standard logic, one is led to introduce a predicate expressing a disposition property by a ‘definition in use’ having, more or less, the following form: x is flexible (at temperature T) =df if x is subjected to a pressure, then x changes its shape.

10 P.A. Schilpp (ed.), The Philosophy of R. Carnap (La Salle, 1963), p. 951.

48

Dispositions and Causal Powers

In the usual symbolism, we get: ∀x[Dx ↔(Px→Qx)] As it is well known, such a definition is not satisfactory because, in virtue of the properties of the ‘material’ implication, the definiens is true every time the antecedent of the implication is false (here: every time ‘x is subjected to a pressure’ is false); which has the counterintuitive consequence that every object which is not subjected to a pressure could be legitimately said to be flexible. In other words, whenever ‘Pb’ is false for a given b, the definiens is true and so ‘Db’ is equally true. To cope with this difficulty, Carnap, in ‘Testability and Meaning’, substitutes for this definition in use, what he calls a ‘reduction sentence’ allowing to introduce ‘D’ in the following way: ∀x{Px→[Qx↔Dx]} that is to say: ∀x{{[Px ∧ Qx]→ Dx} ∧ {[Px ∧~Qx] → ~ Dx}}11 Here, in this particular case, we have what Carnap calls a bilateral reduction sentence. It may be the case that ‘P’ and ‘Q’ are not the same predicate in the two members of the conjunction. We have then two distinct reduction sentences, which are mere material conditionals: ∀x{[P1x ∧ Q1x]→ Dx} ∀x{[P2x ∧ Q2x]→ ~Dx} If those two sentences are valid (analytic or P-valid12), we have a ‘reduction pair’. Carnap adds the obvious condition that the conjunctions in the premises ought not to be contravalid (contradictories or P-contravalid); in other words, it must not be the case for: ∼∃ x[P1x ∧ Q1x]

11 This formula is equivalent to: ∀x{[P(x) → {Q(x) → D(x)}] ∧ [P(x) → {~Q(x)→ ~D(x)}]} 12 In Carnap’s terminology at this time, a statement S is valid iff S can be deduced, in virtue of the inference rules, from the axioms laid down in the system. There are two kinds of axioms: logical axioms and empirical postulates; if S follows from the logical axioms alone, S is L-valid or analytic. If, in the deduction of S, at least one empirical postulate is used, S is P-valid. If S is not valid, it may be L-contravalid i.e. contradictory (there is a deduction for ~S from the logical axioms alone), or P-contravalid (there is a deduction for ~S using empirical postulates), or indeterminate (there is a deduction neither for S nor for ~S).

Dispositions and Counterfactuals

49

to be analytic or P-valid. In such a case, no object, either on logical or empirical grounds, would satisfy the conditions of application of the property D, which would become itself inapplicable. So, unlike the first attempt to introduce ‘D’ by means of a definition in use, when we are dealing with an object b which does not satisfy ‘P’, for example, the implication ‘[P1b ∧ Q1b] → Db’ is indeed true, but the falsehood of the antecedent does not allow to detach ‘Db’ and so, it is not possible to attribute the property D to b. Concerning a bilateral reduction sentence, this results in attributing the property D to each object satisfying both ‘P’ and ‘Q’ and the property ~D to those satisfying both ‘P’ and ‘~Q’; but to an object not satisfying ‘P’, we can attribute neither D nor ~D. In this last case, to say of b that it is D or that it is not D, is neither true nor false, but merely meaningless. As Carnap says: ‘If we establish one reduction pair (or one bilateral reduction sentence) as valid in order to introduce a predicate “D”, the meaning of “D” is not completely established, but only for the case in which the test condition is fulfilled.’13 Of course, introducing new reduction pairs or establishing new laws, allows reducing this ‘region of indeterminateness’, but without being sure that this region could be completely reduced. An explicit definition appears then as a special case of reduction sentence: if a bilateral reduction sentence as ∀x{Px → [Qx ↔ Dx]}’ is analytic, with ‘∀xPx’ analytic too, then ‘Q(x) ↔ D(x)’ is analytic and has the same status as a usual explicit definition of ‘D’ in term of ‘Q’. It is for this reason that Carnap considered that introducing a predicate by means of reduction sentences(s) is just an extension of introducing it by means of explicit definition. So, one can avoid the predicament arising from the introduction of ‘D’ by means of a definition in use, but at the price of not complying with the Fregean requirement of sharp boundaries since one cannot decide the truth or falsehood of the proposition ‘Db’ when ‘b’ names an object belonging to the ‘region of indeterminateness’. There was nevertheless a way out, which had been taken by Frege: when it seems that a concept have no meaning for an object (Is J. Cesar a prime number?), it is enough to stipulate that the proposition resulting from filling the argument place of the conceptual word by a name of this object, is false. Carnap refused this trick because ‘[it] is not in accordance with the intention of the scientist concerning the use of the predicate [‘D’]. The scientist wishes neither to determine all the cases of the third class positively, nor all of them negatively; he wishes to leave this question open until the results of further investigations suggest the statement of a new reduction pair’.14 Later on, after having given up this solution, Carnap will stress the fact that, even if ‘this simple form [sc. reduction sentences] cannot suffice to introduce concepts of theoretical science’, it was nevertheless ‘useful because it exhibited clearly the open character of the scientific concepts,

13 R. Carnap, ‘Testability and Meaning’, §9, p. 148. 14 R. Carnap, ‘Testability and Meaning’, §10, p. 151.

50

Dispositions and Causal Powers

i.e., the fact that their meanings are not completely fixed’.15 There was a second formal drawback in introducing a predicate by means of reduction sentences: in most cases, such sentences have factual consequences, i.e. by laying down as valid a reduction pair, for example, it is asserted that some state of affairs is realized. Indeed, let the reduction pair ∀x{[P1(x) ∧ Q1(x)] → D(x)} ∀x{[P2(x) ∧ Q2(x)] → ~D(x)} It is easy to see that by contraposing the second formula and using transitivity, one obtains the formula: ∀x~{P1(x) ∧ Q1(x) ∧ P2(x) ∧ Q2(x)} This formula, which Carnap calls ‘the representative sentence’ of the reduction pair, expresses merely the fact that the same object cannot have, at the same time, the four properties P1, P2, Q1, et Q2 since this would mean that it would be possible to attribute to it both D and ~D. The only circumstances in which the representative sentence has no factual content is when we are dealing with a bilateral reduction sentence as: ∀x{P(x) → [Q(x) ↔ D(x)]} whose representative sentence is: ∀x~{P(x) ∧ Q(x) ∧ ~Q(x)}, which is obviously ‘analytic’ (‘L-valid’). 16 Finally, it follows from all this that the introduction of a predicate by means of reduction sentences makes such a predicate ineliminable, which means a new and 15 P.A. Schilpp, The Philosophy of R. Carnap, p. 59, cf. also, ‘The Methodological Character of Theoretical Concepts’, in Minnesota Studies in the Philosophy of Science, vol. 1 (1956): 67. 16 Nevertheless, it has to be noticed with Hempel (Schilpp, The Philosophy of R. Carnap, p. 704) that this holds only if ‘D’ is introduced by one single bilateral; if two or more bilaterals are laid down for ‘D’, there are again factual consequences. For two bilaterals, ∀x(Px → (Qx ↔ Dx)) and ∀x(P’x → (Q’x ↔ Dx)) the representative sentence is: ∀x~(Px ∧ Qx ∧ P’x ∧ ~Q’x) ∧ ∀x~(P’x ∧ Q’x ∧ Px ∧~Qx) that is to say ∀x[(Px ∧ Qx ∧ P’x) → Q’x] ∧ ∀x[(P’x ∧ Q’x ∧ Px) → Qx)]

Dispositions and Counterfactuals

51

serious departure from the introduction of a term by means of explicit definition and has important consequences on the question of ‘analyticity’. After the Second World War, these two features of the introduction of predicates by means of reduction sentences – that is to say incomplete meaning and factual consequences – have been severely judged by Chisholm, Pap, and Goodman. They thought it was too high a price to cope with the difficulty, noticed by Carnap, raised by the treatment of disposition predicates in the frame of extensional logic. Chisholm17 was surprised that in a reduction sentence like ‘if body b is placed in water at time t then b is soluble iff b dissolves at time t’, the first and third sub-propositions are meaningful (supposing obviously that the predicates ‘x is placed in water at time y’ and ‘x dissolves at time y’ comply with the Fregean orthodoxy) while the second one is, or is not, meaningful, depending on the truth or falsehood of the first. In the second case, it is the whole of the statement which becomes meaningless, but to know that, we must wait until we know that one of its components (the first) is false, which seems very strange! Goodman, for his part, considered it impossible to put on the same footing ‘introduction by means of reduction sentences’ and ‘introduction by means of definition’: to introduce a predicate by means of reduction sentence is, indeed, to introduce it by postulate, as primitive, hence ineliminable.18 More generally, he blamed Carnap for having tried to reproduce in his formal machinery for disposition predicates, the order of their ‘presystematic adoption’, rather than having tried to ‘explicate’ them: ‘While explication must respect the presystematic application of terms, it need not reflect the manner or order of their presystematic adoption.’19 These criticisms are just worth what their presuppositions are worth and they are precisely the ones that Carnap allows himself to give up in the name of the principle of tolerance. The point for him is to show how in a ‘liberalized’ framework it is still possible to distinguish the analytic from the synthetic and allowable predicates from ‘metaphysical’ ones. The general idea is the following: according to the interpretation that people of the Circle had provided of Wittgenstein’s Tractatus, a proposition (or a statement) is meaningful only if it is either an elementary proposition, whose truth or falsehood is directly decidable on the basis of ‘experience’, or a ‘molecular’ proposition, whose truth or falsehood is a ‘function’ of the truth or falsehood of its component elementary propositions. In this perspective, it was admitted that an elementary proposition like ‘this table is black’ (as far as it is elementary) is 17 In his 1946s paper ‘The Contrary-to-Fact Conditional’, reproduced in H. Feigl and W. Sellars (eds), Readings in Philosophical Analysis (New York, 1949), p. 488. 18 Later on, Carnap will acknowledge this point and will consider reduction sentences (in a slightly different form) as ‘meaning postulates’, cf. ‘Meaning Postulates’, in Meaning and Necessity, (2nd edn, Chicago, 1956), p. 228; cf. also P.A. Schilpp, The Philosophy of R. Carnap, p. 948. 19 N. Goodman, Fact, Fiction and Forecast, (Indianapolis, 1973), p. 47.

52

Dispositions and Causal Powers

meaningful because it is determinately true or false, and that, in principle, it is possible to ascertain it definitively, in other words, to ‘verify it completely’. It was likewise admitted that a proposition like ‘men are bad’ ought to be analysed as a conjunction of elementary propositions: ‘Hitler is bad and Stalin is bad and…and Joe Dalton is bad and…etc.’ and that the quantifiers were not to be introduced as primitive logical constants. The distinction between analytic and synthetic statements is then theoretically harmless as is the distinction between allowable and metaphysical predicates, with the proviso, however, that the ‘theoretical’ predicates be replaceable by primitive ‘observable’ predicates, which means that they have been introduced by explicit definitions only (as Carnap did in his Aufbau). It is very difficult to follow these two theses to their ultimate consequences: an elementary statement like ‘this table is black’ is not ‘completely verifiable’ since the correctness of the observations which have led to assert it is always open to doubts. The best we can do is admit that this statement is firmly confirmed and that this is enough to affirm it. In the same way, a universal statement, with a few exceptions, cannot be transformed into a conjunction of singular statements, and then, cannot be conceived as a ‘truth-function’ of these last ones. Hence, such universal statements cannot be completely verified; at best, they are well confirmed as far as their instances are, for their own, well confirmed. All this ends in the view that ‘every (synthetic) sentence is a hypothesis, i.e. can never be verified completely and definitively’.20 How, then, is it possible to distinguish the allowable predicates from the nonallowable ones, and the synthetic sentences from the analytic ones? Concerning the first point, Carnap’s solution rests on the idea that confirmation is conveyed by the relation of ‘consequence’ in both directions, i.e. down to the consequences and back to the premises. It is a question of generalizing the two following situations: 1. from an elementary sentence ‘Pb’ follows logically the conclusion ‘∃xPx’ and in this case ‘∃xPx’ is said to be completely confirmed by the premise ‘Pb’: by way of generalization, the confirmation of a sentence E is said to be completely and directly reducible to the confirmation of a finite set of sentences C if and only if E is a consequence of C. 2. from a universal sentence ‘∀xPx’ follows an infinite set C’ of sentences of the form ‘Pa’, ‘Pb’, etc. (with ‘a’, ‘b’, etc. constants); the more of these last sentences are confirmed, the more ‘∀xPx’ can be said confirmed, in so far as no sentence of the form ‘~Pc’ has been confirmed; in this case ‘∀xPx’ is said to be incompletely confirmed. By way of generalization, the confirmation of a sentence E is said to be incompletely and directly reducible to the confirmation of an infinite set of sentences C’ if and only if the sentences belonging to C’are consequences of E by substitution alone.

20 Carnap, ‘Testability and Meaning’, §4, p. 136.

Dispositions and Counterfactuals

53

These two definitions make it possible to define the corresponding relations of indirect reducibility, complete or incomplete, along the same lines as the passage from the fatherhood relation to that of ancestor.21 Now, these definitions extend themselves to the reduction of a predicate to a class of predicates. Following Carnap lets us call ‘full sentence’ of the one place predicate (say) ‘P’ a sentence of the form ‘Pb’, and ‘CP’ the set of the full sentences of ‘P’ and of their negations. If we are dealing with a set of predicates D, we will set up in the same way the set CD of the full sentences of the predicates belonging to D and of their negations. Now, one can define the reduction (complete or incomplete) of the confirmation of a predicate P to that of the set of predicates D in the following way: The confirmation of ‘P’ is reducible (comp. or incomp.) to that of D iff the confirmation of each full sentence of ‘P’ with some constant c is reducible to that of a consistent subset of ‘CP containing only full sentences or negations of full sentences of predicates belonging to D with the same constant c. If this holds also for ‘~P’ (i.e. if the confirmation of ‘~P’ is reducible to that of D), it will be merely said that ‘P’ is reducible to D. Let us suppose now that a predicate ‘R’ has been introduced by the reduction pair: ∀x{[P1x ∧ Q1x] → Rx} ∀x{[P2x ∧ Q2x] → ~Rx} and let D = {P1, P2, Q1, Q2}. It is plain then that ‘R’ is completely reducible to D. Indeed, the confirmation of each full sentence of ‘R’, i.e. of each sentence of the form ‘Rb’, with b constant, as well as of each sentence ‘~Rb’, is completely reducible to the confirmation of a consistent subset of {P1b, ~P1b, P2b, ~P2b, Q1b, ~Q1b, Q2b, ~Q2b}, i.e. ‘Rb’ is consequence of {P1b, Q1b} and ‘~Rb’ is consequence of {P2b, Q2b}. In fact, we have for each constant ‘b’: P1b, Q1b, ∀x[P1x → (Q1x → Rx)] ⏐= Rb and, the same way: 21 Easy example of ‘completely and indirectly reducible’: consider the sentence ∃x∃yPxy and the class C of sentences: {Pab}; we get: the confirmation of ∃x∃yPxy is completely reducible to the confirmation of the class C’ = {∃yPay} and the confirmation of ∃yPay is completely reducible to the confirmation of the class C = {Pab}; hence the confirmation of ∃x∃yPxy is completely and indirectly reducible to the confirmation of C = {Pab}. Easy example of ‘incompletely and indirectly reducible’: let the sentence ∃x∀yPxy and the infinite class C of sentences {Paa, Pab, Pac, Pad, etc…}; we get: the confirmation of ∃x∀yPxy is completely reducible to the confirmation of the class C’={∀yPay} and the confirmation of ∀yPay is incompletely reducible to the confirmation of the class C= {Paa, Pab, Pac, Pad, etc…}, hence: the confirmation of ∃x∀yPxy is incompletely and indirectly reducible to the confirmation of the class C.

54

Dispositions and Causal Powers

P2b, Q2b, ∀x[P2x → (Q2x → ~Rx)] ⏐= ~Rb22 Now, if the predicates belonging to D are observable,23 then, in virtue of the definition of the confirmability of a predicate,24 it can be said that ‘R’ is completely confirmable. In case of ‘soluble’, for example, it can be even added that it is a testable predicate (but no more on this here). One sees that it is the mere logical relation of consequence which gives the predicate ‘R’ its confirmability (in case ‘P1’, ‘P2’, ‘Q1’, and ‘Q2’ are confirmable predicates or reducible to observable predicates). Whatever is the factual truth of ‘Pb’, etc., it remains that the two micro-inferences above are valid and permit us to say, in virtue of the definitions of reducibility and confirmability, that ‘R’ is confirmable. That is why the confirmability of ‘R’ is formally ascertained, even if, at worst, no object is P (which is excluded by Carnap in §8). So the confirmability of ‘R’ is independent of the fact that its meaning is not ‘completely determinate’ since the fact that neither R or ~R can be attributed to what is not P, does not matter for the confirmability of ‘R’. That is why, in spite of this indeterminateness, ‘R’ is allowable for an empiricist and has a meaning, unlike other predicates which prove to be non-reducible (comp. or incomp.) to a class of observable predicates. The comparison with what Carnap was doing in his Aufbau, results in what follows: In the Aufbau, the only use of explicit definitions, which made it possible to eliminate the introduced (‘constructed’) predicates, made sure that these predicates were ‘completely verifiable’, once it was admitted that it was always possible to ascertain that some state of affairs was realized. There were no needs to elaborate a particular notion of ‘confirmability’ or of ‘testability’: the mere eliminability of the introduced predicates, warranted by the procedure of ‘construction’, was sufficient. The ‘liberalization of empiricism’ led to dissociate the introduction procedure of a predicate from its confirmability/testability: from this time onward, the problem for Carnap was to try to formalize an introduction procedure which could reflect ‘the manner or order of their presystematic adoption’, as said Goodman, while working out at the same time a notion of factual content (in ‘Testability and Meaning’ this was done via the notions of confirmability and testability) which could provide a clear cut separation between allowable and unallowable predicates. Afterwards, Carnap will not blame himself for having violated the Fregean orthodoxy, but, on the contrary, for having been too cautious in his way of formalizing the introduction

22 That is what is expressed by theorem 6, p. 147 of Carnap, ‘Testability and Meaning’. 23 ‘Observable’ is a primitive predicate that Carnap only explicates, cf. Carnap, ‘Testability and Meaning’, §11. 24 A predicate is confirmable (comp. or incomp.) if it is reducible (comp. or incomp.), in the sense above, to a class of observable predicates, cf. Carnap, ‘Testability and Meaning’, df. 18, p. 157.

Dispositions and Counterfactuals

55

of predicates in sciences; hence, to have ruled out scientific practices which are epistemologically fruitful, even if not logically respectable. Now, what about the second point mentioned above, that is to say: what about the distinction between analytic and synthetic sentences; more precisely, for all that concerns us here, are sentences in which there occur terms introduced by means of reduction sentences, analytic or synthetic? Generally, when we are dealing with sentences containing descriptive predicates, primitive or introduced by means of explicit definitions, the criteria of analyticity, as expressed by Carnap in The Logical Syntax of Language, §51 and in ‘Testability and Meaning’ §10, are the following: -

-

A sentence S containing only primitive descriptive predicates is analytic iff it is valid and all the sentences obtained from S by substitution for descriptive predicates, wherever they occur in S, of other ones of the same type, are valid. A sentence S, containing descriptive predicates introduced by explicit definitions is analytic iff the sentence obtained from S by elimination of the defined predicates is analytic.

Problems clearly arise when a sentence contains one or more predicates introduced by means of reduction sentences, since these predicates, as we saw, are not eliminable. Carnap’s solution to this problem is in conformity with the following pattern: it can be admitted generally that all that follows from an explicit definition is analytic (for example if ‘Qx =df Px∧Rx’ is laid down, then ‘Qa→Ra’ is obviously analytic). When predicates introduced by means of reduction sentences are taken into account, then such a characterization is no more possible since, as we saw, from a reduction pair (or a set of reduction pairs) follows logically a sentence with factual content, the so-called ‘representative sentence’. So, it is not enough to accept as analytic, sentences that follow logically from reduction pairs. It must be added that only those sentences are analytic, which follow logically from a set of reduction pairs without factual consequences. This means that the only sentences taken as analytic are those following from one bilateral, since from one bilateral follows only a sentence without factual content. Nevertheless, there remains an apparent difficulty which has been raised by some of Carnap’s readers: what about sentences which look like logically valid sentences but which contain predicates introduced by means of reduction sentences? To take an easy example, what about the sentence ‘∀x(Rx∨∼Rx)’ when ‘R’ has been introduced by a reduction pair? The predicament is that the meaning of ‘R’ is not completely determinate and that if one takes into account an object b belonging to the ‘region of

56

Dispositions and Causal Powers

indeterminateness’ of ‘R’, it seems senseless to assert ‘Rb∨∼Rb’ since ‘Rb’ (as well as ‘~Rb’) is supposed to be meaningless.25 Carnap’s way of speaking in ‘Testability and Meaning’ led obviously to ask this question;26 in emphasizing that for an object b not belonging to the class (P1 ∧ Q1) ∨ (P2 ∧ Q2), ‘neither the predicate nor its negation can be attributed to b ’ and then, that in this case, ‘R’ ‘is without meaning’,27 he was inciting his readers to think that the Excluded Middle Principle could not hold for this kind of predicate. Nevertheless, this was missing the point of ‘Testability and Meaning’, which was precisely to provide a notion of meaning for non-Fregean predicates, those which have no ‘sharp boundaries’. As we have seen above, that a predicate introduced by means of reduction sentences has no sharp boundaries, is quite compatible with the fact that it is confirmable (or even, completely confirmable), and, hence, on the basis of adequate decisions, that it can be admitted as empirically meaningful. As a consequence, the Excluded Middle, or Excluded Contradiction Principle holds for such a predicate. As Carnap emphasized in his answer to A. Pap, ‘if a disposition predicate is introduced as a primitive constant […], then, although only partially interpreted, it is significant in a wider sense since possibilities of confirmation and of disconfirmation are given. This significance is not depending upon contingent facts but only upon the rules of the given language, including the postulates […]. Therefore the logical principles, e.g. that of the excluded middle and that of the excluded contradiction, hold unrestrictedly for terms introduced by reduction sentences.’28 However ingenious may be Carnap’s position in ‘Testability and Meaning’, to introduce predicate by means of reduction sentences is obviously a rather odd procedure. Can it be claimed that such a procedure is no more than an extension of the usual procedure of introduction by means of explicit definition, even though it amounts not only to introduce a new predicate, but also to ‘postulate’ that such and such state of affairs is the case? The problem of the ‘factual consequences’ of reduction sentences is obviously crucial and it has led more than one to remark that there is no good reason to try to distinguish reduction sentences (or, later on, ‘meaning postulates’) and ordinary empirical postulates. We do not want to pursue this discussion here, but it can be remarked that a procedure avoiding reduction sentences to the benefit of an honest introduction by means of explicit definitions would have been welcomed. It was perhaps Carnap’s hope as he answered to A. Pap: ‘Once the problem of the explication of nomic form has been solved and a logic of causal modalities has been constructed, it will be possible to use these modalities 25 Let us recall that the Fregean requirement of ‘sharp boundaries’ is nothing more than the requirement that the Excluded Middle will hold without limitations. 26 cf. A. Pap, ‘Disposition concepts and extensional logic’ in Minnesota Studies in the Philosophy of Science, vol. II (1958): 211, and ‘Reduction sentences and disposition concepts’, in Schilpp, The Philosophy of R. Carnap, p. 593, as well as G. Hempel, ‘Carnap and the philosophy of science’, in Schilpp, The Philosophy of R. Carnap, p. 694. 27 Carnap, ‘Testability and Meaning’, §9, p. 149 and §10, p. 152. 28 Schilpp, The Philosophy of R. Carnap, p. 948.

Dispositions and Counterfactuals

57

for the explication of subjunctive and in particular of counterfactual conditionals. Presumably it will then also be possible to introduce disposition terms by explicit definitions.’29 We still lack a satisfactory logic of ‘causal modalities’, but there have been some interesting attempts to construct a logic for counterfactuals and we may consider what happens to Carnap’s predicaments when a disposition term is explicitly introduced using counterfactual conditionals. Disposition and Counterfactual Conditional Oversimplifyingly, it may be said that all logics for counterfactuals have been constructed on the basis of what has been called the ‘Ramsey test’, which can be stated as follows: ‘In general we can say with Mill that ‘If p then q’ means that q is inferrable from p, that is, of course, from p together with certain facts and laws not stated but in some way indicated by the context.[…] If two people are arguing about ‘If p will q’ and are both in doubt as to p, they are adding p hypothetically to their stock of knowledge and arguing on that basis about q’.30 The idea is the following: we can determine the truth-value of a counterfactual like ‘if Hitler had not invaded Russia, Germany would nothave lost the war’ only if we admit that between the antecedent ‘Hitler has not invaded Russia’ and the consequent ‘Germany has not lost the war’, there is a stronger tie, or connection, than the one expressed by the philonian conditional of standard logic; and this is so because the antecedent alone does not imply the consequent but the antecedent plus a set H of supplementary conditions implicitly admitted. First, in the ‘derivabilist’ point of view of Chisholm and Goodman, the aim was ‘to render a subjunctive conditional […] into an indicative statement which will say the same thing’,31 in other words, to be able to express in classical terms what ordinary language expresses using terms like: ‘if p were (had been) the case, then q would be (have been) the case’. This amounts to determine if there are classical logical criteria for the selection of the conditions H to be adduced to the antecedent to obtain the consequent, respecting the minimal requirement that H have not, as consequence, the negation of the antecedent. The chief merit of Goodman’s paper (1955, written in 1947) was to show that one came against three big difficulties in this research: 29 Ibid., p. 952. 30 This is the quotation of Ramsey made by Chisholm in his 1946 paper, ‘The Contraryto-Fact Conditional’, p. 489. Stalnaker took again this reference to Ramsey in his 1968’s paper which is at the beginning of the treatment of counterfactuals in terms of ‘possible worlds’ and stated in his own fashion this ‘test’: ‘First add the antecedent (hypothetically) to your stock of belief; second, make whatever adjustments are required to maintain consistency (without modifying the hypothetical belief in the antecedent); finally, consider whether or not the consequent is then true’ (‘A Theory of Conditionals’, in W.L. Harper, R. Stalnaker and G. Pearce (eds), Ifs ( Dordrecht, 1981) p. 44. 31 R. Chisholm, ‘The Contrary-to-Fact Conditional’ p. 486.

Dispositions and Causal Powers

58

-

non-deductivism: one cannot express in purely deductive terms (classical) the content of H; circularity: to express H it has to be made use of counterfactuals; it is the ‘cotenability’ problem; shifting: for the same antecedent, the set H can change according to the consequent (the kangaroo problem, context sensitivity).32

With the introduction of ‘possible worlds’ semantics, things do not change radically and one finds again approximately the same worries. The success of this kind of approach depends more, in fact, on its attraction for imagination than on its intrinsic qualities, even if some interesting work has been done on ‘deviant’ logics along this line of thought. We shall satisfy ourselves with taking again some basic ideas of David Lewis’s ‘ptolemaic’ semantics. Instead of the idea borrowed from Ramsey that to determine the truth-value of a counterfactual, minimal adjustments to our beliefs, when theantecedent is added, have to be considered, things can be represented as follows: one supposes that the set of possible worlds is arranged in a sequence of concentric spheres centred around our world, so that we can represent the ‘distances’ of possible worlds from ours. Each sphere contains ‘just those worlds that resemble [ours] to at least a certain degree’33 and so contains all the worlds belonging to its inner spheres. If one writes a counterfactual conditional ‘ϕ □→ ψ’ (‘ϕ’ and ‘ψ’ being metavariables) and calls ϕ-world, a world in which ϕ is true, we have,34 as truth-condition for ‘ϕ □→ ψ’ (in our world): -

‘ϕ □→ ψ’ is true in our world iff some sphere S contains at least one ϕ-world w and ψ is true in every ϕ-world belonging to S (in other words: ϕ → ψ is true in all the worlds belonging to S, ‘→’ being the classical conditional). More intuitively, that amounts to say that ψ is true in all the ϕ-worlds that are at least as close to ours as w.

It results from this presentation of a logic for counterfactuals that important properties of the classical notion of consequence are lost: weakening (monotony), transitivity 32 We do not follow here Goodman’s presentation but the clear and precise one given by S.O. Hanson, ‘The Emperor’s New Clothes’ in G. Crocco, L. Fariñas de Cerro and A. Herszig (eds), Conditionals: from Philosophy to Computer Science (Oxford, 1995), pp. 13-31. 33 D. Lewis, Counterfactuals (Cambridge, 1973), p. 14. 34 We disregard here the case where no sphere contains at least one ϕ-world (in this case, ϕ □→ ψ would be vacuously true); in such a case, we find again the same problem posed by the explicit definition of the form ∀x[Dx ↔ (Px → Qx)]. Let b be an objet such that no sphere around our world contains a Pb-world, then the counterfactual ‘ Pb □→ Qb’ would be true in our world and so the disposition D could be attributed to b. Lewis leaves aside such counterfactuals as ‘unentertainable’ (Counterfactuals, p. 16).

Dispositions and Counterfactuals

59

35

and contraposition. So it is a kind of deviant logic and we shall see below that this ‘deviance’ causes some effects stranger than those of the introduction of disposition predicates by means of reduction sentences. Let us examine quickly what the two ‘defects’ attributed to the Carnapian introduction procedure of disposition predicates (region of indeterminateness and factual consequences) when such predicates are defined in terms of counterfactuals. In addition, we shall see what happens to the analytic/synthetic distinction in the counterfactualist perspective. In what follows, we shall admit that in a ‘linguistic frame’ containing a logic for counterfactuals (which is all the more in keeping with Carnapian tolerance), a disposition predicate ‘D’ can be explicitly defined as: ∀x[Dx ↔ (Px □→ Qx)] For example: x is soluble in water (‘Dx’) =df if x were put into water (‘Px’) then x would dissolve (‘Qx’). The Problem of the ‘Region of Indeterminateness’ One remembers that Carnap takes in ‘Testability and Meaning’ §9, an example aiming to show that, even if one can further determine the meaning of a predicate introduced by a reduction pair, there may remain a ‘region of indeterminateness’. Let us consider again this example, which had troubled Chisholm. Let ‘soluble’ be introduced by means of the bilateral: if x is put into water, then x is soluble iff x dissolves. As it has been noticed, this leads to allow the application of the predicates ‘soluble’ and ‘insoluble’ only in case the antecedent is true, i.e. only in case an object is put into water. Now, suppose that on the basis of various investigations, a general law has been established stating that: ‘two bodies of the same substance are either both soluble, or both insoluble’. Now consider an object O1 of substance S1, not be put into water. If another object of the same substance is or has been put into water and dissolves, ‘soluble’ can be attributed to O1: O1 is soluble. Now let an object O2 of substance S2 put into water. If no object of the same substance is or has been put into water, neither soluble not insoluble can be attributed to O2. O2 is in the region of indeterminateness in which the meaning of soluble is not fixed, that is: it cannot be said that O2 is soluble or that it is not. Here we have what upsets the Fregean super-ego of every serious logician.

35 Noting the consequence relation by ⏐=, the weakening rule amounts to admit that if we have ∆ ⏐= ϕ, we have also ∆, ∆’ ⏐= ϕ, ∆ and ∆’ being sets of formulas; transitivity means that if we have ∆ ⏐= ϕ, and ∆, ϕ ⏐= ψ , we have also ∆ ⏐= ψ (if ∆ is empty, this is the classical ‘modus ponens’).

60

Dispositions and Causal Powers

Now, the question is: what does this indeterminateness become in the counterfactualist’s perspective. Instead of the bilateral, we have now the explicit definition: x is soluble in water =df if x were put into water then x would dissolve. Let us suppose, as above, that object O2 of substance S2 is not put into water and that no object of the same substance, till now, has been put into water. What about ‘O2 is soluble’? Has this sentence truth conditions? Let us try to answer this question in the ptolemaic terms of D. Lewis. We have to evaluate the sentence: ‘if O2 were put into water, O2 would dissolve’. In order to do this, one has to look at the nearest worlds in which O2 is put into water, i.e. those which differ minimally from ours, The question is: is a world in which O2 is put into water and dissolves closer to or not from ours than a world in which O2, put into water, does not dissolve? It can be supposed that in those worlds, the general law about the solubility or insolubility of objects of same substance still holds, since they differ from ours only minimally. The predicament is that to answer this question, we would have to establish whether a world in which objects of substance S2 dissolve is closer to ours than a world in which they do not dissolve. However, it appears at once that to answer this question, we should need data about the solubility in water of objects of substance S2 in our world; but this is precisely what we lack. So, we are not able to decide whether a world in which objects of substance S2 dissolve is closer to ours than a world in which they do not dissolve: therefore, we cannot, in principle, determine the truth value, in our world, of the counterfactual: ‘if O2 were put into water, then O2 would dissolve’. The meaning of the predicate ‘soluble’ is no more determinate in the case of O2, than it was in the Carnapian perspective. It is easy to see that if, on the contrary, we had been dealing with an object O1 of substance S1, the evaluation of the counterfactual would have been unproblematic: it is clear enough that a world in which O1 is put into water and dissolves is much closer to ours than a world in which, put into water, O1 does not dissolve: for this would mean that in this second world, the general law about substances does not hold. Generally speaking, these considerations amount to make clear that, in order to be able to determine the comparative similarity of the worlds in which the antecedent of a counterfactual is true, we need data about our world; but those data are precisely those that we lack when we are in the ‘region of indeterminateness’ which makes the predicate inapplicable for Carnap. This is just a rewording of this characteristic of Lewis’s models that there is no non-trivial purely logical measure of the proximity of a world to another one. In ‘derivabilist’ (or ‘metalinguistic’) terms, we find again what had been shown by Goodman: it is not possible to state precisely in pure deductive (classical) terms the content of the set H to be added to the antecedent in order to infer the consequent. So, we see that we are not in a better position whether we enjoy the delights of the ‘counterfactualism’ or accept the Carnapian scandalous indeterminateness:

Dispositions and Counterfactuals

61

a disposition predicate is applicable in Carnap’s sense only if the corresponding counterfactual can be evaluated, and reciprocally. In other words: the ‘region of meaning’ of a disposition predicate is exactly the one in which the truth-value of the corresponding counterfactual can be determined. The Question of the Factual Consequences In contrast with what happens with one bilateral, the introduction of a disposition predicate by two bilaterals has factual consequences. Formally this means that if D is introduced by ∀x(P1x→(Q1x ↔ Dx) ∀x(P2x → (Q2x ↔ Dx ) we have the following consequence: ∀x~(P1x ∧ Q1x ∧P2x ∧~Q2x)) ∧ ∀x~(P2x ∧ Q2x ∧ P1x ∧~Q1x) that is to say: ∀x[(P1x ∧ Q1x ∧ P2x) → Q2x] ∧ ∀x[(P2x ∧ Q2x ∧ P1x) → Q1x] These sentences do not include any more the disposition predicate ‘D’ and in order to simplify, it can be admitted that the predicates ‘P1’, ‘Q1’, ‘P2’, ‘Q2’ are primitive descriptive. Let the disposition predicate ‘irritable’ be introduced by the two bilaterals: B1: for all x, if x is jostled, x is irritable iff x shouts and B2: for all x, if x is under the rain, x is irritable iff x weeps The factual consequences are, on one hand that if someone is jostled, under the rain, and shouts, then he, or she, weeps; on the other hand, if someone is jostled, under the rain and weeps, then he, or she, shouts. It does not seem unreasonable, for a counterfactualist to express this by something like: ‘if x were jostled, under the rain and shouting, then he, or she, would weep’. Now, let us suppose that we meet a person, say Adele, who is jostled, under the rain and who weeps but who does not shout, then we would have to admit that the extension of the predicate ‘irritable’ as introduced by B1, does not coincide with the extension of the predicate ‘irritable’ as introduced by B2. Indeed, Adele under the rain, weeps: so we may apply to her ‘irritable’ in virtue of B2. Adele is jostled but does not shout: we cannot apply to her ‘irritable’, in virtue of B1. So we must admit

62

Dispositions and Causal Powers

that there are two predicates, ‘irritable1’ and ‘irritable2’ where we thought there was only one. As we saw, this is just the price of introducing predicates by means of reduction sentences: we are not a priori sure that the introduced predicate can be attributed consistently; it depends on the factual truth of the consequences above. As a result, we could be led to modify the extension of the predicates so introduced if further investigations were to show the falsehood of those factual consequences. The counterfactualist account of this little affair is rather odd. The first bilateral becomes an explicit definition of ‘irritable’: x is irritable =df if x were jostled, x would shout ∀x[Dx ↔ (P1x □→ Q1x)], the second: x is irritable =df if x were under the rain, x would weep ∀x[Dx ↔ (P2x □→ Q2x)] The troubling thing is that it is quite possible that the two interior counterfactuals are true, whereas the counterfactuals corresponding to those mentioned above, that is: ∀x[(P1x ∧ Q1x ∧ P2x) □→ Q2x] or ∀x[(P2x ∧ Q2x ∧ P1x) □→ Q1x] would be false; in other words, for a constant ‘b’, the three sentences: P1b □→ Q1b P 2 b □ → Q 2b ~[(P1b ∧ Q1b ∧ P2b) □→ Q2b] or ~[(P2b ∧ Q2b ∧ P1b) □→ Q1b] could be simultaneously true (and even the counterfactual (P1b ∧ Q1b ∧ P2b) □→ ~Q2b, for example). This means, for example, that the two counterfactuals C1: ‘if Adele were jostled, Adele would shout’ and C2: ‘if Adele were under the rain, Adele would weep’ could be true (which allows us to say consistently that Adele is irritable), while the counterfactual

Dispositions and Counterfactuals

63

C3: ‘if Adele were jostled, under the rain and were shouting, Adele would weep’ would be false. This seems quite odd and paradoxical, and from the classical point of view, it is really so because C3 is no more than a weakening of C2 and, classically, from the truth of C2 follows necessarily that of C3. But this is precisely not the case in the ‘logic’ for counterfactuals: the weakening rule does not hold in it, as we noticed above, and this means merely that C2 can be true and C3 false since nothing warrants that the weakening of the antecedent in C2 preserves the truth of the counterfactual conditional. All this can be represented in the following diagram:

64

Dispositions and Causal Powers

On account of A, ‘P1b □→Q1b’ is true w. On account of B, ‘P2b □→Q2b’ is true in w. On account of C, ‘(P1b ∧ Q1b ∧ P2b) □→Q2b’ is false in w and ‘(P1b ∧ Q1b ∧ P2b) □→~Q2b’ is true in w.

Q1b

P1b

P2b

C: P1b, Q1b, P2b, ~Q2b

A

B

w

Figure 1

Q2b

Dispositions and Counterfactuals

65

Analyticity It remains to have a glance at the question of the Excluded Middle. As we remarked, ‘Testability and Meaning’ aims to conciliate two seemingly antinomic perspectives: on one hand, to introduce predicates non ‘determinate’ for all objects and so violating the Fregean requirement of sharp boundaries; on the other hand, to accept only predicatea with ‘empirical meaning’. It is by defining various kinds of ‘reductibility’ for predicates (and by choosing the corresponding languages) that Carnap may pretend that even if a predicate ‘R’ has no full determinate ‘meaning’ (in a Fregean sense), it is nonetheless meaningful in a wider sense and so falls under the Excluded Middle Principle. Things are not so simple in the counterfactualist account of this affair. In the counterfactualist frame, ‘R’ can be eliminated and so we have to come back to the standard criteria holding for statements whose predicates are introduced by means of explicit definitions. The predicament, as it is well known, comes from the way the negation of a counterfactual is understood. To say that the Eiffel Tower is not soluble in water, is probably to say that if the Eiffel Tower was put into the Seine, it would not dissolve; in other words, to deny a counterfactal, is not to deny the whole of the formula ‘Pb □→ Qb’ (i.e. ‘~(Pb □→ Qb’)); such an external negation would only mean, in ptolemaïc terms, that in the sphere (around our world) containing at least one Pb-world, the Eiffel Tower does not dissolve in at least one Pb-world of this sphere; but this does not at all rule out that in other Pb-world of this same sphere, the Eiffel Tower dissolves. Now, it seems reasonable to think that, in asserting that the Eiffel Tower is not soluble in water, the case of Eiffel Tower put into water and dissolving, is ruled out. This amounts to deny the consequent of the counterfactual while preserving the antecedent, i.e. ‘Pb □→ ~Qb’. If this analysis is endorsed, then the negation of ‘Pb □→ Qb’ is not ‘~(Pb □→ Qb)’ but rather ‘Pb □→ ~Qb’. Goodman as well as Chisholm underlined this fact and Stalnaker followed them in his foundational paper of 1968.36 Concerning our little problem, once eliminated ‘R’ in ∀x(Rx∨~Rx), we get the sentence (‘conditional excluded middle’): ∀x[(Px □→ Qx) ∨ ( Px □→ ~Qx)]. The predicament is that for such a formula to be valid, we ought to be able, in terms of ‘possible worlds’, to select the single closest world (from the world in which the counterfactual is evaluated) in which the antecedent is true (it is what Stalnaker’s ‘selection function’ does) in order to evaluate the material conditional corresponding to the counterfactual to be evaluated. We have here a most disputable hypothesis that Lewis did not accept for good reasons (what is the closest world in which Nantes 36 By Stalnaker’s thesis that the evaluation of a counterfactual has to be made in the nearest antecedent-world, the two negations are equivalent.

66

Dispositions and Causal Powers

would be ‘less than 400 km from Paris’?). If such a hypothesis is admitted, the formula above is ‘analytic’; if not, as for Lewis, it is no more analytic.37 So, what was analytic from the Carnapian point of view, becomes problematic from the counterfactualist point of view and depends on the admission, or not, of the conditional excluded middle. Concluding Remarks The working out of the logics of counterfactual did not aim at escaping the unpleasant characteristics of the Carnapian solution to the problem raised by disposition terms, consisting of introducing them by means of reduction sentences; so, it cannot be concluded from the remarks above to a failure of these attempts. Since these logics assume the overtly deviant character of the conditional counterfactual, it is not quite a surprise that they only find again the same difficulties, in disguise. One of the chief merits – but a negative one! – of the work on the logics of counterfactuals is, for what concerns us here, to have shown that it is really possible to come back to the orthodoxy of the usual explicit definitions, but that this does not result in a real improvement. To the ‘region of indeterminateness’ of the predicate introduced by means of reduction sentences corresponds, as we saw, the impossibility, in principle, to evaluate the counterfactual definiens. It could be argued that the introduction of a disposition predicate by means of explicit definition, would make the theory of confirmation developed by Carnap in chapter six of ‘Testability and Meaning’, useless, since it would be enough to require, for the meaningfulness of such a predicate, that the predicates occurring in the counterfactual definiens be observable (or eliminable to the benefit of observable predicates). But, here again, nothing is achieved: contrary to what is the case for the hidden side of the Moon, we have no logical insurance that the notion of an observable predicate keeps the same meaning when we ‘go’ from one world to another one. As we noticed, Carnap tried to conciliate the ‘open’ character of the concepts used by scientists with what he thought to be the requirements of empiricism. In so doing, he was going along what seemed to Quine, for example, a slippery road. At least, it was rather less slippery than the one followed, later on, by the logics for counterfactuals. This chapter tried to show no more than that.

37 Let ‘ϕ □→ ψ’: it may be the case that among the ϕ-worlds of the ‘nearest’ sphere, some are ψ-worlds and other are ~ψ-worlds; in this case, both counterfactuals ‘ϕ □→ ψ’ and ‘ϕ □→ ~ψ’ are false.

Chapter 2

Filled In Space Stephen Mumford

Dispositions, their Grounds, and Humeanism An argument has been presented elsewhere – the Ungrounded Argument – with the conclusion that there are ungrounded dispositions.1 An ungrounded disposition is one with no grounding at all, not categorical nor dispositional, other than itself. Some time ago, David Armstrong2 proposed the view that all dispositions have categorical bases. Prior, Pargetter and Jackson3 called this the causal base. Hugh Mellor4 countered that the base of a disposition need not be categorical but could be, instead, a further disposition. There has been recent work, however, that has challenged the assumption of global groundedness, the thesis that a disposition must have a ground, of some kind, other than itself.5 C.B. Martin6 is acknowledged as a pioneer who first raised the prospect of challenging the received view but the idea of there being ungrounded dispositions was not overtly endorsed until Molnar7 and Ellis,8 though it was also considered by Blackburn9 without an endorsement. The issue is one of the key battlegrounds of contemporary metaphysics as it crystallises the differences in ontological outlook between Humeanism and anti-Humeanism. This paper concentrates on that difference in outlook. Neo-Humeans will find the 1 S. Mumford, ‘The Ungrounded Argument’, Synthese 149(3)(2006): 471-489. 2 D.M. Armstrong, A Materialist Theory of the Mind (London, 1968), pp. 85-88. 3 E. Prior, R. Pargetter and F. Jackson, ‘Three Theses about Dispositions’, American Philosophical Quarterly, 19 (1982): 251-257. 4 D.H. Mellor, ‘In Defence of Dispositions’, Philosophical Review, 83 (1974): 157181. 5 J. McKitrick allows that a disposition could qualify as its own ground, which is discussed in Mumford. J. McKitrick, ‘The Bare Metaphysical Possibility of Bare Dispositions’, Philosophy and Phenomenological Research, 64, No. 2 (2003): 349-369. S. Mumford, ‘The Ungrounded Argument’. 6 C.B. Martin and J. Heil, ‘Rules and Powers’, Philosophical Perspectives, 12 (1998): 283-312. 7 G. Molnar, ‘Are Dispositions Reducible?’, The Philosophical Quarterly, 49 (1999): 1-17. G. Molnar, Powers: a Study in Metaphysics, S. Mumford (ed.) (Oxford, 2003). 8 B. Ellis, Scientific Essentialism (Cambridge, 2001). B. Ellis, The Philosophy of Nature: a Guide to the New Essentialism (Chesham, 2002). 9 S. Blackburn, ‘Filling in Space’, Analysis, 50 (1990): 62-65.

Dispositions and Causal Powers

68

conclusion of the Ungrounded Argument unacceptable so will be obliged to look for an error in one or more of its premises. I will argue that the conclusion is acceptable and that Humeans do not find it so because they place unjustified strictures on what can count as actual and what can count as a property. All the premises of the Ungrounded Argument seem strong and have support from philosophy or science or both. If the argument appears sound, then Humeanism is in serious difficulty as the required broadening of the concepts of actuality and of a property is the abandonment of the Humean strictures. The ultimate acceptance of irreducible modal properties would go against one of the defining commitments of Humean metaphysics. The Ungrounded Argument The Ungrounded Argument (abbreviated to UA) was an attempt to make explicit the direct and positive case for the thesis that there are ungrounded dispositions, which had previously been implicit or advocated by intuitions alone. It was an attempt to form a demonstrably valid argument from premises that appear true. The full merits of the argument were assessed elsewhere, which I will not repeat in any detail. As a context for the discussion in the present paper, UA is re-presented in brief here. The argument has four premises. 1. 2. 3. 4.

There are subatomic particles that are simple. That which is simple has no lower-level components or properties. The properties of subatomic particles are (all) dispositional. The grounds of a dispositional property can be found only among the lowerlevel components or properties of that of which it is a property.

By modus ponens from 1 and 2, we deduce that subatomic particles have no components 2a. 3 tells us only that subatomic particles have dispositional properties but from it and the modus tollens using 2a and 4, we can conclude that: 5. The dispositional properties of subatomic particles have no ground. What really interests us is the existential generalization of 5, which is: 6. There exist some ungrounded dispositions.10

10 I also presented a simplified version of UA, without quantifiers but demonstrably valid. Where sp’s is short for subatomic particles: S = there are sp’s that are simple; L = sp’s have lower-level components or properties; D = the properties of sp’s are dispositional; G = the properties of sp’s are grounded, the following argument can be presented with proposition numbers corresponding roughly to those in the fuller version of UA: 1. S 2. S → ¬L 2a. ¬L [1, 2 MPP] 3. D

Filled In Space

69

The significance of 6 is that it is contrary to the thesis of global groundedness. Premises 1 and 3 are a posteriori claims of science, but have a weight of evidence and theory behind them. 1 is necessary and 3 is apparently so. Premises 2 and 4 are philosophical. 2 is a priori. 4 is to a degree stipulative to the extent that by grounds we mean properties or components at a level relatively micro to that which it grounds. Hence, for the micro-reductive explanation that motivates global groundedness, we do not allow things to ground themselves, in this sense of ground, nor do we allow relatively higher-level grounding of which Harré’s ultra-grounding11 is an example. The stipulation in 4 is thus based on the micro-reductive programme. This programme is one that has had some success and appears effective for common macro-level dispositions such as solubility or fragility. These dispositions are explained in terms of their grounds, which are identifiable lower-level properties or components of the soluble or fragile thing. If UA is successful, however, such micro-level reductive explanation cannot be universal. Universal micro-reduction is the overt target of the argument, but not the only target. If the argument is sound, it lays down a challenge to Humean supervenience as it has been developed by neo-Humeans. The Humean Response: Empty Space In later sections, I will consider the Humean strictures on what can count as actual and as a property. To understand the importance of this, and thus the importance of the Ungrounded Argument, we need first to see what would be the likely response to UA from a Humean. While the response I consider is not the only one that is possible for Humeanism, it is the one that I think strongest. This response challenges the conclusion as being counterintuitive if not incoherent. There must, therefore, be some fault in one or more of the premises of the argument. The fault may be in premise 3, as the properties of subatomic particles – spin, charge and mass – might not qualify as dispositional after all. It might be, against premise 1, that there are no simples. Or it might be that the restriction on grounds in premise 4 is too strong, hence ultra-grounding of even ‘dispositions all the way round’12 might come into the reckoning. However, I will not consider all the options the Humean might pursue in order to counter one or more premise. Instead I will consider what is most likely their reason for seeking weaknesses in the premises, namely that 6 is unacceptable to them. 4. G → L 5. ¬G [2a, 4 MTT] 6. D & ¬G [3, 5 &I] 11 R. Harré, Varieties of Realism (Oxford, 1986) pp. 295-301. 12 R. Holton, ‘Dispositions All The Way Round’, Analysis, 59 (1999): 9-14. I am not accusing Holton of defending Humeanism. On the contrary, he defends the claim that every property could be grounded even if all were dispositional. But the position he describes would be ruled out by [4] as long as ‘being a lower-level property’ is taken as transitive and irreflexive.

70

Dispositions and Causal Powers

Grounds for dispositions were originally sought as a means of accounting for the proposed truth that a dispositional property can be possessed in between its manifestations or even if never manifested at all. We distinguish, for instance, between being soluble, which is the disposition, and being dissolved, which is the manifestation. Disposition ascriptions are categorical in the sense that dispositions are ascribed unconditionally. We are prepared to say that x is soluble unconditionally, rather than soluble if something else is the case. If anything is conditional, it is only the manifestation of the disposition; where a manifestation is simply an event, which need not be manifest to an observer. There are two main accounts of what it is for a property to be dispositional. The first, and most commonly held, is the view that a dispositional property is one whose ascription entails a stronger-than-material conditional.13 Grounds are assistance to this account. We might wonder, where a disposition is unmanifested, why such a stronger-than-material conditional is true. If every true contingent statement needs a truthmaker, what would be the truthmaker of such a conditional where it is counterfactual, which it would be where the disposition was unmanifested? The ground of a disposition is invoked at this stage as being just such a truthmaker for the counterfactuals that hold. Where there is a true counterfactual, of the relevant kind, a disposition ascription is justified. Hence, a disposition ascription may remain true when the disposition is unmanifested ultimately because its ground persists in the subject of ascription throughout such times. The disposition ascription is true if and only if the ground is present. The Humean might well want to then say that all such grounds are categorical or even that dispositions are nothing more than their categorical bases, but we can leave aside that discussion here.14 Where a disposition is ungrounded this account appears to be unavailable. There is no ground, so there is no truthmaker for any associated counterfactual conditional. It might be objected, therefore, that there is no such true conditional and thence that there is no disposition present at all. This raises what I call the question of the being, actuality or existence of the ungrounded, unmanifested disposition. The manifestations of dispositions are in events or states, which will be accepted as actual or real enough by all. The actuality of unmanifested dispositions was thought, however, to reside only in their grounds. UA has led us to a case of a disposition with no ground. Where it is unmanifested, therefore, it will seem that there is nothing actual to it at all. The disposition seems like nothing: just empty space. This was why Simon Blackburn, a neo-Humean who was nevertheless sensitive to the possibility of ungroundedness, wondered how space might be filled in. The second account of dispositions, which has emerged more recently, is that dispositions are properties that are directed at their manifestations. They are

13 See W. Malzkorn, ‘Realism, Functionalism and the Conditional Analysis of Dispositions’, The Philosophical Quarterly, 50 (2000): 452-469 for a recent defence. 14 It is discussed in S. Mumford, Dispositions (Oxford, 1998), ch. 8.

Filled In Space 15

16

71 17

intentional states (Place, Martin and Heil and Molnar ). Molnar has argued in detail that the Brentano thesis is false and intentionality is the mark of the dispositional, not of the mental. One of the marks of intentionality that dispositions meet, even physical dispositions, is intentional inexistence. Dispositions are directed towards manifestations that might never be. Nevertheless, the disposition is still there even if its object is never actualized. Molnar, however, allows irreducible and ungrounded dispositions.18 The same question of being will strike against such an account. If a disposition has no ground and is essentially a property that is directed towards a nonexistent state, where it is unmanifested, it might be questioned what sort of existence such a property has. Such existence would seem Meinongian at best. There seems nothing actual about such a property. Again, there seems nothing but empty space. Those of us who endorse the Ungrounded Argument now have a duty. We must show that the Humeans’ worries about the being or actuality of ungrounded dispositions can be met. We must show that we are able to fill in the spaces that they claim would be left empty. In doing so, however, we will have to argue that it is Humeanism itself that has generated the problem. While Humeans would have to seek a weakness in the Ungrounded Argument, to show that we need not accept the counterintuitive conclusion 6, I will go on to argue that 6 appears counterintuitive only because of the inadequate resources that Humeanism will grant us. If one rejects Humeanism, is no longer counterintuitive. There would then remain nothing that ruled against ungrounded dispositions being fundamental, properties and real. Humean Supervenience The objectionability of 6 is based on a metaphysic that rules out modal properties and permits only causally inert, occurrent properties and particulars. Claims can truly be made about causation and about modality, in this metaphysics, but their meaning is provided in terms of relations between types of events, for causation, or between worlds, for modality. In this section, therefore, I examine the elements of the position that is now widely known as Humean supervenience. This metaphysical stance has been articulated better than anyone by David Lewis. He has characterized the position in a number of ways, for example: Humean supervenience is named in honor of the greater [sic.] denier of necessary connections. It is the doctrine that all there is to the world is a vast mosaic of local matters

15 U.T. Place, ‘Intentionality as the Mark of the Dispositional’, Dialectica, 50 (1996): 91120 and ‘Intentionality and the Physical: a Reply to Mumford’, The Philosophical Quarterly, 49 (1999): 225-231. 16 C.B. Martin and J. Heil, ‘Rules and Powers’. 17 G. Molnar, Powers: a Study in Metaphysics, ch. 3. 18 G. Molnar, ‘Are Dispositions Reducible?’.

Dispositions and Causal Powers

72

of particular fact, just one little thing and then another … We have geometry: a system of external relations of spatio-temporal distance between points. … And at those points we have local qualities: perfectly natural intrinsic properties which need nothing bigger than a point at which to be instantiated. For short: we have an arrangement of qualities. And that is all. … All else supervenes on that.19

This statement does not appear explicitly to rule out dispositions or modal properties from being among the ‘local matters of particular fact’. But the following passage, and indeed much of Lewis’s work, is evidence of his view that in the ideal theory they would be among the supervenient things in a world. … if two worlds match perfectly in all matters of particular fact, they match perfectly in all other ways too – in modal properties, laws, causal connections, chances, ….20

Lewis’s position sets up a divide between ‘all there is to the world’ and what supervenes on it. The presumption is that ‘all there is’ is non-modal, non-causal, nonchancey. The properties to be found are thereby occurrent, fully manifest, carrying no hidden intrinsic possibilities. Hence features such as chances, laws and causes have a right to exist only if they are supervenient on the Humean base. They are not suited, in their own right, for inclusion within the base. We need to know a little more about this Humean, subvenient base and what justifies the initial presumption that it would ideally be non-modal. Is this just a prejudice or is there good reason for it? Where Lewis discusses the nature of the base in more detail, it is not at all obvious why modal properties such as dispositions would be excluded. For example: There may be properties of mass, charge, quark colour and flavour, field strength and the like; and maybe others besides, if physics as we know it is inadequate to its descriptive task. Is that all? Are the laws, chances, and causal relationships nothing but patterns which supervene on this point-by-point distribution of properties?21

Similarly: We may be certain a priori that any contingent truth whatever is made true, somehow, by the pattern of instantiation of fundamental properties and relations by particular things. In Bigelow’s phrase, truth is supervenient on being.22 If two possible worlds are discernible in any way at all, it must be because they differ in what things there are in them, or in how those things are. And ‘how things are’ is fully given by the fundamental, perfectly natural, properties and relations that those things instantiate. … we may reasonably hope that physics – present-day physics, or anyway some not-too-distant improvement thereof

19 20 21 22

D. Lewis, Philosophical Papers II (Oxford, 1986), pp. ix-x. Ibid., p. 111. D. Lewis, On the Plurality of Worlds (Oxford, 1986), p. 14. J. Bigelow, The Reality of Numbers (Oxford, 1988), pp. 132-133 and 158-159.

Filled In Space

73

– will give us the inventory of all the perfectly natural properties and relations that ever appear in this world.23

Further : It [Humean Supervenience] says that in a world like ours, the fundamental properties are local qualities: perfectly natural intrinsic qualities, or of point sized-occupants of points.24

Nothing in these passages indicates why Humean Supervenience excludes dispositions – even ungrounded and unmanifested dispositions – from being among the natural properties that have being. Properties that are acceptable for inclusion in the Humean subvenient base are characterized in these excerpts as being a) those described by physics, b) intrinsic, c) point-sized occupants of points or local qualities, d) fundamental. The dispositional properties invoked in UA seem precisely these. It is hard to see, therefore, what prompted the search for an account of the supervenience of modal properties such as these. They are, after all, the properties physics invokes. Even orthodox accounts of the fundamental particles characterize their properties as spin, charge and mass and give a dispositional interpretation of each. The Ungrounded Argument uses such scientific accounts in support of premise 3. Dispositions can also be intrinsic properties as they require nothing else, not even their manifestations, for their own existence. Lewis and Langton allow this possibility.25 Dispositions can be point-sized occupants of points, as spin, charge and mass evidently are. They are, thus, fundamental and local. While, on the evidence Lewis presents, there is nothing in the statement of Humean supervenience that excludes dispositions from the subvenient base, it is clear from the way Lewis carried out his project that dispositions, along with laws, causes and chances were regarded as supervening. His project was to show how these things were not a fundamental part of being but, rather, supervenient on it. Most importantly, there could still be modal truth in a world without modal properties. Is there any answer, then, to why Lewis thought that modal truths were supervenient? Why were there no modal properties in their own right? The explanation, as indicated in the first quotation, is that Lewis accepted Hume’s denial of necessary connections in nature and regarded it as such a fundamental commitment that he named his position after him. In Hume, we find the following statement of this view:

23 D. Lewis, ‘Humean Supervenience Debugged’, in D. Lewis, Papers in Metaphysics and Epistemology (Cambridge, 1999), p. 225. 24 Ibid., p. 226. 25 Though whether dispositions are intrinsic depends on the correct theory of laws as they see a disposition as obtaining only in virtue of an intrinsic basis together with laws of nature, and such laws might turn out to be extrinsic. D. Lewis and R. Langton, ‘Defining “Intrinsic”’, pp. 123-124, in D. Lewis, Papers in Metaphysics and Epistemology.

74

Dispositions and Causal Powers The necessity of power, which unites causes and effects, lies in the determination of the mind to pass from one to the other. The efficacy or energy of causes is neither plac’d in the causes themselves, nor in the deity, nor in the concurrence of these two principles; but belongs entirely to the soul, which considers the union of two or more objects in all past instances.26

This was a central claim of Hume’s philosophy that he repeats and develops at length. Hence, a Humean will not accept powers or dispositions as suitable items to be among the fundamental intrinsic properties of things. They are projected on to things by minds only. But if this is the main pillar upon which the case against UA rests, then it seems that there is ample room for debate. That all necessary connections are projected on to the world by the mind or soul is a strong and highly contentious claim. It seems vulnerable to attack, though as it involves such major, fundamental areas of philosophy, so I shall not pursue such a line here. There are to be allowed no necessary connections between the discrete elements in the subvenient base, according to Humean supervenience. In what sense, therefore, would dispositions violate this stricture? All elements must exist in total independence of all other elements. But dispositions, it is often said, are necessarily connected to their manifestation. Hence the properties being soluble and being dissolved would have some degree of necessity connecting them. Such necessity is evidently there, but not easy to specify accurately. It is not a necessity such that anything which is soluble must be dissolved. We have already seen that a disposition might not be manifested. It is not even the case that anything soluble will necessarily dissolve if in liquid. Because other factors can interfere with the process of dissolution in liquid (the solvent may be saturated or frozen, etc.), it might not occur. This has led some to say that any conditionals entailed by a disposition ascription can be ceteris paribus only27 or that they hold only in unspecifiable ideal conditions.28 The connection between a disposition and its manifestation is necessary in some sense, however. Perhaps it is this. It is dissolving that is the manifestation of solubility, and necessarily not anything else. It is not breaking, attraction, stretching or any other type of event. Being soluble and being dissolved seem, in this sense, to violate the principle of no necessary connections between distinct existences. There is a modal tie between them that does not exist between being soluble and being broken, for instance. Therefore, there is not complete unconnectedness between all the events or properties in the vast mosaic. Some, clearly, are more connected than others. There remains an opportunity for the Humean to give a different interpretation of the points just raised. Any necessity of a power, they might maintain, is merely conceptual: it is an analytic truth that dissolving is the manifestation of solubility. Hence this necessity is only in the minds of language users, not in the world. The Humean further claims that all that can be known of a power, and therefore all we 26 D. Hume, A Treatise of Human Nature (1739-40), ed. L.A. Selby-Bigge, (Oxford, 1888), Book I, Part III, Sect XIV, ‘Of the idea of necessary connection’, p. 166. 27 See Schrenk, Chapter 11, this volume, for further discussion of ceteris paribus. 28 S. Mumford, Dispositions, pp. 88-92.

Filled In Space

75

can claim of its being, is its manifestations. There is certainly scope for a point-bypoint discussion of all these issues but that would again take us too far afield from the present argument. The result of this initial investigation seems to be, therefore, that as Humean supervenience is explicitly stated, there is nothing in principle that would exclude dispositions from the subvenient base as they are fundamental and can be at a point. That dispositions are excluded by the Humean seems more the result of other views, seemingly fundamental to the position, about what counts as a property and what can count as actual. I will explore these issues more fully next. Actuality and Properties Might we exclude dispositions from a list of real properties on the ground that they are nothing more than possibilities? Although their manifestations may at times be real, the dispositions themselves might not quite pass the test for being actual properties. What would lead anyone to say that a manifestation, such as being dissolved, was a genuine and acceptable property but being soluble was not? In the case of the fundamental properties, is there any case to answer that only the manifestations of fundamental dispositions are real and the dispositions themselves are not? Of course, being dissolved is hardly likely to be among the fundamental properties, as it could not be instantiated at a point. The general issue, more accurately, is why might one allow non-dispositional properties into the subvenient base but not dispositional properties. In this section, I argue briefly that there is nothing in the concept of a property that would exclude dispositions from being properties. Hence it cannot be assumed that an ungrounded, unmanifest disposition has insufficient being. Charge produces attraction or repulsion to other objects, depending on their charges. It meets this mark of being a property. Are we to suppose that an elementary particular loses its charge just because nothing is in its electromagnetic field of influence? If so, why is another particular attracted or repulsed as soon as it comes in to the field? Isn’t the best explanation that the charge was there all along, an actual real property that persists whether or not it is manifested? There might be a parallel drawn here between our attitudes to ungrounded dispositions and phenomenalism about physical objects. To say that an object exists only when perceived is repellent to our intuitions and phenomenalism is often rejected on this ground alone. Why, then, not reject the view that a power exists only when manifested? A run through some of the main accounts of properties shows that there is nothing about dispositions that would stop them qualifying as such, even if they are ungrounded and unmanifested. Most importantly, such dispositions still are instantiable, capable of identity through each of their instantiations, and make a difference to the causal role of the particulars that instantiate them.

76

Dispositions and Causal Powers

To be a universal, as conceived by Russell29 (following Plato) in its transcendent form and Armstrong30 in its immanent form, the requirement is that something be instantiable – capable of instantiation – in exactly the same way but at different times and places. Immanent universals must be instantiated at some space time in order to be real, while transcendent realism permits properties that are uninstantiated. Dispositions can be properties as universals, if this is the correct theory of properties, whether the immanent or transcendent theory is best. To meet the instantiation requirement of immanent realism would not require manifestation of the disposition but only that something be soluble, charged, fragile, massive, and so on. There is no requirement that a property be non-dispositional, nor that it be grounded. The instantiation requirement is taken a stage further by David Lewis. Here properties are taken to be just the set of all their instances: ‘all of them, this – and other-worldly alike’.31 Lewis thus avoids the problem of accidentally coextensive properties. There will be none such if all the worlds in which properties are instanced are taken into the reckoning. Again, there is nothing in this theory that would explicitly rule dispositions out. So it was not Lewis’s own theory of properties that led him to exclude dispositions from his subvenient base. Trope theory offers a radically alternative account of properties and, indeed, replaces such a notion with one of exactly resembling tropes, which are qualitative particulars: abstract particulars as Campbell32 has said. This theory is blighted by the problem that resemblance must be either taken as a universal itself, so universals cannot be reducible to trope particulars, or an infinite series of resemblance tropes would have to be permitted (Daly,33after Russell34). Such a regress argument might be taken as a reason why there are no tropes but its real force is against the reduction of properties to particulars, which tropes are. Some, therefore, accept that there can be both properties and tropes, for example Lowe,35 who follows Locke in accepting both attributes (qualities, properties) and modes (property instances, tropes).36 Attributes would be universals, while modes would be their spatiotemporally located instances. On any of these accounts, dispositions are as fit to be tropes, modes or property instances as is squareness, being broken, dissolved or attracted. A disposition is as good a trope as anything. Sydney Shoemaker has said that properties are nothing but causal powers. Dispositions would be, on such an account, paradigmatic properties. To be a property,

29 B. Russell, The Problems of Philosophy (London and New York, 1912), ch. 9. 30 D.M. Armstrong, A Theory of Universals (Cambridge, 1978). 31 D. Lewis, On the Plurality of Worlds, section 1.5, p. 50. 32 K. Campbell, Abstract Particulars (Oxford, 1990). 33 C. Daly, ‘Tropes’, in D.H. Mellor and A. Oliver (eds), Properties (Oxford, 1997). 34 B. Russell, ‘On the Relation of Universals and Particulars’, Proceedings of the Aristotelian Society, 12 (1911-1912): 1-24. 35 E.J. Lowe, ‘Recent Advances in Metaphysics’, Facta Philosophica, 5 (2003): 3-24. 36 This is not the classical (Cartesian) distinction between attributes and modes. See the Introduction to this volume.

Filled In Space

77

37

according to Shoemaker, is just to be powerful. There is no essential requirement of groundedness here. Indeed, if a property is just a causal power, any such ground would have to be a further power anyway. Either way, dispositions, or powers, are properties. Nominalism has been catalogued and assessed in all its main forms by Armstrong.38 It is not a theory of properties or universals but, rather, a claim that there are no properties at all. In their place, the nominalist says there are only predicates, concepts, classes or resemblances that, the nominalist hopes, can be accounted for without appeal to universals. Dispositions make for predicates, concepts, classes or resemblances as good as any others. Self-Grounding and Truthmaking Having shown, first, that no adequate case has been made against dispositions being among the subvenient base properties and, second, that there is nothing merely in the concept of a property that excludes dispositions from being properties, we are now in a position to reassess the Humean charge that dispositions that are ungrounded would have insufficient being when not manifested. The claim we must make is that a disposition is sufficiently existent on its own, even if ungrounded and unmanifested. But then in what does its existence consist? Nothing, we must admit, but itself. Such a disposition provides its own grounding in actuality, hence McKitrick is right to speak of an ungrounded or ‘bare’ disposition as one that has no grounding except for itself. With such fundamental properties as these must be, we have reached the bedrock. We cannot say that such a property F supervenes on a more fundamental property G because the case is one where there are no further subvening properties. Neo-Humeans tell us that any such a property would have to be perfectly occurrent or categorical: it must manifest, at all times, all of which it is capable. But to accept this is to accept a fundamental level of properties that is utterly inert and powerless. Everything is passive and nothing active. Why would such a world be better credentialled to support our evidently active, powerful world than a world that has powers there at the bottom of it all? How could such a dead world support that which is built upon it? How could it produce a world in which things happened?

37 S. Shoemaker, ‘Causality and Properties’, (1980), repr.in D.H. Mellor and A. Oliver (eds), Properties (Oxford, 1997). Shoemaker’s account is slightly more complicated than this. To be a disposition is to be a function from properties to powers (pp. 233-234), which makes properties second-order powers: powers to produce first-order powers. This technicality need not divert us. Few distinguish between dispositions and powers and even when Shoemaker does, dispositions remain powers of a kind. If the present discussion pertains to Shoemaker’s first-order powers only, we could easily change terminology so that our subject is ungrounded powers. 38 D.M. Armstrong, Nominalism and Realism (Cambridge, 1978), chs 2-5.

78

Dispositions and Causal Powers

The contrasts between the dead world of mechanism and the active world of causal powers have been drawn by Ellis.39 He argues in detail for the advantages of powerful particulars. If the world really does contain such things, then there is no need to give an account of their being in terms of the being of something else. They are not to be reduced away or supervenient. If they are actual, they can ground their own behaviour, they can be directed towards manifestations, they can be truthmakers of counterfactual conditionals. They will be precisely what Hume and his followers sought to deny: modal properties. Modality would thus be a feature of our world, not of the relation between our world and others similar to it. Necessity and possibility would be intrinsic to our world. Hence, dispositions would not have to be accounted for in terms of counterfactuals, as many have tried. Dispositions would be among the truthmakers of counterfactuals and the whole elaborate construction of possible worlds would not be needed. Being such a truthmaker of a counterfactual, being that which supports a possibility or necessity, is more than enough to give a disposition the required being. We will not be able to find the verification of such truthmakers in anything that occurs, because ungrounded, unmanifested dispositions are not occurrences. But this brings us to the crux of the matter. What should we adopt as our reality test? Humeans look for events or occurrences. Their properties must be occurrent in the sense that nothing about them must be hidden; all must be manifest. But there is another reality test, that derives from Plato’s Eleatic Stranger’s.40 To be real, on this test, is to be powerful, so dispositions pass with flying colours. Under one interpretation of the test, manifestation of the power is not required. The Stranger said: I suggest that anything has real being that is so constituted as to possess any sort of power either to affect anything else or to be affected, in however small a degree, by the most insignificant agent, though it be only once.

The final clause, ‘though it be only once’, could be dropped by anyone realist about causal powers. It appears to be a requirement that a power be manifested at least once.41 There is no reason in principle why we could not place the same requirement on our ungrounded dispositions: they need not be manifested at all times or at any particular time but they must be manifested at least one time. But to be realist about dispositions is to accept them as real qua powers. It is to accept, pace Hume, a

39 B. Ellis, Scientific Essentialism (Cambridge, 2001). B. Ellis, The Philosophy of Nature: a Guide to the New Essentialism (Chesham, 2002). 40 Plato, ‘Sophist’, in E. Hamilton and H. Cairns (eds), The Collected Dialogues of Plato (Princeton, 1961), pp. 957-1017, at 247d-e. 41 There is an ambiguity in the statement of the test between (A) ‘having the power to affect or be affected at least once’ and (B) ‘having the power to affect or be affected, which actually does affect or is affected at least once’. Either interpretation fits the bare statement as it stands. The realism I advocate accepts (A) as a good reality test but not (B).

Filled In Space

79

42

distinction between a power and its exercise. Being powerful does not require that the power be exercised. To state that the power be exercised, ‘though it be only once’, looks like backtracking on, or at least having second thoughts about, realism. I would support, therefore, a form of the Stranger’s test that has no proviso that the power be exercised. Which reality test is right? It will be hard to find a principled, non-circular answer to this question, which seems to demand a reality test for reality tests. Perhaps we can only answer by saying that the test is right if it delivers the right answers. But Humeans and realists disagree over what the right answers are. Each ontology will bring with it a reality test. Given the strength of the Ungrounded Argument, with its scientific and philosophical appeal, together with the other work that powers can do for us, a reality test that rules out ungrounded dispositions seems to rule out the world we inhabit. This looks bad for the occurrence test and the Humean metaphysic that demands it. At the very least, we could claim that there remains an onus of proof on the Humean to show that their reality test that rules out ungrounded, unmanifested dispositions, has a sufficiently credible justification. Conclusion I conclude that there is no good reason why ungrounded powers or dispositions fall short of being genuine, actual properties. There is no reason, therefore, why they cannot be accepted into any subvenient base if such a thing there be. Indeed, current science appears to teach us that if there is such a subvenient base, dispositional properties are a part of it. We saw in the third section of the paper that Humean supervenience, as stated by Lewis, does not automatically exclude dispositions. It was then shown how Humean presuppositions about properties and actuality decree dispositions unacceptable and how Lewis has attempted to secure modal truths from worlds that have no modal properties. This view was reassessed in the last two sections of the paper, in which I showed that there was nothing in any of the main conceptions of properties that would rule dispositions any less real or actual than any other property and that being powerful and truthmaking of counterfactuals is enough to pass at least one reality test. Any empty space there is thought to be is filled in by the dispositional properties themselves. They are the grounds of their own manifestations. They are the being. They are as much qualified to fill in space as any other properties. We can say, therefore, with reference to the two main theories of what it is to be a disposition, that it is the ungrounded dispositions that are the truthmakers for the associated conditionals. Or, on the intentionality account, it would be they that were directed towards manifestations of a certain kind. In being so directed, in containing possibilities within themselves, they meet the Stranger’s test of reality. In being instantiable and often instantiated, they meet the other most common test of a real property. 42 ‘The distinction, which we often make betwixt power and the exercise of it, is equally without foundation’, D. Hume, A Treatise of Human Nature. Book I, Part III, Sect. XIV.

80

Dispositions and Causal Powers

The purpose of this discussion has been to show that the notion of an ungrounded disposition is coherent, even where it is one that remains unmanifested. There is no reason, therefore, to reject the Ungrounded Argument purely on the grounds that its conclusion is counterintuitive. The argument is counterintuitive only if we uphold Humean strictures on what it is to be actual and what it is to be a property. Such strictures, even though they seemingly have a powerful grip on some of our metaphysical intuitions, have yet to be given any serious justification. That they rule wrongly in the current case, counts against them.43

43 I am grateful to audiences at Cambridge (HPS), Cardiff, Durham, Lund and Paris where earlier versions of this much-changed chapter were presented. For comments that have contributed to this evolution, I would like to thank Helen Beebee, Nancy Cartwright, Rom Harré, James Ladyman, Anna-Sofia Maurin and Alexander Miller. For points raised in correspondence, I would like to thank Brian Ellis. For letting me see her paper in advance of publication, I am grateful to Jennifer McKitrick.

Chapter 3

Dispositions and Essences Claudine Tiercelin

From some time now, dispositions have no longer been viewed as ethereal or occult powers that have a pre-scientific status, threats or ghostly forces in need of redemption, mere promissory notes, some lazy or inaccurate way of talking that no self-respecting scientist should ever use.1 The pregnant spinsters have now acquired the respectable status of unmarried mothers, up to the point that they do lead a life of their own. Indeed, not only are they viewed as possible accidental properties of things: they are taken as essential properties of things in nature. Even more, for some philosophers, all things in nature are essentially dispositional. I have tried elsewhere2 to sketch the historical and philosophical reasons why there might have been such a revolution from pure elimination (as may be found under various forms in Boyle,3 Quine,4 Goodman,5 or Ryle6) to inflation (Ian Thompson,7 Rom Harré8, Nancy Cartwright9,

1 As D.H. Mellor already observed, some thirty years ago, in his famous ‘In Defence of Dispositions’, Philosophical Review, 83 (1974): 157-181. 2 ‘Sur la réalité des propriétés dispositionnelles’, in Cahiers de l’Université de Caen, nos. 38-39 (2002): 127-157. 3 R. Boyle, ‘The Origins and Forms of Qualities’, in M.A. Stewart (ed.), Selected Philosophical Papers of Robert Boyle (Manchester: Manchester University Press, 1979) pp. 1-96. 4 Cf. for ex. W.V. Quine, The Roots of Reference (La Salle, Ill., 1974), p. 11, or The Ways of Paradox and Other Essays, revd. edn (Cambridge, Mass., Harvard University Press, 1966), pp. 71-72. 5 N. Goodman, Fact, Fiction and Forecast (Indianapolis, 1955). 6 G. Ryle, The Concept of Mind (London, 1949). 7 I.J. Thompson, ‘Real Dispositions in the Physical World’, British Journal for the Philosophy of Science, 39 (1988): 67-79. 8 R. Harré, ‘Powers’, British Journal for the Philosophy of Science, 2 (1970): 81-101; R. Harré and E.H. Madden, ‘Natural Powers and Powerful Natures’, Philosophy 48 (1973): 209-230; R. Harré and E.H. Madden, Causal Powers: A Theory of Natural Necessity (Oxford, 1975). 9 N. Cartwright, Nature’s Capacities and their Measurement (Oxford, 1989): ‘It is not the laws that are fundamental, but rather the capacities’, p. 181; also see The Dappled World (Cambridge, 1999).

Dispositions and Causal Powers

82 10

Karl Popper , Hugh Mellor11, Simon Blackburn12) as regards dispositions, powers or capacities (with, in some cases, their idealistic consequences13). In what follows, I would like to concentrate on some less extreme versions of dispositional essentialism such as have been proposed by some authors, especially Brian Ellis, and in an even less extreme version, I think, by Stephen Mumford. I shall begin by presenting Ellis’s main lines of argument and then try to see whether some of the difficulties raised by his account and the answers suggested by Mumford are satisfactory, especially as far as a dispositionalist account of laws is concerned. Favoring myself some form of dispositional realism14 rather than essentialism, I shall finally less present here some arguments in favor of such a position than merely make some suggestions as to what seem to me the major difficulties which any kind of dispositionalism should be ready to face. Brian Ellis’s Conception of Dispositional Essentialism: the Main Arguments15 Ellis’s conception of dispositional essentialism (which he first exposed in an article written with Caroline Lierse in 1994) is part and parcel of the more general view he now entitles scientific essentialism, which can be summarized in three basic theses:16 1. Scientific essentialism deals with natural kinds such as events and processes and not only with objects and substances. According to Ellis, ‘the world is structured into hierarchies of natural kinds of objects and processes: it is not an amorphous world on which we must somehow impose our own system of 10 K. Popper, A World of Propensities (Bristol, 1990). 11 D.H. Mellor, especially ‘In Defence of Dispositions’; ‘Counting Corners Correctly’, Analysis, 42 (1974): 96-97 and ‘The Semantics and Ontology of Dispositions’, Mind, vol. 109, 436 (2000): 757-780. 12 For S. Blackburn, science finds dispositions ‘all the way down’, ‘Filling In Space’, Analysis, 50 (1990): 62-65. 13 Cf. S. Blackburn, Ibid. 14 I have tried to spell out some arguments in favor of such a position in ‘Sur la réalité des propriétés dispositionnelles’. 15 B.D. Ellis and C.E. Lierse, ‘Dispositional Essentialism’, Australasian Journal of Philosophy, 72 (1994): 27-45; B.D. Ellis, ‘Natural Kinds and Natural Kind Reasoning’, in P. Riggs (ed.), Natural Kinds, Laws of Nature and Scientific Methodology (Dordrecht, 1996), pp. 11-28; ‘Causal Powers and the Laws of Nature’, in H. Sankey (ed.), Causation and Laws of Nature (Dordrecht, 1999), pp. 21-42; ‘Bigelow’s Worries about Scientific Essentialism’, in H. Sankey, ed. Causation and Laws of Nature pp. 77-97; ‘Response to D. Armstrong’, in H. Sanky ed., Causation and Laws of Nature, pp. 49-55; Scientific Essentialism (Cambridge, 2001); and for a good summary of his position, see ‘The New Essentialism and the Scientific Image of Mankind (NESIM); see also The Philosophy of Nature: A Guide to the New Essentialism, Acumen, 2002. 16 Cf. NESIM, 3 ‘Scientific essentialism’.

Dispositions and Essences

83

categories. There is a pre-existing grid of objective categories, and it is the aim of natural science to reveal and describe them’.17 Scientific essentialism is concerned with natural kinds which range over events and processes as well as with the more traditional sorts which range only over objects or substances.18 The most general laws of nature describe the essential properties of all these global kinds,19 and therefore hold necessarily of all objects, events and processes: ‘The law of conservation of energy, for example, states that every event or process of this global kind is one that is intrinsically conservative of energy. Hence, any event which was not intrinsically conservative of energy could not be one of a kind that would occur in our world. The laws we think of as causal laws are generally more specific in their direct application. The laws of electromagnetism, for example, apply directly to all electromagnetic radiation, and hold necessarily of all such radiation. Therefore, if there is any radiation which is not propagated according to these laws, it cannot possibly be electromagnetic’.20 2. Secondly, scientific essentialism claims that ‘the essential properties of the most fundamental kinds of things are not just the passive primary qualities of classical mechanism, but also include a number of causal powers, capacities and propensities – powers to act, and powers to interact’,21 namely properties which are essentially dispositional in nature, implying dispositions to act and react in various ways, depending on the circumstances. Traditionally, such properties have been thought to be ontologically dependent on underlying categorical (i.e. non dispositional) properties and on the laws of nature. According to the new essentialism, however, at least some of these dispositional properties are fundamental, and not dependent upon any other properties. In other words

17 NESIM, 3. 18 Ibid., and Ellis, Scientific Essentialism, p. 3. 19 For Ellis this means that the natural kinds in these two categories occur in natural hierarchies. ‘The most general kind in the category of substances includes every other natural kind of substance existing in the world. This is the global kind, for our world, in the category of substances. It defines the range of natural kinds of objects or substances that can exist in our world. The most general kind in the category of events is the global kind, which includes every other natural kind of event or process which occurs in the world. This global kind effectively defines the range of kinds of events or processes that can occur in our world’. B. Ellis, Scientific Essentialism, p. 3. 20 NESIM, 3 and B. Ellis, Scientific Essentialism, p. 6. 21 NESIM, 3 and B. Ellis, Scientific Essentialism, pp. 106-144. This is what D. Armstrong objects to. See his ‘Reply to Ellis’, in H. Sankey ed., Causation and Laws of Nature, pp. 43-48. He argues that such intrinsic causal powers are Meinongian properties, and thus objectionable. In principle, he says, there could be causal powers which happen never to be exercised. If such properties have no categorical bases, as Ellis allows is possible, then such powers can be defined only by relationships between non-existent objects, i.e. between the kind of circumstances which would trigger them and the kind of display which would then result. For Ellis’s answer see Scientific Essentialism, pp. 133-135.

84

Dispositions and Causal Powers

(one will note the Aristotelian and/or Leibnizian flavor), the basic things in the world are essentially active and dynamic. ‘They are not just passive objects obeying blindly the commands of God, as most seventeenth and eighteenth Century philosophers believed, a world of things having only the attributes of extension and impenetrability, as Descartes’s and Locke’s worlds were; rather it is a world in which things have their own internal dynamics, which are essential to their natures, and which are determinative of their behaviour’.22 In that respect, ‘all things are essentially active and reactive: at the most basic level, what they are intrinsically disposed to do is what makes them the kinds of things they are. Things of given kinds must always be disposed to behave in certain kinds of ways, just by virtue of being things of these kinds. Their identities as members of these kinds depend on their being so disposed to act’.23 Hence, ‘the real defining characteristics of the most fundamental kinds of things that we know about, namely, things like protons and electrons, would all appear to lie in their laws of interaction. Things of these kinds would appear to have no real defining characteristics at all apart from their causal roles. A proton, for example, might be defined (by way of real definition) as any particle which behaves as protons do. For no proton could possibly fail to behave in these ways, and no particle other than a proton could possibly imitate this behavior. Its identity, qua proton, might thus be defined by its causal role. Similarly, one might say that an electron is, by real definition, any particle for which the laws of interaction are precisely those of electrons. Nothing other than an electron could possibly behave in such a way, and whatever does behave in this way has to be an electron’.24 3. Hence a third thesis: the laws concerning the behavior of protons, and their interactions cannot be just accidental, i.e. laws which could well have been otherwise. On the contrary, it is essential to the nature of the proton that it is disposed to interact with things of various kinds precisely as it does. The proton’s causal powers, capacities and propensities are not just amongst the accidental properties of protons, which depend on what the laws of nature happen to be, but amongst their essential properties, without which there would be no protons, and which protons could not lose without ceasing to exist.25 Accordingly, ‘the traditional view that the laws of particle physics are imposed on intrinsically passive things which have kind identities which are independent of the

22 NESIM, 3. Ellis claims that it is this second tenet of scientific essentialism which sets it apart most strongly from other theories of the nature of reality. 23 Ibid. 24 NESIM, 4. 25 B. Ellis, Scientific Essentialism, p. 2.

Dispositions and Essences

85

laws of their behaviour is thus implausible from the point of view of modern science. Essentialism is a much more plausible position to take’.26 As Brian Ellis observes, the claim that the world is structured into hierarchies of natural kinds of objects and so on, could in principle be accepted by philosophers who were otherwise sympathetic to mechanism:27 things of different natural kinds, they might say, are just things made up of different basic ingredients, or of the same ingredients, but put together in different ways, although there is nothing in their natures which requires that they should behave in one way rather than another. How they are disposed to behave depends on what the laws of nature happen to be. Now it is precisely such a contingency thesis or Humean view of the laws of nature that Ellis denies.28 First, the relationship of dependence between the causal powers and the laws of nature is the other way round: the laws depend on the properties, not the properties on the laws; second implication: the laws of nature are metaphysically necessary and hence true in all possible worlds.29 That is, ‘it must be metaphysically impossible for things, constituted as they are, to behave other than in accordance with the laws of nature. Even God (assuming Him to exist and be all powerful) couldn’t make them behave contrary to their natures. He might change their natures, perhaps, so that they might become, or be replaced by, different kinds of things. But there is no possible world in which things, constituted as they are, could behave any differently. For them to behave differently, they would have to be or become things of different kinds, or be made up of things of different kinds’.30 Hence laws of nature are both descriptive and prescriptive: ‘What science observes and codifies are the manifestations of these dispositions. Hence, laws which describe how dispositional properties act will, at the same time, tell us what things which have these properties essentially must do in virtue of being the kinds of things they are’.31 The natural dispositions are the truthmakers for these laws.

26 NESIM, 4. 27 B. Ellis, Scientific Essentialism, p. 107 ff. 28 Ibid., p. 4, p. 7, and p. 44 ff, p. 211 ff., p. 229 ff., p. 261 ff. 29 Ibid., part four, chs 6 and 7, p. 203 ff. 30 ‘The most elementary kinds of things all have fixed causal powers, i.e. their dispositional properties are all fixed by their essential natures. A copper atom, for example, has the same dispositional properties wherever or whenever it might occur. The same is true of a proton or an electron. They are things which belong to what might be termed “fixed natural kinds”. Their distinguishing feature is that you cannot change any of their dispositional properties. They do what things of these kinds always do, and you cannot teach them any new tricks. There can be no question of a copper atom, for example, being disposed to behave in one way at one time, but in a different way at another time. Nothing with such variable powers could possibly be a copper atom’. NESIM, 5. 31 B. Ellis and C. Lierse, ‘Dispositional Essentialism’, p. 40.

86

Dispositions and Causal Powers

Mumford’s Dispositional Essentialism Even if Mumford characterizes his own position as being mainly a functionalist32 view of dispositions coupled with a neutral monistic ontology, at several places, especially in his comment on Ellis’s and Lierse’s paper33 as well as in the last chapter of his book Dispositions,34 he admits that he endorses dispositional essentialism for certain basic kinds and agrees with many points raised by Ellis and Lierse: only, he takes them to task for the ontology they deem necessary for their position. So, let me try to spell out what are the main arguments Mumford advances in favor of such a dispositional essentialism: 1. First of all, Mumford thinks that ‘the explanatory value of appeal to dispositions typically resides in them being causally efficacious and being properties’.35 It means that they cannot be reduced, for example, to a different class of entities such as events or occurrences, and that their criterion of existence is the causal criterion which can be defined in the following way: ‘The causal criterion of property existence: for any intrinsic abstract property P, P exists if and only if there are circumstances C in which the instantiations of P have causal consequences’.36 2. Ellis and Lierse, at least in their 1994 paper, were very clear that they did not seek to ‘restore the reputation of dispositions by attacking the status of categorical properties, or by arguing that all properties are basically dispositional, as some philosophers have done’,37 but mainly to show that ‘their lowly status was a 32 I have shown in ‘Sur la réalité des propriétés dispositionnelles’, the difficulties raised by a ‘functionalist’ account of dispositions. On the problems linked with ‘functions’, see also J. Searle, The Construction of Social Reality (New York, 1995), p. 29 ff. 33 S. Mumford, ‘Ellis and Lierse on Dispositional Essentialism’, Australasian Journal of Philosophy, 73 (1995): 606-612. 34 S. Mumford, Dispositions (Oxford, 1998), p. 234 ff. 35 Ibid., p. 14. 36 Ibid., p. 122. This means, in particular, that mere ‘Cambridge changes’ are not allowed. S. Shoemaker also agrees on that criterion; see his ‘Causality and Properties’, in P. Van Inwagen (ed.), Time and Cause (Dordrecht, 1980), pp. 109-135, repr. in D.H. Mellor and A. Oliver (eds), Properties (Oxford: 1997), pp. 228-254, p. 232. cf. Evan Fales in Causation and Universals, p. xiii: ‘A physical universal exists if and only if it is a member of the causal “web”; if it is then it exists whether or not it is ever instantiated’, quoted by S. Mumford, Dispositions, p. 123. Let us note that some philosophers do not accept the idea that the causal criterion should be a criterion for the existence of a property. For example, A. Oliver, ‘The Metaphysics of Properties’, Mind, 105 (1996): 1-80, p. 8. 37 Ellis and Lierse, ‘Dispositional Essentialism’, p. 27. The philosophers they cite are R. Harré and E.H. Madden, as well as Evan Fales, who ‘has argued that the essential properties of the most fundamental natural kinds are their monadic properties, and, in so far as these properties are dispositional, things of these kinds must behave as these properties prescribe’, (E. Fales, ‘Essentialism and the Elementary Constituents of Matter’, Midwest Studies in Philosophy, 11 (1986): 391-402; Causation and Universals (London, 1990). Nevertheless,

Dispositions and Essences

87

consequence of their traditional affiliation with an inadequate ontology based on a Humean metaphysic, and a flawed semantics of dispositional terms’ (ibid.). So, what they sought to provide was ‘a more adequate semantics, and an ontologically more satisfactory theory of dispositions – one which allows at least some dispositions to be counted as genuine properties existing in their own right’ (ibid.). Mumford, together with most philosophers trying to rescue dispositions, agrees with the first part of the program: a new semantics for dispositions is needed, which in particular should provide an adequate analysis of the relation (which is not one of equivalence, but of irreducibility38), between disposition ascriptions and subjunctive conditionals. 3. Again, as Ellis and Lierse, and unlike, for example, Hugh Mellor, who thinks that such properties as triangularity which at first sight look categorical, may be viewed as dispositional,39 Mumford does not so much attack the CategoricalDispositional distinction as he tries (1) to explain why it is relevant only as a conceptual (though not a real)40 distinction between ascriptions (ch. 4), (2) Ellis and Lierse admit that ‘in the case of the fundamental particles, the dispositional properties may indeed be all that we have’, (Ellis and Lierse, ‘Dispositional Essentialism’ p. 29). Mumford agrees with this. Not only does he admit the possibility of ‘ungrounded dispositions’, but his functional analysis makes him give importance to such functionalist systems as ‘thermostats’ or thermometers whose essence is clearly entirely dispositional (ibid., p 8), i.e. that ‘the exact mechanisms that equip a thermostat with the dispositions it has are, in a significant way, irrelevant to the fact that it is a thermostat. Roughly, for something to be a thermostat, it must be sensitive to changes in temperature and be able to trigger a switch if a pre-calibrated temperature threshold is crossed’ (ibid.). 38 S. Mumford, Dispositions, p. 63. See chs 3 and 4 for more details: ‘What I suggest is the rejection of a solely conditional analysis of dispositions and that we treat them as real instantiations of properties which afford possibilities rather than just being shorthand ways of talking about certain combinations of events’. 39 D.H. Mellor, ‘Counting Corners Correctly’, p. 97. Ellis and Lierse note that Mellor’s dispositional foundationalism, which purports to be Humean, faces difficulties: ‘By embracing the Contingency Thesis, Mellor, at the very least, has to subscribe to the view that the laws in this world are only contingently related to the entities that exist in it. Thus, the behaviour of an entity must be logically distinct from the kind of thing it is. Now, if Mellor wants to 1) deny the existence of categorical properties, 2) avoid embracing a behaviourist theory of dispositions, and 3) have an ontology of occurrent properties, then his only option is to identify a disposition with some occurrent property or properties which are not dispositional. But what could these be? Could they be, perhaps, those structural or other properties which must exist if the disposition is to exist?’ (Ellis and Lierse, ‘Dispositional Essentialism’, p. 35). 40 He is very careful to distinguish the two levels. See for ex. p. 145: ‘The thesis of property dualism tries to take this distinction (i.e. C-D) further into ontology. Property Monism, on the other hand, is the claim that the conceptual distinction is the only distinction because it denies that there is an ontological division in reality which the conceptual division maps’. ‘There are not distinct categorical and dispositional types of properties but there are distinct categorical and dispositional ways of talking about instantiated properties’.

88

Dispositions and Causal Powers

to clarify the confusions attached to the concept ‘categorical’,41 (3) to fix some criterion for the distinction, namely the following: both dispositional and categorical properties have causal roles. But ‘disposition ascriptions are ascriptions of properties that occupy a particular functional role as a matter of conceptual necessity and have particular shape or structure characterizations only a posteriori’ whereas ‘categorical ascriptions are ascriptions of shapes and structures which have particular functional roles only a posteriori’; In other words, ‘our knowledge of the causal roles of the various categorical properties is a complex matter of scientific investigation’.42 Indeed, ‘what is crucial to property monism (the position Mumford adopts) is a justification of the claim that dispositions and their categorical bases are actually the same states of particular things, characterized in two different ways, rather than distinct states in the world’.43 Which, incidentally means that ‘the identity conditions for properties are separated from the identity conditions for predicates’.44 Hence ‘I can know what it is for x to be soluble though I need not know what it is for x to have the particular molecular structure which is a categorical base of solubility, say, xyz. I need not know which molecular structure is involved or even what a molecular structure is, even though “x is soluble” may be true in virtue of the possession of the same state that makes “x has molecular structure xyz” true’. 45 4. Again, another argument by which Mumford seems to agree with Ellis’s version of dispositional essentialism is his rejection of the Continuing existence argument raised by categorical realists to justify the need of categorical bases for dispositions. According to the continuing existence argument, dispositions have existence in between their manifestations only if they are grounded in categorical properties. Indeed, as Ellis and Lierse hold, dispositions need to be based in reality: ‘They must at least be properties of real things. Moreover, it is often the case that things have the dispositions they do because of their internal structures; and in all such cases, we may say that the dispositions are grounded in these structures. However, it is not clear that the basis of any given disposition must, or must, ultimately be non-dispositional. For, without begging the question against non-Humeans, it cannot be assumed that the basis of a disposition does not, or does not ultimately, include other dispositions. For example, the dispositions of an object might well depend on 41 ‘In particular, if “categorical” means “unconditional”, dispositions are, in a very obvious way, categorical: when I say that a particular sugar-cube is soluble, I am in no way making an ascription conditionally for I am saying that it is actually soluble now, not that it could be soluble if some other conditions obtained’, p. 64. 42 S. Mumford, Dispositions, p. 77. 43 Ibid., p. 145. 44 S. Mumford, Dispositions, p. 146. At the same time, it implies also, as Armstrong puts it, ‘the emancipation of the theory of universals from the theory of semantics’, D.M. Armstrong, A Theory of Universals (Cambridge, 1978), p. 6 45 S. Mumford, Dispositions, p. 148.

Dispositions and Essences

89

the causal powers of its parts, as well as on how these parts are arranged’.46 So a power could well be grounded in a further power: not all powers need categorical properties as a basis.47 Nevertheless, Mumford is very cautious to distinguish ‘ordinary’ from ‘ultimate’ dispositions. First, because ‘most dispositions are of the ordinary variety’,48 so that ultimate dispositions are ‘limited’ (examples of which would be the dispositions of subatomic particles which apparently have no internal constitution that could explain their behavior). Second, because the history of science shows that mistakes are possible in this respect: ‘We cannot be sure, for any putative ungrounded disposition, whether it is genuinely ungrounded or actually just an ordinary disposition and only thought to be ultimate because of the incomplete state of our physics’.49 Nonetheless, in terms of explanatory requirements and sound methodology we cannot either rule out their existence a priori.50 Even more: ‘The nature of explanation is such that ungrounded dispositions will always have to be posited in order to avoid a regress of explanation and, further, we have every reason, as part of the atomist strategy, to assume that there are genuine ungrounded dispositions. So the function of ungrounded dispositions is to “fill the space” of our process of explanation’ (ibid.). 5. Another important point of agreement in both versions of essential dispositionalism has to do with the indeterminate51 character of most

46 Ellis and Lierse, ‘Dispositional Essentialism’, p. 31. 47 The most detailed analysis of such questions is to be found in G. Molnar, Powers: A Study in Metaphysics, ed. S. Mumford (Oxford, 2003) especially chapter 8. 48 S. Mumford, Dispositions, p. 229. 49 Ibid., p. 231. 50 Ibid., p. 233. 51 In a deterministic world, this indeterminacy is due to linguistic vagueness. Hence, ‘To define fragility, for example, we cannot do much better than say that a fragile object is likely to break if dropped, or otherwise handled roughly. Any more precise definition might capture some more specific concept of fragility. But it would not be the broad but vague concept with which we are familiar’. On the contrary, in an indeterministic world such as ours, ‘the dispositions are causally indeterminate, not because of vagueness, but because of the indeterminacy of the underlying physical processes. For example, the probability that a radium atom existing at t will have decayed by t + ∂ is, for any given frame of reference, a precisely specifiable function of ∂, and this probability is independent of the circumstances in which the radium atom exists. Hence, we cannot, even in principle, eliminate this causally indeterminate disposition in favour of any more precisely defined dispositions which are causally determinate’ (Ellis and Lierse, ‘Dispositional Essentialism’, p. 42).

90

Dispositions and Causal Powers

dispositions, either causal52 or stochastic.53 6. But most of all, Mumford agrees with Ellis not only in letting disposition ascriptions play an explanatory role and in conceiving dispositions as units of explanation, but in favoring a dispositionalist account of laws, instead of what he labels the ‘laws view’, namely, the view that laws, conceived as constant conjunctions or natural necessitations of events, are the basic building blocks of explanation.54 Instead, ‘the dispositionalist attempts to dispense with laws of nature by grounding natural necessities within the individual instantiated properties themselves. This gives us the model of instantiated properties as real powers to do things. It is also useful to speak of them as enablements or affordances where this conveys the dispositional notion of causal mediation between events. Thus we have natural necessities without commitment to general laws ranging over classes of events. Instead of general necessities, natural necessity occurs at the level of the particular. It is not in virtue of a general law that sugar dissolves when in a liquid, for instance, it is in virtue of a particular state or instantiated property possessed by that sample that it does so’.55 In various places, Mumford has, quite convincingly, outlined the advantages offered by such a dispositionalist account of laws over both the Humean supervenience view as most strongly defended by David Lewis56 and the Dretske57-Tooley58-Armstrong59 (DTA) view of laws, which, despite its challenging the Humean view by taking laws to be real things in nature involving necessary connections between properties, still retained some vestiges of the Humean metaphysic.

52 ‘Where we have a causal disposition, there is typically a certain pattern of causeand-effect or stimulus-and-response which anything having the disposition would normally display if it were appropriately caused or stimulated to do so’ (Ellis and Lierse, ‘Dispositional Essentialism’, p. 40). 53 ‘A Stochastic disposition is a propensity of some kind, in which the antecedent condition is not strictly the cause of its manifestation, but only a necessary condition for it.’ For example, the disposition of a radium atom to decay in a certain way is a stochastic disposition. If this species of radioactive decay is to occur, it is a necessary condition that radium atoms should exist. But events of radioactive decay are not caused by the existence of such atoms. Not, as far as they know, are they caused by anything else. There is just a certain objective probability p that within any given-time interval ∂ such an event will occur’ (ibid., p. 40). 54 S. Mumford, Dispositions, p. 217. 55 Ibid., p. 221. 56 D. Lewis, Counterfactuals (Oxford, 1973) and ‘Humean Supervenience Debugged’, Mind (1994). 57 F. Dretske, ‘Laws of Nature’, Philosophy of Science, 44 (1977): 248-268. 58 M. Tooley, ‘The Nature of Laws’, Canadian Journal of Philosophy, 7 (1977): 667698; Causation: a Realist Approach (Oxford, 1987). 59 D.M. Armstrong, What is a Law of Nature? (Cambridge, 1983) and A World of States of Affairs (Cambridge, 1997), chs 15 and 16.

Dispositions and Essences

91

Now, what are indeed the advantages of seeing laws to be determined by dispositions instead of dispositions being determined by laws? The problems with some laws (I do not deal here with general a posteriori identity statements such as ‘water is H20’ or ‘heat = mean molecular motion’60 but with laws such as ‘all water boils at one hundred degrees centigrade’ or ‘all ducks have webbed feet’) are well known: the identification of them with universally quantified conditional statements is doomed to failure. The main challenge to regularity accounts of law is to distinguish coincidental from genuinely lawlike regularities61 and, secondly, virtually no statements of this form are true.62 As Nancy Cartwright has argued, laws of nature qua general statements would be, for the most part, idealizations and abstractions that should not be regarded as literally true. And the ceteris paribus strategy, in her view, almost amounts to triviality. Getting away from the notion of the statements of laws being general truths, the DTA or ‘nomic necessitation’ version, by construing them as causal relations holding between universals constitutes a progress insofar as it allows realist causal claims that go beyond empirical evidence.63 But the nomic necessitation view does not explain how nomic necessitation entails the corresponding generalizations – it has to stipulate that it does. Moreover, at least in the case of Armstrong’s conception, according to Mumford, although it deals with exception cases (the ‘oaken’ or defeasible laws as opposed to the ‘iron’ laws), it too, like Hume’s and Lewis’s accounts, is ‘committed to reading off the laws from the actual history of events at a world’64 and is caught up in some form of actualism.65

60 I have analyzed the difficulties raised by such statement in ‘La métaphysique et l’analyse conceptuelle’, Revue de Métaphysique et de Morale, numéro spécial: Métaphysique et ontologie: perspectives contemporaines, no. 4, (2002): 559-584, pp. 575-581. 61 See A. Bird, ‘The Dispositionalist Conception of Laws’, Foundations of Science (forthcoming) p. 2. (Preprint available on A. Bird’s web page). The best of these is the systematic regularity theory of Lewis (following from Ramsey), Counterfactuals, pp. 7277: ‘According to this view, a generalization states a law only if it is deducible from that axiomatic systematization of the facts that optimally combines strength and simplicity’. But, as Bird notes, ‘While this reduces the pressure of the objection from accidental regularities, it does not remove it altogether. For we could imagine a system of laws that was itself rather complicated and weak but which generated an accidental regularity, whose addition to the axiomatic system might in fact add considerably to its strength without detracting much from its overall simplicity’, p. 2. This objection (and others addressed to the regularity view) is also to be found in Armstrong, What is a Law of Nature?, pp. 66-73. 62 S. Mumford, Dispositions, p. 122 ff. 63 When the relation holds between two universals, it entails the corresponding generalization but the reverse entailment does not. 64 S. Mumford, ‘L’état des lois’, in J.-M. Monnoyer (ed.), La Structure du Monde (Paris, 2004), pp. 235-262, p. 235. 65 Ibid., p. 256, i.e. ‘the only existence of the law is in what particular causal sequences occur, so the laws are determined by the events. But then comes a twist. Singular causation

92

Dispositions and Causal Powers

Mumford thinks that a right account of laws must reside somewhere ‘twixt regularity and regulation’, namely between the descriptive regularist view and the theological view concentrating on regulation with laws as God’s prescriptions. In order to qualify as the right interpretation of laws, both somewhat descriptive and prescriptive, viewing these as that ‘which bind the possible and necessary but not the actual in the natural world at which they apply’ (ibid.), a dispositionalist standpoint seems the most appropriate. For example, instead of saying ‘all ravens are black’, one may drop the requirement that the conditionals associated with law statements must be universally quantified and say instead ‘ravens are black’ without quantification.66 Now there is no need for quantification if we understand such a statement as being not explicitly about a numerable quantity of ravens but as being about the normal dispositional coloring of the species. According to Mumford, such a view fits with scientific practice: namely, the scientist finds a typical representative of some species or kind of thing and then, in carefully controlled conditions, finds out some natural dispositional property of that thing. He is not concerned with amassing more and more cases of black ravens, nor will he be troubled by the existence of strange cases, such as an albino.67 Thus, such a notion of ‘the natural behaviour of things’ supplies us with ‘as much regulation as we need for our laws, it gives us a sense of objects having natural governing principles’.68 Indeed, Mumford contends, ‘because of the dispositional force of such governing principles, these dispositions need not be manifested: whether or not they are manifested depends on the contingencies of the world’s history; for instance, on whether the right stimulus conditions are ever realized and whether the correct background conditions are in place’.69 In the second place, ‘what the laws are best understood as descriptive of is not actual or possible events but the capacities of things. Thus, the truth of the claim that a fair coin has a 50/50 chance of landing a head or a tail is true not in virtue of the result of any toss or sequence of tosses, but in virtue of the 50/50 propensity the coin has to land head or tail… The probabilistic law can thus be explained by it being descriptive of an was only a ladder enabling us to ascend to the idea of nomicity, which once attained, the ladder can be thrown away. The fundamental causal relation is a nomic one, holding between universals [as Armstrong says in World of States of Affairs: ‘Singular causation is no more than the instantiation of this type of relation in particular cases’, p. 227]. This phrase suggests to me Platonism: causation not between particular instantiations but between universals. Again though, we know that for Armstrong, a universal is nothing more substantial than the sum of its instantiations, so I still do not think that Armstrong escapes actualism. Causation holds between universals, but it can only do so if it has instantiations standing in the right relation’. 66 S. Mumford, ‘L’état des lois’, p. 255. 67 Ibid., pp. 238-239, p. 255; cf. C.B. Martin, ‘Dispositions and Conditionals’, The Philosophical Quarterly, 44 (1994): 1-8. Note that C.S. Peirce makes a similar remark: ‘what the chemist is interested in is not the particular sample, but the molecular structure’, The Collected Papers of C.S. Peirce, 8 vols, (Cambridge, 1939-1958), vol. 4, § 530. 68 S. Mumford, ‘L’état des lois’, p. 258. 69 Ibid.

Dispositions and Essences

93

actual disposition and the account also explains the possibility of unrealized laws. The disposition ascription is true of the coin if it is never tested and thus there can be two true but different ascriptions in two worlds that coincide in all their events’.70 In the third place, ‘there is plenty of logical contingency in these laws: I need not even appeal to another possible world to illustrate this: the laws of nature could change, gradually or instantaneously, on my account... The laws are fixed only by what things (can) do at present. Modalities can be tensed. Thus ravens are black now, meaning that it is true of the raven kind that its normal members are dispositionally black, but tomorrow, through a divine intervention, normal ravens may become dispositionally pink. More naturally, over a period of time, normal ravens may become grey through evolutionary processes... My account leaves open the possibility that even the fundamental laws could change…’71 Finally, such a theory meets the ‘desideratum of placing laws somewhere “twixt regularity and regulation”. There is a sense in which laws are about governing principles – that should naturally be the case – but these governing principles are not wholly detached from the world. They are enshrined in the properties natural for class, kind or sort members, hence this can still be called a descriptivist theory of laws. The laws are descriptions of what is natural for classes of particulars but they are not descriptions of events, rather descriptions of what things can do, with a dispositional sense of can’.72 So far so good. But are such arguments wholly convincing? Some Suggestions about the Main Difficulties Dispositional Realism Should be Ready to Face Mumford as well as Ellis or other proponents of the redemption of dispositions intend to confer them a real and not only a fictional status. But what kind of reality have dispositions actually gained? More precisely, are their respective versions of dispositional essentialism or functionalism realistic enough? There are a lot of convincing aspects in Mumford’s account but some are also somewhat puzzling. I shall take them in turn. The Merits of Ontological Deflation or ‘Neutrality’: From Properties to Predicates. It is surely an interesting part of the neutral monistic understanding of dispositions to show how the dispositionality or non dispositionality of a property or state is not a question which is (mainly) grounded in ontology, but is, rather, a matter relative to the predicate with which the property or state is denoted and to the connection between this predicate and the behavior which is mediated;73 or again that ‘it is 70 S. Mumford, Dispositions, p. 231. ‘Given the realist understanding of such dispositions, this would be something more than a counterfactual difference’. 71 S. Mumford, ‘L’état des lois’, p. 259. 72 Ibid., p. 260. 73 S. Mumford, Dispositions, p. 210.

94

Dispositions and Causal Powers

not a truth of ontology whether a property is dispositional or not; it can only be a truth about concepts’.74 All the same, as both Ellis and Mumford have observed, such a conceptual reading of the issue does not prevent you from being a realist, as Peirce once also noted, who was one of the first philosophers to attach a particular importance to the reality of dispositions: ‘Anybody may happen to opine that “the” is a real English word; but that will not constitute him a realist. But if he finds that, whether the word ‘hard’ itself be real or not, the property, the character, the predicate hardness is not invented by men, as the word is, but is really and truly in the hard things and is one in them all, as a description of habit, or behaviour, then, he is a realist’.75 So indeed, any correct account of the role and nature of dispositions should start with a semantic account of dispositional predicates, as Hugh Mellor also reminded us,76 and, in particular, should try to deal with the reasons why the reduction of dispositional ascriptions to conditionals does not seem to work (for example, due to finkish or antidote cases), or why reduction sentences may or may not tell us ‘all’ that dispositional predicates mean. Suppose, for example, that we know all the manifestations of courage: are we going to be able to say what ‘courageous’ means by summing up all the reduction sentences that say what those manifestations are? Or course not: as Peirce insisted on, a disposition is, by essence, irreducibly general and indeterminate, and cannot be reduced to the mere conjunction of its occurrences. Dispositions as Properties or How should one Understand the CategoricalDispositional Distinction? But, as Mellor, Ellis or Mumford also insist on, properties are not (or are not given) simply by the meaning of our predicates.77 Nobody would ever think that the planet Mars is, or is part of, or is defined by the meaning of the ‘word’ Mars which we use to refer to it. On the contrary, what gives consistency to the referential capacity of our predicate, when we use it, is that it takes for granted the existence and the identity of planet Mars. In other words, we want a conditional and non verifunctional statement such as ‘if x was dropped, it would break’ to have a truthmaker. This is why Mellor, following Armstrong,78 rejects the Rylean and counterintuitive view according to which when a fragile thing a and a non fragile thing b are not dropped, there is no factual difference between them.79 On the contrary, one should say that if a is fragile and b is not, then a and b, whether dropped or not, must differ in some

74 Ibid., p. 213. 75 C.S. Peirce, Collected Papers, vol. 1, § 27, note 1. 76 D.H. Mellor, ‘The Semantics and Ontology of Dispositions’, pp. 757-779. 77 D.H. Mellor, Matters of Metaphysics (Cambridge, 1991), p. 171. See also B. Ellis, Scientific Essentialism, p. 18, note 1. 78 D. Armstrong, A World of States of Affairs, ch. 6.6. 79 G. Ryle, The Concept of Mind, ch. 5.

Dispositions and Essences

95

respect, the most obvious difference consisting of the fact that a has the property of being fragile, which b has not. But how should one explain what that property consists of? To a certain extent, Mumford refuses to commit himself too far. ‘The functionalist theory of dispositions states that nature consists of hierarchically ordered levels where at each level there are entities or properties that can be characterized either dispositionally or structurally and whether they are dispositional or structural characterizations depends on their explanatory relations to other entities or properties within the whole of which they are a part.’ Hence: Relative to the functional role of causing dissolving when in liquid, a denotation of a property P that has that causal role by conceptual necessity – solubility – is dispositional. Denoted in such a way that does not conceptually necessitate that causal role, perhaps in terms of molecular structure, that same property comes out as non-dispositional. By the argument from the identity of causal role, however, these two denotations are of the same state or property instance.80

In a similar vein, Mellor suggests that properties are neither categorical nor dispositional: they just are. But this is a bit unsatisfactory. To a certain extent, Avicenna had a similar formula when he said that equinitas est equinitas tantum.81 Now everybody knows that the whole problem of universals started from the questions raised by grounding the objective indeterminacy of such a ‘tantum’. It seems too easy, or a way to have one’s cake and eat it, to deny a real distinction between the Categorical and the Dispositional while agreeing that a conceptual distinction is needed, meaning by this that it must have some objective foundation, or that it is more than a distinction of reason, namely, more like what Duns Scotus would have named a formal distinction, that is ‘a distinction from the nature of the thing occurring between two or more really identical formalities of which one, before the operation of the intellect, is conceivable without the others though inseparable from them even by divine

80 S. Mumford, Dispositions, p. 215. 81 Crucial to Avicenna’s position is the view not so much that essence, as such, may be considered under two headings, in things and in the intellect, as the fact that it may be viewed as such, in its pure essentiality, neither universal nor singular. The essence (which Duns Scotus, following Avicenna, is to name ‘The Common Nature’ is characterized by such neutrality or indifference towards all its possible determinations, which allows, in particular, to conceive it apart, as a distinct object of thought: ‘ipsa equinitas non est aliquid nisi equinitas tantum’ (Avicenna, Liber de Philosophia prime sive Scientia divina, ed. S. Van Riet, Louvain, E. Peeters - Leiden, E.J. Brill, I-IV (1977), V-X (1980), Book 5, chap. 1 (fol. 86a), pp. 227229. For more details, cf. C. Tiercelin, ‘Le problème des universaux: aperçus historiques et perspectives contemporaines’, in J.-M. Monnoyer (ed.), La Structure du monde (Paris, 2004), pp. 329-353, p. 335ff.

Dispositions and Causal Powers

96 82

power’. One way or other, some deep metaphysical issues will have to be handled here with great care. Substantival or Relational Realism? Again, the ‘neutral’ monism is in danger of committing itself to more ontology than it would surely be prepared to, at least in the following respect: In Hume’s ontology, there is a sharp distinction between the way a thing is, and how it is disposed to behave: the way it is depends on what its properties are. How it is disposed to behave depends on what the laws of nature are. The two are supposed to be independent of each other.83 Mumford’s general view has the great advantage of trying to account for the changes that occur in the world, from the inside: as he somewhere says, a world described entirely in categorical terms is a static world, and hence, not ours.84 Again, the familiar Humean point that it is a logically contingent matter what effect follows from any antecedent event, and that laws of nature are what is added to animate, so to speak, the world, has something unconvincing about it. This is precisely one of the confusions which Peirce diagnosed in the scholastics’ treatment of dispositions. As is well known, the scholastics (with the maxim operari sequitur esse) claimed that the operations or activities of a thing reflected a corresponding perfection in that thing, which could account for its ability to perform that sort of operation. It was on the basis of this principle that they were able to specify natures and natural classes. If a man performs an action that other animals cannot, such as reasoning, then the man has a power that other animals do not possess. If such a capacity is the distinguishing characteristic, it is dignified with the title ‘nature’. But why did Peirce finally object to the scholastic operari sequitur esse, i.e. to the substantial form as a dispositional character – aside from the question of how scientific one is in determining distinguishing characters?85 Because it failed to reveal the relational structure which is ultimately involved. That is, the scholastics were right, as far as they went, but their limited logic did not allow them to see that the nature, power, or disposition represented in these monadic predicates was only a truncated image of a relational law. Apparently, it was the old logic’s inability

82 M.J. Grajewski, The Formal Distinction of Duns Scotus (Washington, 1944), p. 93. Such is the best definition given of this difficult Scotistic notion, mid-way between a real distinction and a distinction of reason, through which Duns Scot thinks he was able to 1) avoid the difficulties raised both by extreme (or Platonist) realism and nominalism, 2) understand how individuals (haecceities) can participate to the Common Nature (by contraction) without being reduced by it, 3) ensure the objectivity of our concepts which remain founded in things and not mere flatus vocis. But this notion of the Subtile Doctor has been severely attacked (by Occam in particular). 83 Cf. Ellis and Lierse, ‘Dispositional Essentialism’, p. 28. 84 S. Mumford, Dispositions, p. 214. 85 C.S. Peirce, Collected Papers, vol. 6, § 361.

Dispositions and Essences

97 86

to handle abstractions properly that was mainly at fault. The scholastics did not realize that their most important abstractions were really hypostatized relations, and therefore that real abstractions indicated real relations – laws and not forms. I am not sure whether Peirce’s substitution of a relational realism to a substantival realism is necessarily the correct or only way of handling the problem of dispositions, but at any rate, it is true that there is, in Mumford, a very unelaborated commitment to a view of what an object is in terms of a plain substance, and not much elaboration either on the ‘something’ which justifies or constitues the truthmaker for a dispositional ascription: is it mere epistemic indeterminacy? Or is this indeterminacy linked with some irreducible, metaphysical features of, for instance, some Common Nature?87 Again, Mumford is obviouly committed to a trope theory view of instantiated particulars but neither does he provide an account of the process of instantiation or individuation nor does he present a justification of the reasons why he should rather favor such a view instead of another,88 while admitting that there are notorious difficulties, raised by Russell for example, for a tropist account.

86 C.S. Peirce, Collected Papers, vol. 3, § 642. I have analyzed this in ‘L’influence scotiste dans le projet peircien d’une métaphysique scientifique’, Revue des Sciences Philosophiques et Théologiques, numéro spécial: Jean Duns Scot et la métaphysique classique, no. 1 (1999): 117-134, pp. 128-129. 87 See notes 74 and 75. 88 He rejects E.W. Prior’s contention that some identification of properties as universals is needed. (cf. E.W. Prior, ‘The Dispositional/Categorical Distinction’, Analysis, 42 (1982): 93-96; Dispositions (Aberdeen, 1985), p. 76; E.W. Prior, R. Pargetter and F. Jackson, ‘Three Theses about Dispositions’, American Philosophical Quarterly, 19 (1982): 251-257. ‘Being D need not be the same, in every case, as being C, so long as each instance of the disposition is identical to some instance of a categorical base’. The only problem becomes whether there is some plausible notion of a property instance that can warrant our consent. One candidate notion is that of trope, being an ‘abstract particular’ (cf. K. Campbell, Abstract Particulars (Oxford, 1990) or J. Bacon, Universals and Property Instances (Oxford, 1995), such as the redness of a particular apple or the squareness of a particular window. According to the ontology of tropes, squareness and redness in general do not exist, but only a sum of particular rednesses and squarenesses of things. But, as Mumford himself admits, the problem with such an account is notorious, and was noted by B. Russell: the trope theorist cannot explain how a number of tropes resemble each other. How is it that ten red things are all red, for instance? Or ‘what causes a square peg to fit a square hole? It is not a timeless universal that exists nowhere, rather it is something about this hole and this peg, regardless of what else exhibits a similar quality elsewhere’. (S. Mumford, Dispositions, p. 161). Indeed, but it is precisely on the ‘something about’ that we expect some explanation. Some authors have recently tried to answer such objections addressed to Resemblance Nominalism, in particular G. Rodriguez-Pereyra, Resemblance Nominalism – A Solution to the Problem of Universals (Oxford, 2002).

98

Dispositions and Causal Powers

The Generality Objection: Contingency or Necessity There are advantages in adopting a dispositionalist view of laws, as Mumford has clearly outlined.89 But the dispositionalist also faces problems: namely, how can he, without general laws, explain why generalities in behavior are true of kinds? For example, all sugar dissolves in liquid and tastes sweet to the tongue. Is it merely a cosmic coincidence that each sample of the kind carries the same set of capacities? Could we have rogue samples that possessed different dispositions? Here we face a very serious issue to which I am not sure Mumford provides a satisfactory answer: this has to do, it seems to me, with the contingency of the laws he still wants to maintain in his dispositionalist account. In the last chapter of Dispositions, Mumford clarifies the charge he had addressed in a previous article, namely that Ellis and Lierse argued that dispositional essentialism entailed ‘a definite answer to this debate: that dispositions were ultimate, rather than laws. This involved them making a number of controversial (and poorly justified) claims, such as the laws of nature were logically necessary instead of contingent as the Humean tradition takes them to be’. But, Mumford contends, dispositional essentialism could be true even if it were laws rather than dispositions that were in the relevant sense ultimate or basic. Which does not mean, he hastens to add, ‘that the basic laws view is proved’ or that he was offering a proof, even though in places he ‘pushed that view for dialectical purposes’.90 Further, ‘although basic dispositions as truthmakers of laws are not necessary for the claim of dispositional essentialism about certain fundamental kinds in science, as Ellis and Lierse suggested, …dispositional essentialism may be required for the credibility of the basic dispositions view because it suggests a solution to the generality problem that exists for the basic dispositions view. Hence, basic dispositions are not required by dispositional essentialism but dispositional essentialism may be required for basic dispositions’ (Ibid.). Intricate (or dialectical) as it may sound, such a formulation points to some difficulties one has in understanding Mumford’s exact position as far as the status of laws is concerned, as well as regards the explanatory role he wishes himself to confer to ‘basic’ dispositions. In particular, is Mumford really subscribing to the contingency thesis (CT: the laws of nature are logically contingent) about laws? In Dispositions, he explicitly says he endorses it: ‘Although dispositions bring natural necessity in the world there still exists the kind of logical contingency that is required. The replacement of laws with real dispositions is no threat to the view that the way nature behaves could have been otherwise but it does mean that CT will have to be revised accordingly’. Indeed, ‘a modified version of the CT can be given that is framed in terms of the contingency of a particular’s capacities, rather than the contingency of laws’. And he pursues: ‘There exist relevant conceptual necessities of dispositional essentialism, such as that described above where an electron must, 89 Cf. S. Mumford, Dispositions, p. 221 ff. 90 S. Mumford, Dispositions, pp. 235-236.

Dispositions and Essences

99

conceptually, have certain dispositions if it is to be an electron. Such conceptual necessities do not threaten natural contingency in the sense that the only logical necessities are conceptual. Hence it is logically contingent that a particular entity is a particular with the capacities it actually has. That particular could have had different dispositions to the ones it actually has, even though this would entail that the particular belonged to a different kind in virtue of that fact. Such contingency does not threaten the identities for such entities across possible worlds however, for our disposition ascriptions are actual-world relative’.91 Hence Mumford denies Ellis and Lierse’s move according to which dispositional essentialism requires the logical necessity of physical laws. ‘While it is certainly true that an electron would not be an electron if its behaviour were different from the behaviour it has in the actual world, this necessity is purely conceptual. That it is in virtue of behaviour B that a particular x is classified within a kind K does not entail that x necessarily has behaviour B’ (ibid.). According to Mumford, this involves a misunderstanding of the scope of the logical necessity involved. ‘That a particular possesses any disposition is logically contingent even though some particulars, such as electrons, would not have been classed as such if they had a different behaviour. To deny this would be to claim that an electron’s behaviour is dictated by logic, and, presumably, physics is a trivially analytic human folly’ (ibid.). But I do not think that such a criticism is justified, at least, as far as Ellis’s dispositional essentialism is concerned, which indeed does imply some necessitarist account of dispositions, if it aims at presenting kinds of processes as discovered principles active in nature and not as invented classifications. But to a certain extent too, I do not see how Mumford can maintain, even in a reformulated way, the CT about a particular’s capacities. What he may maintain at most (and in fact, it seems to be mainly what he provides as a further account of laws in ‘L’Etat des lois’) are some contingent elements understood as changes or departures from the law. At all events, I doubt that one could go so far as admitting such contingency as the one Mumford alludes too: ‘Thus ravens are black now, meaning that it is true of the raven kind that its normal members are dispositionally black, but tomorrow, through a divine intervention, normal ravens may become dispositionally pink’. Indeed this is making fun of physics or of the laws of evolution. It seems that the contingency Mumford wants to introduce has more to do with what Peirce called ‘conditional necessity’,92 in order to account both of the descriptive and regulative ‘would-be’ (which was not merely a ‘may be’ or a ‘will be’) character of dispositional laws and

91 S. Mumford, Dispositions, pp. 236-237. 92 C.S. Peirce, Collected Papers, vol. 2 § 664. This is a point recently emphasized by A. Bird, ‘Necessarily Salt Dissolves in Water’, Analysis 61 (2001): 267-275, and ‘On Whether Some Laws are Necessary’, Analysis 62 (2002): 257-274 in response to S. Psillos, ‘Salt dissolves in water, but not necessarily’, Analysis 62 (2002): 255-257. Cf also M. Kistler, ‘The Causal Criterion of Reality and the Necessity of Laws of Nature’, Metaphysica, vol. 3, no. 1 (2002): 57-86 and ‘Necessary Laws’, in J. Faye, P. Needham, U. Scheffler and M. Urchs (eds), Nature’s Principles (Dordrecht:, 2005), pp. 201-227.

100

Dispositions and Causal Powers

of the tychistic or chance elements he deemed necessary to introduce in his both synechistic and evolutionary metaphysics.93 Anyway, should that really be an advantage for the dispositionalist account of laws to try, by all means, to avoid any kind of necessity? For one thing, as we know from Kripke’s and Putnam’s works, a point also recently emphasized by F. Jackson,94 many necessary facts can be known to be true by a posteriori means only: certain identity statements, including identity statements concerning scientific and natural kinds, express propositions that are necessarily true, but that can be known only a posteriori. Hence, in general, as A. Bird has shown,95 the thought that laws might be necessary but knowable only a posteriori is not objectionable and should be familiar from other cases. Furthermore, it can be shown that some96 laws are necessary but have every appearance of contingency: ‘Precisely this relationship can be shown to hold between the existence of salt (sodium chloride) and the law that salt dissolves in water. The underlying law in this case is Coulomb’s law which governs both the electrostatic attraction required for salt to exist and also is sufficient to ensure that salt dissolves in water. Clearly the law that salt dissolves in water is a posteriori and at first sight it seems entirely contingent. But it can be shown to be necessary, even if we assume that the underlying laws are contingent’.97

Whatever we decide in favor of contingent or necessary laws, it remains that one should be careful, when adopting such and such form of dispositional essentialism or realism, not to indulge in what Popper thought was the major defect of essentialism: to aim at providing an ultimate explanation of what reality consists in. In that respect, and although Mumford’s recourse to ‘ungrounded’ dispositions may sometimes sound a little bit like a virtus dormitiva, I prefer to view it less as an invitation to quietism (something like Wittgenstein’s ‘justification must end somewhere’) than as

93 This is how Peirce intends to explain how, while following the principle of continuity (synechism), laws, in the sense of general indeterminate habits or dispositions do not follow an ironlike necessity, in the course of evolution, but a ‘conditional necessity’, which allows not only some radical spontaneity or chance (tychism) as that of the germinal chaos from which they come, but departures or ‘habit-changes’, these being always conceived, nonetheless, by means of probabilistic laws (on a propensity (or would-be) more than frequency model such as a die’s). Cf. C. Tiercelin, ‘C.S. Peirce et le projet d’une métaphysique scientifique évolutionnaire’, Publications de la Sorbonne, Philosophie, 5 (2000): 453-463. 94 F. Jackson, From Metaphysics to Ethics: a Defence of Conceptual Analysis (Oxford, 1998). I have analyzed and commented such views in C. Tiercelin, ‘La métaphysique et l’analyse conceptuelle’, Revue de Métaphysique et de Morale, 107, no. 4 (2002): 559-585. 95 A. Bird, ‘The Dispositionalist Conception of Laws’, Foundations of Science, 10 (2005: pp.353-370. 96 This qualification is important: it means that, in such a view, one does not have to deny that the fundamental laws are contingent. 97 Cf A. Bird, ibid., p. 8, § 7.1; see also A. Bird, ‘On whether some laws are?? necessary’, Analysis, 62 (2002): 257-274.

Dispositions and Essences

101

98

a principle (hence always possibly revisable) of sound, fallibilistic methodological procedure.

98 Indeed, Mumford seems in more recent texts to be coming to a more critical attitude toward ‘ungrounded’ dispositions and to a revision of his position concerning the contingency thesis about laws, mid-way between Humean lawlessness and nomological realism, in favor of realist lawlessness. See S. Mumford, Laws in Nature (London, 2004).

This page intentionally left blank

Chapter 4

The Causal Efficacy of Macroscopic Dispositional Properties Max Kistler

Introduction A vase falls from high on hard ground. Small wonder it breaks. Is its fragility among the properties causally efficacious for its breaking? It is controversial whether a property can be both dispositional and causally efficacious. Fragility is a dispositional property because its identity depends essentially on a counterfactual conditional: an object is fragile to the extent that it would, in otherwise normal circumstances, break if it were to fall from high on hard ground. The antecedent of this conditional need never be actually satisfied: dispositional properties can be possessed while they do not manifest themselves, and even without ever manifesting themselves. The identity of categorical properties does not so depend on what would happen in certain counterfactual situations: whether an object is spherical is determined by a criterion that can be spelled out in terms of actuality: all points on the surface of the object are at the same distance from its centre. Among others, Armstrong1 holds that dispositional properties can be causally efficacious, first of all by bringing about their own manifestation, in a situation where they are put to the test. In the case of the vase, a fall from high on hard ground is a test situation2 for fragility, and breaking its 1 D.M. Armstrong, A Materialist Theory of the Mind (London, 1968), p. 88; D.M Armstrong, A World of States of Affairs (Cambridge, 1997), pp. 70-71. 2 In a test situation, which is otherwise normal (more on this restriction later), the disposition manifests itself. Cf. R. Carnap, ‘Testability and Meaning’, Philosophy of Science, 3 (1936): 420-471; N. Goodman (1955), Fact, Fiction and Forecast, 3rd edition (Indianapolis, 1973), and the introduction to this volume. Many dispositional properties give rise to different characteristic manifestations in different types of test situation. D.H. Mellor, ‘The Semantics and Ontology of Dispositions’, Mind, 109 (2000): 760. Mellor mentions mass as an example of such ‘multi-conditional’ dispositional properties: the mass m gives its bearer both the disposition to accelerate with f/m, in a situation in which it is subjected to a force f, and the disposition to exercise the force m²/r² on another mass m, in a situation in which they are at a distance r. Another case is fragility, which also gives its bearer many dispositions: the disposition to break in a certain type of situation, the disposition to fissure in another type of situation. In a similar way, the high temperature of an inflammable gas is a dispositional property that gives the gas the disposition to be at a certain pressure, but also the disposition

104

Dispositions and Causal Powers

manifestation. Prior, Pargetter and Jackson3 argue on the contrary that dispositional properties (or dispositions) are in principle incapable of causal efficacy, in other words that they are epiphenomenal. However, there is more agreement among the parties in this debate on the interpretation of the nature of dispositions than these opposite theses suggest. It is generally presupposed in the debate that the dispositional properties at issue are macroscopic, and in principle reducible to a microscopic reduction base. By ‘macroscopic’ I simply mean that they are properties of objects that have parts: the properties of the whole are macroscopic relative to the properties of its parts which are by definition, microscopic. In this sense, ‘microscopic’ and ‘macroscopic’ are relative terms. It is common ground shared by both positions that the only causally efficacious properties are the microphysical base properties. Because it is usually taken for granted that these properties are not themselves dispositional, the reduction base is often called the ‘categorical base’ of the disposition. The disagreement is only over whether the macroscopic disposition inherits this efficacy by being identical to its reduction base (which is Armstrong’s position)4 or whether it is epiphenomenal

to explode. Thus, in the case of such properties, the distinction between dispositions and dispositional properties is important. Nevertheless, I shall often speak indifferently of dispositions and dispositional properties, in contexts in which there is no danger of confusing the dispositional property with the different dispositions to manifest itself the property gives its bearer. 3 E. Prior, R. Pargetter and F. Jackson, ‘Three Theses about Dispositions’, American Philosophical Quarterly, 19 (1982): 251-257. 4 In the general context of the analysis of the logic of reducing one scientific theory to another, many authors follow Causey’s thesis that reduced and reducing properties are identical. R.L. Causey, Unity of Science (Dordrecht, 1977). In Locke, one can find both the doctrine of the identity of dispositions (or, as he says, powers) with their microscopic categorical basis, and the idea that the former only depend on the latter, which is the thesis I shall defend. Locke seems to express the former doctrine when he says that ‘whiteness or redness are not in it [i.e. in the Porphyre] at any time, but such a texture, that hath the power to produce such a sensation in us’ (Locke 1689, II, 8, 19). The colours are identical with a texture, which is a microscopic categorical base property; and the latter can be causally efficacious in producing sensations in us when we look at a coloured object. But on the other hand, Locke’s view may be interpreted as being compatible with the second doctrine when he says: ‘Colours and Smells, … and other the like sensible Qualities … are in truth nothing in the Objects themselves, but Powers to produce various sensations in us, and depend on those primary Qualities, viz. Bulk, Figure, Texture, and Motion of parts’ (Locke 1689, II, 8, 14); here, Locke says that sensible qualities depend on the texture and other ‘primary qualities’, i.e. microscopic categorical properties, rather than being identical with them. I will express essentially the same idea by saying that the microscopic categorical properties determine the macroscopic dispositional properties. To say, as Locke, that these secondary qualities are ‘Powers to produce various sensations in us’, seems compatible with saying, as I shall do, that they are macroscopic properties that can be conceived as dispositions which can cause (produce) certain effects (e.g. sensations).

The Causal Efficacy of Macroscopic Dispositional Properties

105

because it is not identical to its reduction base (which is the ‘functionalist’ position defended by Prior, Pargetter and Jackson).5 In this chapter, I challenge the consensus about the monopoly of microproperties in matters of causal efficacy, by showing that it is coherent, and at least sometimes also plausible, to conceive of dispositional macroproperties as causally efficacious and nevertheless distinct from the microscopic properties in their reduction base. ‘Is fragile’ designates a macroscopic property of the vase, which is dispositional in the sense that its bearer has (by virtue of the meaning of the word ‘fragile’) the disposition to break when falling from considerable height on hard ground. However, the possibility of conceiving of this property in a dispositional way does not prevent it from causally contributing to the breaking. One can conceive of one property in both dispositional and in categorical ways. The second conception is in play when one attributes causal efficacy to the property. Part of the importance of this issue lies in its bearing on the nature of mentality: If I am right about macroscopic dispositions in general, this gives legitimacy to the idea that our desires, beliefs and other psychological properties give us dispositions to think and act, and contribute thus causally to the actions by which these dispositions manifest themselves. There are several arguments believed to show that dispositional properties in general are not causally efficacious as such. First, I shall briefly provide some considerations showing how to resist these arguments. Then I will turn to a group of arguments that Armstrong and the functionalists have offered specifically against the idea that dispositional macroscopic properties, which are micro-reducible, have their own causal efficacy. General Arguments Against the Efficacy of Dispositions Many philosophers deny the efficacy of dispositions in general, for one or several of the following reasons: 1. According to one such argument, dispositional properties lack causal efficacy because they are unobservable. The issue of realism with respect to theoretical properties lies beyond the scope of this paper. According to scientific realism, theoretical predicates used in science refer to real objects and properties even if they are not directly observable. However, to disarm this argument against the efficacy of dispositions, it suffices to note that the impossibility of observing them directly does not distinguish the dispositional property of being brittle from theoretical properties, such as being magnetized or having spin ½. By parity of reasoning, the fact that the former are unobservable is not a reason to doubt their reality and causal efficacy any more than it is a reason to doubt the reality and causal efficacy of the latter. Now, from a realist point of view, 5 F. Jackson, R. Pargetter and E.W. Prior, ‘Functionalism and Type-Type Identity Theories’, Philosophical Studies, 42 (1982): 209-225.

106

Dispositions and Causal Powers

explanations mentioning unobservable theoretical properties are accepted as causal explanations; therefore, to the extent that the fact that a piece of iron is magnetized is accepted as causally responsible for the movement of the iron filings it attracts, the fact that brittleness is not observable does not constitute a legitimate reason to deny that it can causally contribute to the vase’s breaking. 2. A second argument consists in pointing out that non-occurrent properties cannot be efficacious and that no property can be both dispositional and occurrent. However, the idea that dispositions are non-occurrent seems to result from the following fallacious reasoning. Dispositional properties seem to be non-occurrent (and therefore incapable of causal efficacy) because they are conditional or hypothetical. The property (or disposition) that every French citizen over 25 possesses to be potentially, or conditionally, president of France is not sufficient for giving every French citizen the causal powers of the president. All one may legitimately conclude from this is that the disposition does not have the causal powers possessed by its manifestations. This does not preclude the possibility that the dispositional property itself lends other powers to its possessor, e.g. to become a candidate for presidency. It is the manifestations of the disposition, not the disposition itself, that exist only conditionally or hypothetically. It is fallacious to argue from the nonoccurrence of the manifestation to the non-occurrence of the disposition. Therefore, nothing prevents us from allowing that dispositions are occurrent even while they are not being tested, and thus while they do not manifest themselves.6 3. According to a third argument, dispositions are permanent states, or static properties, whereas only changes can be causes. The correct reply seems to me to follow Mill7 in refusing the obligation of philosophical analysis to stick to common sense at any cost. True, in most ordinary circumstances, common sense prefers to pick out changes as ‘the cause’ among many causal factors. However, from a scientific and according to Mill a philosophical point of view, stable factors can play exactly the same causal role in producing an effect as changes. This gives us ground enough for holding that a stable factor

6 This point has been made by C.B. Martin, ‘Final Replies to Place and Armstrong’, in D.M. Armstrong, C.B Martin and U.T. Place, Dispositions: A Debate, (ed.) T. Crane (London, 1996) and S. Mumford, Dispositions (Oxford, 1998). Martin observes that the disposition itself is ‘something that is fully real and actual (unlike some of the manifestations) […]. Dispositions are actual continuants that predate, outlast, and may exist entirely without the existence of their manifestations.’ (Martin, ‘Final Replies to Place and Armstrong’, p. 166; Cf. also S. Mumford, Dispositions, p. 74). Mumford makes the same point even more explicitly: ‘“Categorical” means “unconditional” and this casts doubt on the putative distinction that is being drawn [between the dispositional and the categorical] because dispositions are, in a very obvious way, categorical.’ (Mumford, Dispositions, p. 64). 7 J.S. Mill, A System of Logic, Ratiocinative and Inductive (Honolulu, Hawaii, 2002).

The Causal Efficacy of Macroscopic Dispositional Properties

107

8

can be a ‘cause […] philosophically speaking’. If I introduce tension into a copper wire, the change in tension is certainly a causally efficacious factor in producing electrical current flowing through the wire. But the resistance of the wire, although a stable disposition that does not undergo any change, is also a causally efficacious factor determining, together with the tension, the intensity of the current. The wire’s resistance contributes causally to determining that intensity by microscopic interactions between the conduction electrons and the microscopic constituents of the wire. However, the fact these microscopic interactions bring about microscopic changes does not imply that the macroscopic resistance undergoes any change whatsoever. Thus, it is a stable factor that is nevertheless causally efficacious.Two other arguments against the efficacy of dispositional properties have affinities with the argument that permanent states cannot be causes. 4. Dispositions are facts whereas only events can be causes. Helen Steward9 has recently argued that by attributing a disposition to an object one can only refer to a fact, not to an event. Now, according to Steward, facts are entities that exist only by virtue of the linguistic expressions whose meaning they are; therefore, they lack causal efficacy although it can be relevant to mention them in an explanation. This argument has the same source as the preceding one because the reason why it is linguistically awkward to form expressions referring to events with dispositional predicates, stems from the permanent nature of dispositions and from the fact that it is more natural to conceive changes as events rather than permanent states. However, this argument presupposes the Davidsonian conception of the distinction between events and facts, according to which events are particulars whereas facts are linguistic entities whose identity is determined by virtue of the meaning of the words expressing them.10 This is not the right place to go into the details of the complex debate on the nature of events and facts.11 Let me just note that Davidson’s account has the implausible consequence of denying that there is any difference among the properties of a given cause event, with respect to their contribution to bringing about a given effect event, over and above a pragmatic difference between good and bad explanations. Consider a red billiard ball that hits a white billiard ball at rest by a central elastic shock, thereby transferring the red ball’s momentum M to the white ball. Let us now compare two causal explanations of the fact that after the shock, the white ball carries momentum M. The first explains this fact by saying that the red ball’s carrying M at the moment it hit the white ball with an elastic shock is causally 8 J.S. Mill, ibid., III, V, 3, p. 217. 9 H. Steward, The Ontology of Mind (Oxford, 1997). 10 Cf. D. Davidson, Essays on Actions and Events (Oxford, 1980). 11 Some efforts for clarifying this distinction can be found in M. Kistler, ‘Causes as Events and Facts’, Dialectica, 53 (1999): 25-46 and M. Kistler, Causation and Laws of Nature (London, 2006).

108

Dispositions and Causal Powers

responsible; the second explains it by saying that the red ball’s being red at the moment it hit the white ball with an elastic shock is causally responsible. It is clear that the first explanation is not only good and relevant but also true and that the second is not only bad but false. There must surely be something about that causal relation that makes the former true and the latter false. According to one account, the truth-maker of the first explanation is the fact that there is a relation of causal responsibility between the fact that the red ball carries M before the shock and the fact that the white ball carries M after the shock. However, there is no analogous truth-maker for the second explanation: there is no relation of causal responsibility between the fact that the red ball is red before the shock and the fact that the white ball carries M after the shock.12 If this account is correct, it is legitimate to attribute a causal role to facts: they are terms of relations of causal responsibility. Therefore, that attributions of dispositions normally have a factive rather than an eventive format is no reason for denying that dispositions can contribute causally to bringing about their manifestations. 5. Squires13 argues that the hypothesis that dispositions can be causes leads to an infinite regress. According to Squires, in order to explain why a disposition manifests itself at some times but not always, it is necessary to postulate a further disposition: the disposition which the original disposition has to manifest itself. But then, this new disposition itself can also manifest itself or not. Therefore, we must suppose that it has yet another disposition to manifest itself, and so on to infinity. To this, Armstrong14 has given a convincing reply. Armstrong compares Squires’ infinite regress to the infinite series of facts existing superposed on any fact p: the fact that it is true that p, the fact that it is true that it is true that p, etc. One can account for this infinite series of facts in at least two ways: either one distinguishes a linguistic concept of fact, according to which this is really an infinite series of different facts – because their linguistic expressions differ – from a ‘Russellian’15 conception according 12 I have developed this argument for the existence of facts, from the truth-makers of causal explanations elsewhere. Cf. M. Kistler, ‘Erklärung und Kausalität’, Philosophia Naturalis, 39, Heft 1 (2002): 89-109. 13 R. Squires, ‘Are Dispositions Causes?’, Analysis 29 (1968): 45-47. 14 D.M. Armstrong, ‘Beliefs as States’ (1973), in R. Tuomela (ed.), Dispositions (Dordrecht, 1978), p. 419. 15 This terminology is due to J. Bennett, Events and their Names (Cambridge, 1988), p. 41. According to Bennett, the identity of a ‘Fregean’ fact is determined by the meaning of the linguistic expression used to express it. The facts designated by two expressions are identical only if their linguistic expressions may a priori be derived from each other, by virtue of their meaning (Cf. Bennett, Events and their Names, pp. 35-37). However, ‘we sometimes use definite descriptions as though they were Russellian, regarding them merely as pointers to their referents’ (Bennett, Events and their Names, pp. 39-40). In this sense, two different statements can express one ‘Russellian’ fact even if their senses are not equivalent, i.e. if their expressions cannot a priori be derived from one another.

The Causal Efficacy of Macroscopic Dispositional Properties

109

to which a unique fact underlies all these linguistic facts. Or one can, as Armstrong proposes to do, distinguish between the linguistic expression of a fact and what makes such an expression true, its truth-maker. According to the first way of putting it, there is only one real disposition, which belongs to the one Russellian fact; according to the second way, just as the infinite series of facts superposed on p all have the same truth-maker, i.e. p, the one truthmaker of the infinite series of higher order dispositions is simply the firstorder disposition. Both analyses allow us to reject the objection of the infinite regress, by arguing that the apparent infinite series of dispositions described by Squires is only an artefact of language, whereas only one (Russellian) fact containing a unique (first-order) disposition makes them all true. 6. According to the traditional objection Molière has made famous by making fun of the alleged causal and explanatory power of the ‘dormitive virtue’ of opium, dispositions cannot be among the causes of their manifestations because they are related to them by an analytical and therefore necessary link, whereas causation is an essentially contingent relation. It often happens, at least in my kitchen, that a fragile object breaks after having fallen on hard ground. Now, it is part of the meaning of the predicate ‘is fragile’ that objects to which it applies break when, in otherwise normal circumstances, they fall from high on hard ground. Therefore, goes the argument, given that this fragile vase has fallen from high on hard ground, the judgement that is has broken after its fall is analytic. This implies that, in the sentence ‘this vase has broken after having fallen because it is fragile’, the word ‘because’ does not express a cause-effect relation but rather an analytic relation of implication, based on the meaning of the word ‘fragile’. So it seems that the vase’s fragility cannot be among the causes of its breaking. However, this argument is fallacious because it overlooks, in the statement of the relation between the disposition, a test condition and its characteristic manifestation, the essential reference to normal conditions. It is part of the requirements of the correct attribution of a disposition that it does not always manifest itself in the characteristic way in a given type of test situation, but only most of the time, in a statistical sense. Even a fragile vase that falls from high on hard ground does not always, or necessarily, break, but only most of the time, or in ordinary circumstances. One can imagine extraordinary circumstances in which the hard surface rests on springs absorbing the shock and other extraordinary circumstances in which the vase and the ground contain strong magnets repelling each other. True, such situations are far fetched. However, to show that there is no necessary (because analytic) link between the fact that a given object falls on hard ground and the fact that it breaks, it is sufficient to show that there are situations even if they are very rare, in which the first is a fact but not the second. In other words, what is analytically implied by the fact that the fragile vase falls on hard ground is only that it breaks in otherwise normal circumstances. However, that fact does not analytically entail that it breaks, tout court. Therefore, the fact that it

110

Dispositions and Causal Powers

breaks remains a contingent fact and nothing prevents the fragility of the vase from being among the factors causally contributing to its breaking. It has often been observed that dispositions do not always give rise to their typical manifestations in characteristic test conditions. Martin16 has shown it by carrying out the following simple thought experiment. He imagines a mechanism that is so conceived that, when connected to an object possessing a given disposition, prevents it from showing its characteristic manifestation in a given test situation. A copper wire under tension has the disposition to give an electric shock to anyone touching it (if the person’s feet touch the ground and the person isn’t wearing rubber soles). Now, Martin’s imagined ‘electro-fink’ is a mechanism that takes the disposition away if and only if it is put to the test, i.e. when someone touches it. In this type of test situation, the disposition, although really there, never manifests itself in the characteristic way. Relatively to the pair consisting in: , the fink takes care that circumstances are never ‘normal’. Martin’s electro-fink is not as far fetched as it appears at first sight: in fact, it is just an extreme case of a quite ordinary phenomenon. The electric circuit in every modern house contains a ‘circuit breaker’, a mechanism working just like the electro-fink. More generally, as Bird (1998) has noted, for every disposition and every relevant test situation, there are ‘exceptional’ circumstances where an ‘antidote’, to use Bird’s expression, prevents the manifestation although the disposition is present.17 The springs under the hard surface, which absorb the shock received by the falling object, act as an antidote against the vase’s fragility in this sense. Before we return to the analysis of the ‘ceteris paribus’ clause that accompanies the expression of the link between the disposition, the test condition and its manifestation, we must look at a group of arguments that question more particularly the efficacy of macroscopic dispositional properties. The Epiphenomenalist Trilemma for Macroscopic Dispositions All dispositions have manifestations. By the ‘causal base’ (or ‘categorical basis’) of a disposition possessed by an object, I mean the set of intrinsic (i.e. non-relational) 16 C.B. Martin, ‘Dispositions and Conditionals’, Philosophical Quarterly, 44 (1994): 1-8. 17 M. Johnston, ‘How to Speak of the Colors’, Philosophical Studies, 68 (1992): 221263 and G. Molnar, ‘Are Dispositions Reducible?’, Philosophical Quarterly, 49 (1999): 1-16 call them ‘maskers’: they mask the original disposition by preventing it from manifesting itself. D. Lewis, ‘Finkish Dispositions’, Philosophical Quarterly, 47 (1997): 143-158 and W. Malzkorn, ‘Realism, Functionalism and the Conditional Analysis of Dispositions’, Philosophical Quarterly, 50 (2000): 452-469 have tried to analyse the meaning of disposition attributions with counterfactual conditionals that do not contain any ‘ceteris paribus’ clauses, thereby avoiding refutation by cases such as Martin’s.

The Causal Efficacy of Macroscopic Dispositional Properties

111

properties of the object that contribute causally to bringing about the manifestations of the disposition. Insofar as the bearer of the disposition contributes causally to bringing about its manifestations, the properties of the bearer of the disposition contain the causal basis for that disposition. Within this conception of the causal base, the properties belonging to the causal base are categorical properties, because they are causally efficacious. However, in the case of macroscopic dispositional properties, there are different ways of conceiving of the relation between a disposition and its causal base, which lead to denial that the disposition has its own causal powers. 1. According to the ‘functionalist’ conception of dispositions, a disposition is a second-order property. However, only a first-order property can be causally efficacious. There are two reasons for thinking that dispositions are not identical to any first-order properties. First, one disposition may be grounded on different microscopic bases; second, it is contingent which base underlies a given disposition. 2. According to an important conception of intertheoretical reduction,18 the reduction of a macroscopic property to underlying microscopic properties discloses an identity. To the extent that the disposition is identical with its categorical reduction base, it is causally efficacious because the latter is. 3. Kim’s19 conception of functional reduction combines elements of the other two conceptions. According to this third conception of the relation between a disposition and its base, attributing a disposition to an object means to attribute a second-order predicate to it: the object possesses a property that plays the role of causing the disposition’s characteristic manifestation in a given test situation, when circumstances are otherwise normal. The predicate specifying the role is second order to the extent that the reference to a property playing the role is equivalent to an existential generalization over first-order properties. Only the property playing the role is causally efficacious, but not the disposition corresponding to the role itself. However, Kim also considers that the property playing the role is necessarily microscopic, even when the disposition is attributed to a macroscopic object. None of these conceptions allows macroscopic dispositions to have their own distinctive causal powers. This is a dilemma, in fact a ‘trilemma’, because 1) there does not seem to be any further possibility and 2) neither of these possibilities seems to be compatible with our common sense intuition that our own mental properties, at least, are macroscopic properties that are causally responsible for our own actions without necessarily being identical with any microscopic property of our brain.20 I 18 R.L. Causey, Unity of Science (Dordrecht, 1977). 19 J. Kim, Mind in a Physical World (Cambridge, 1998). 20 Whether this intuition is correct is of course controversial. It is the object of a now classical debate between Wittgenstein (1953) and his followers on one side, who deny conceptual coherence to entities (the so-called ‘mental representations’) whose content justifies

112

Dispositions and Causal Powers

call the trilemma ‘epiphenomenalist’ because all alternatives end up denying that the macroscopic dispositional properties have any causal powers of their own. As far as they have their own identity, they appear therefore as epiphenomenal. According to the first horn of the trilemma, dispositions lack efficacy, whereas according to the second and third, they are efficacious but only by being identical to their microscopic base properties. I shall show that the trilemma can be avoided, and that macroscopic dispositional properties can be conceived as causally efficacious though not identical to their microscopic bases, with the help of two premises: 1. the distinction between the dispositional and the categorical applies to predicates and not properties, and 2. the categorical base is not necessarily the reduction base. Rejecting both of these theses leads to the first horn of the dilemma represented by functionalism, whereas accepting 1 but not 2 leads to the latter two horns of the epiphenomenalist trilemma, represented by Armstrong’s and Kim’s accounts. I shall argue for thesis 1 on pp. 112-120, and for thesis 2 on pp. 120-124. Dispositional and Theoretical Properties Before arguing that the distinction between the dispositional and the categorical bears on predicates rather than on the properties those predicates designate, I shall sketch a general account of the role dispositions play in explanations, and in particular scientific explanations. The logic underlying explanations referring to dispositions is best revealed in exceptional cases where a disposition does not manifest itself although it is subject to one of its characteristic tests. To explain the exceptional behaviour of an object that does not manifest one of its dispositions in a given test situation, one refers to other properties that may be those of the object itself, or may more generally belong to the situation. As we have already seen, in the clause expressing the dependence of the characteristic manifestation with respect to the object’s being in a test situation, it is essential to mention the otherwise ‘ordinary circumstances’ relevant for the case under consideration, because each concrete test situation possesses innumerable other properties that can in principle interfere with the manifestation of the disposition. an action and which are nevertheless also causally responsible for that action, and Davidson (1963) on the other side, who argues that it is necessary for our conception of ourselves as agents exercising causal power on our own actions to suppose that our reasons to act are at the same time causally responsible for our actions, by being mentally represented. The intuition mentioned above is compatible with Davidson’s intuition but not with Wittgenstein’s. L. Wittgenstein, Philosophical Investigations, trans. G.E.M. Anscombe (Oxford, 1953). D. Davidson, ‘Actions, Reasons, and Causes’, in D. Davidson, Essays on Actions and Events (Oxford, 1980).

The Causal Efficacy of Macroscopic Dispositional Properties

113

However, the scientific conception of properties aims at doing away with that ‘ceteris paribus’ clause. Thus, one may turn to a scientific explanation to find out why a given disposition did not manifest itself in a particular test situation; in this case, the scientific explanation must not itself contain any ceteris paribus clause. Take the case of an object in free fall, close to the surface of the Earth. If I let go of an object that is not held up by anything else, it has a disposition to fall a distance of s=1/2 gt² in t seconds. However, in the presence of such ‘antidotes’ as friction with the surrounding air, the disposition does not manifest itself in this way. The discovery of the different antidotes present in a given concrete situation, which allows us to explain an object’s behaviour that deviates from the direct manifestation of the disposition, is a scientific task and requires in general the postulate of properties that are not directly observable. Ideally, if all factors contributing to determine the process have been identified, it is possible to explain the manifest behaviour without any need for a ceteris paribus clause expressing our partial ignorance of the circumstances. Once a complete scientific description of the situation has been given, dispositional properties can be conceived as powers necessitating their effects. However, these effects are not necessarily manifest; they can themselves be powers. In the case of free fall, a scientific description of the situation allows us to substitute for the body’s disposition to fall 1/2gt² meters in t seconds, the force g ⋅ m that produces an acceleration g. Let us call the force and the acceleration ‘constraints’ or ‘powers’ that determine each other by virtue of laws of nature. The force g ⋅ m determines a power to accelerate by g by virtue of Newton’s law, better known in the form F = m ⋅ a . This tendency to accelerate, although it is a necessary consequence of force, does not necessarily manifest itself directly. What is manifest is the result of the superposition (or of the interaction) of all tendencies to move imposed on the body by the different powers that are present. Friction with air is another power present in the situation, which imposes another tendency to accelerate in a direction opposed to that of the acceleration produced by the force of attraction. At this point we must pause to consider an important objection to my thesis that theoretical properties are powers related by laws to other properties that, being themselves powers, do not necessarily manifest themselves directly.21 Suppose Galileo’s law of free fall, taken as a hypothesis, predicts that a body falls 1/2gt2 metres in t seconds but that observation tells us that the distance it really falls is less. My account suggests that this is reason enough to legitimate the postulation of a power to fall 1/2gt2 metres, which does not manifest itself directly, instead of being a case refuting Galileo’s hypothesis. However, if the discrepancy between theoretical prediction and observed result was in itself sufficient to justify the postulation of that power, it would seem that we have legitimized a general strategy of immunization of theories, which would even allow justifying phlogiston theory. Given that empirical observation contradicts the prediction of phlogiston theory that the residue of combustion is lighter than the body before combustion, my account 21 This objection is also discussed by P. Lipton, ‘All Else Being Equal’, Philosophy, 74 (1999): 155-168. Cf. Schrenk, this volume.

114

Dispositions and Causal Powers

seems to legitimize the postulation of a tendency, or power, of combustible bodies to become lighter during combustion, a power which does not, however, directly manifest itself by any measurable loss of weight. This objection can be overcome by requiring that the postulate of a power is subject to the usual conditions of legitimacy for the postulate of theoretical entities: the postulate of a power that does not directly manifest itself is scientifically legitimate only to the extent that it is possible to provide, in each situation in which it is not manifest, an independent explanation of the fact that it is not manifest.22 It is legitimate only if the difference between the postulated power and the manifestation can be accounted for by the interference of other factors whose presence can be independently detected. In other words, the hypothesis explaining the discrepancy between the regular manifestation of the power and what is observed must abide by the Popperian criteria for an empirically meaningful, non ad hoc, hypothesis, as much as any other scientific hypothesis. The hypothesis of a power of combustible objects to lose mass during burning is not scientifically legitimate because the only way to reconcile it with the empirical fact that burning makes objects more massive, is by making another postulate which cannot be independently justified, namely the ad hoc postulate that they gain mass by liberating a substance with negative mass, phlogiston. By contrast, attributing to free falling bodies the power to fall 1/2gt2 metres in t seconds is legitimate insofar as one can find, in each situation empirically investigated, independent grounds for postulating interfering factors (themselves powers) such as friction with air, whose effects add up to, with that of the initial power, the distance fallen, empirically observed. The main conclusion that I propose to draw from this analysis is this. Substituting a force that itself produces a tendency to accelerate, for the disposition to fall 1/2gt² metres in t seconds, corresponds to a change in the way of conceiving of one and the same property. The first conception of the property is grounded on a dispositional predicate whose meaning contains a link to the manifestation of the disposition, which raises the suspicion of analyticity. However, the second conception of the same property on the basis of scientific predicates justifies attributing causal efficacy to it: the link between two properties established by a causal law is never a priori, because laws of nature are discovered a posteriori. The fact that the link is not analytic removes the suspicion that the thus conceived properties are ‘dormitive virtues’. Such a scientific conception thus puts us in a position to conceive in a categorical way the same property that has first been conceived of dispositionally. This analysis suggests that the distinction between the conception of a property as dispositional and its scientific conception as categorical can be grounded on the following criteria. The attribution to an object of a dispositional property exercising a constraint on its evolution is subject to three requirements: 22 Pietroski and Rey show that it is necessary and sufficient to impose such a constraint to save ceteris paribus laws (i.e. laws that allow for exceptional situations) from vacuity. P. Pietroski and G. Rey, ‘When Other Things Aren’t Equal: Saving Ceteris Paribus Laws from Vacuity’, British Journal for the Philosophy of Science, 46 (1995): 81-110.

The Causal Efficacy of Macroscopic Dispositional Properties

115

First, the dispositional property is just one among many properties of the object and the overall situation. Some of these other properties are unknown. A fragile object has a property that imposes a constraint on it to break when it falls from high on a hard surface; however, among its unknown properties some may impose other constraints acting against this first constraint, thereby preventing its breaking after the fall. By contrast, to the extent that the attribution of the property is part of a complete specification of the situation by scientific predicates, which completely determines the evolution of the situation, the attribution (and the property thus conceived) is not dispositional. One wouldn’t call a vase fragile in a situation in which one knows that its tendency to break is counterbalanced by an installation that is sure to absorb the shock.23 The hypothesis according to which some of the other properties of the situation must be unknown in order for it to be appropriate to attribute a disposition allows us to explain why it is impossible to specify explicitly those ‘ordinary circumstances’ in which the disposition, subject to a test, manifests itself in the characteristic way. By contrast, no such restrictions are placed on the scientific conception of properties. The attribution of a property according to its scientific conception can in principle be part of a complete description of the situation. Second, the attribution of a disposition implies the truth of a counterfactual conditional that necessarily contains a ceteris paribus clause. By contrast, the attribution of a scientifically conceived power implies a strict counterfactual conditional.24 Third, a property is conceived as dispositional insofar as it is conceived as establishing a (ceteris paribus) dependency of a manifestation on a test situation, where both are described in observational terms. Falling and breaking are observable conditions, as well as being dropped and falling s metres in t seconds.25 However, the 23 Bird makes the same observation. ‘The combination of [uranium] pile and boron rods, [...] does have a disposition to chain-react when the rods are outside the pile, but loses this disposition when the rods are in the pile. [...] The reactor as a whole, [...], i.e., including the fail-safe mechanism, as long as the mechanism is effective has no disposition to explode at all.’ A. Bird, ‘Dispositions and Antidotes’, Philosophical Quarterly, 48 (1998): 229-230. 24 According to most interpretations of quantum physics, there are fundamental probabilistic laws. However, these laws allow us to make predictions that do not depend on partially unknown aspects of the circumstances, as is the case with predictions grounded on ceteris paribus generalizations, in particular on the attribution of dispositions. 25 The manifestations of some dispositions, which may be called ‘spontaneous’, do not depend on any particular test condition. Radioactive substances have the disposition to decay; however, this manifestation need not be triggered by any test condition, observable or not. Having a belief is having the disposition to act as if the belief were true. No external and observable factor is necessary to trigger an action manifesting the disposition: I can express it by producing an utterance, without being under the influence of any external stimulation. No doubt, Aristotle was the first to insist on the fact that the independence of human utterances of all external stimulation is an essential feature of human language that distinguishes humans

116

Dispositions and Causal Powers

identity of scientifically conceived properties is determined by laws that do not necessarily bear on observable properties.26 The account sketched here implies an important modification of Quine’s theory that a disposition is a ‘partially discerned physical property that will be more fully identified, we hope, as science progresses’27 and of Armstrong’s thesis that ‘dispositions, in fact, are primitive theoretical concepts’.28 Our analysis shows that the dispositional conception of a property can coexist with its scientific conception. Each obeys its own logic and serves its own purposes. The distinction between the dispositional and the categorical is an epistemic distinction and does not introduce any difference between efficacious and non efficacious properties. To the extent that it is possible to conceive of a property in a scientific way, i.e. to conceive of it as a relatum of a law of nature, it is legitimate to consider it to be causally efficacious, even in case there also exists a dispositional conception of the same property. The fact that a disposition does not always give rise to its characteristic manifestations while it is subject to a relevant test, can be explained by the simple fact that it is not the only property of the situation. Furthermore, the fact that such a property is not by itself sufficient to produce a certain effect does not constitute any reason to deny it causal efficacy. For the same is true of clearly efficacious factors, such as the momentum M of the billiard ball which is only causally responsible for the fact that the second ball has M after the shock, because the situation is such that the shock is elastic. The uncertainty about whether the disposition manifests itself in a given situation, which is expressed in the ceteris paribus clause of the conditional linking the test condition to the manifestation, has its root in the partial ignorance of the circumstances, which is a necessary condition for the attribution of a disposition. Neither the fact that a property is not in itself sufficient for producing the effect nor the fact that we do not know whether this effect will come about in a situation we partially ignore, provides any reason to deny it causal efficacy.

from all animals producing sounds or other signs that can have all physical characters of human speech. 26 Theoretical properties have second-order, relational properties in virtue of the laws bearing on them. If a copper cable has conductivity σ, this conductivity constrains other properties of the same object in virtue of the laws governing it. It exercises, e.g. a constraint on the electric current and the electric field, to be related as J/E = σ, in virtue of the law J=σE. If an object O has mass m1, and another mass m2 is in its neighbourhood, O’s property of having mass m1 constrains O’s movement in imposing upon it a force F = γ ⋅ m1m2 r2 , by virtue of the law of gravitation. The complete set of properties instantiated by the body determines, together with the properties of its surroundings with which it interacts, its evolution and causal interactions. For a defence of this thesis, see Kistler (2002a). 27 W.V. Quine, The Roots of Reference (La Salle, 1971), p. 13. 28 D.M. Armstrong, ‘Beliefs as States’, p. 420.

The Causal Efficacy of Macroscopic Dispositional Properties

117

One way to express the thesis that one may conceive of a given property alternatively either in a dispositional or in a categorical way, is to say that the dispositional-categorical distinction applies to the predicates making reference to properties, not to the properties themselves.29 An important argument for this thesis consists in reducing its negation ad absurdum. Suppose that the dispositionalcategorical distinction applies to properties themselves, independently of the predicates we use to make reference to them. All natural properties take part in laws of nature by virtue of which the objects possessing those properties have also other properties. This is also true of properties that we do not intuitively conceive of as dispositions, such as the property of a gas to have a certain temperature, or the property of a stone to have a certain mass, m. By virtue of the so-called ideal gas law, all gases (whose properties are sufficiently close to the properties of an ‘ideal’ gas) that are at temperature T also have, in a volume V, the pressure p=nRT/V (where R represents a constant factor, and n indicates the quantity, in moles, of gas molecules contained in the sample under consideration). Now, the existence of this law provides us with a means of conceiving of the property of being at temperature T in a dispositional way: it gives its bearer the disposition to have another property, to which it is linked by the law. Temperature is a dispositional property insofar as it gives a gas having it the disposition to be at a pressure p, when n moles of it occupy a volume V. Similarly, its property of having a mass m gives our stone, by virtue of the law of gravitational attraction, the disposition to move towards other 29 Alston was one of the first to challenge the ‘assumption that the dispositional and the “occurrent” (“episodic”) interpretations are incompatible’. W.P. Alston (1971), ‘Dispositions, Occurrences and Ontology’, in R. Tuomela (ed.), Dispositions (Dordrecht, 1978), p. 359. The thesis that the dispositional-categorical distinction applies to predicates rather than properties has been defended by Shoemaker, Mellor and Mumford. S. Shoemaker, ‘Causality and Properties’, repr. in S. Shoemaker, Identity, Cause and Mind (Cambridge, 1984), pp. 206-233; D.H. Mellor, ‘The Semantics and Ontology of Dispositions’, Mind, 109 (2000): 757-780; S. Mumford, Dispositions. ‘I think that the term “dispositional” is best employed as a predicate of predicates, not of properties.’ (S. Shoemaker, ‘Causality and Properties’, p. 211). ‘Dispositionality is a feature not of properties but of predicates, namely of those whose application conditions can be stated in reduction sentences. […] Properties in our sense […] need not in themselves be either dispositional or categorical; those that exist can just be.’ (D.H. Mellor, ‘The Semantics and Ontology of Dispositions’, p. 767). Lowe also distinguishes between ‘“occurrent” predication’ and ‘“dispositional” predication’, E.J. Lowe, ‘Dispositions and Laws’, Metaphysica, 2 (2001): 11. However, his conception is incompatible with ours: according to Lowe, dispositional predication attributes a universal property to an object via a kind of object of which it is an instance, whereas occurrent predication attributes a property instance to it. This distinction does not allow us to account for the difference between a dispositional and a categorical attribution of a property, insofar as it does not bring into play the semantic link between a disposition and its manifestation. Lowe puts the distinction in purely ontological terms; however, at an ontological level both types of attribution attribute ultimately the same type of property, though in different ways: the property instance which is the object of ‘occurrent predication’ is an instance of the very universal property which is the object of ‘dispositional predication’.

118

Dispositions and Causal Powers

massive bodies, and in particular to fall when it happens to be close to the surface of the Earth. Therefore, the hypothesis that dispositionality is a property of properties, and not of predicates or of our concepts of properties, leads to the result that all natural properties – i.e. all properties figuring in laws of nature – are dispositional. However, this is clearly incompatible with our intuitive understanding of the concept of disposition according to which some natural properties such as ‘consisting of water’ are not dispositional.30 This thesis that the dispositional-categorical distinction bears on our conceptions of properties and on the predicates expressing them rather than on the properties themselves allows us to offer a plausible interpretation of Popper’s thesis that all properties are dispositional.31 As we have just seen, for every natural property there are dispositional ways to conceive of it. If P is an arbitrary natural property P linked by a causal law to another property R, knowing this law puts us in a position to conceive of P as ‘the disposition to (cause) R’. Interpreted this way, Popper’s thesis loses much of the counter-intuitive character it is usually understood to have: that is, when it is interpreted to mean: all properties are dispositional and therefore, no property is categorical.32 As soon as the distinction is understood as bearing on concepts or predicates, nothing prevents a given property from being conceived as both dispositional and categorical.33 30 S. Shoemaker, ‘Causality and Properties’, S. Mumford, Dispositions, pp. 757-780 and D.H. Mellor, ‘The Semantics and Ontology of Dispositions’ show that one can avoid the paradoxical conclusion that all properties are dispositional by conceiving of the dispositionalcategorical distinction as a semantic distinction between predicates rather than as an ontological distinction between properties. 31 K. Popper, ‘The Propensity Interpretation of the Calculus of Probability, and the Quantum Theory’, in S. Körner (ed.), Observation and Interpretation (London, 1957), pp. 65-70. Cf. the introduction to this volume. Other defenders of this thesis are I.J. Thompson, ‘Real Dispositions in the Physical World’, British Journal for the Philosophy of Science 39 (1988): 67-79, pp. 76-77; S. Blackburn, ‘Filling in Space’, Analysis, 50 (1990): 62-65. R. Harré and H. Madden, Causal Powers (Oxford, 1975); R. Harré, ‘Is There a Basic Ontology for the Physical Sciences?’, Dialectica, 51 (1997): 17-34. N. Cartwright, Nature’s Capacities and their Measurement (Cambridge, 1989). N. Goodman, Fact, Fiction, and Forecast, makes the more modest claim that there are many more dispositional predicates than it appears at first sight. 32 There is a powerful objection against this interpretation of the thesis, variants of which have been put forward by P.J. Holt, ‘Causality and Our Conception of Matter’, Analysis, 37 (1976): 20-29. H. Robinson, Matter and Sense, (Cambridge, 1982). S. Blackburn, ‘Filling in Space’, Analysis, 50 (1990): 62-65; D.M Armstrong, ‘The Causal Theory of Properties: Properties according to Shoemaker, Ellis and others’, Philosophical Topics 26 (1999): 25-37; also published in Metaphysica, 1 (2000): 5-20. However, this objection, according to which the thesis that all properties are dispositional makes us ‘lose the substance of the world’ (P.J. Holt, ‘Causality and Our Conception of Matter’, Analysis 37 (1976): 23) does not bear against the interpretation suggested here. 33 A different way of reconciling the fact that all natural properties can be conceived of in a dispositional way, on the basis of the dispositions they give their bearers, with the paradoxical

The Causal Efficacy of Macroscopic Dispositional Properties

119

Frank Jackson has offered a way to avoid the conclusion that all properties are dispositional, which allows him to stick nevertheless to the traditional idea that the dispositional-categorical distinction bears on properties themselves. He introduces a distinction between properties that are essentially tied to characteristic manifestations in relevant test situations, and others that are only contingently tied to their manifestations. ‘What makes a property a disposition is that it itself is essentially linked to the production of certain results in certain circumstances…’34 However, insofar as the tie between a natural property and its manifestation in relevant circumstances is grounded on a law, it is difficult to justify Jackson’s distinction. It presupposes that there is, among the laws a property figures in, a first set of laws that are essential to the property and a second set of laws that are only contingent relative to the property. It is possible that the property exists even if the laws in the second set do not exist, though it is impossible that it exists without the laws in the first set. I do not know of any independent justification of this distinction between two sets of laws in terms of the modal strength with which they determine the identity of the property.35 Mumford has suggested a promising way to understand the dispositionalcategorical distinction at the level of predicates.36 A predicate D is dispositional if and only if its attribution analytically (in virtue of the meaning of the predicate) entails a counterfactual linking a test condition to a characteristic manifestation. The statement ‘The vase is fragile’ analytically entails: if the vase were to fall from high onto hard ground, all other circumstances being normal, it would break. By contrast, ‘the vase is made of thin terracotta’ entails the same counterfactual conditional, but not analytically: in the case of the latter statement, the entailment appearance of the thesis that all properties are dispositional is to say, with Martin and Heil, that all properties have a ‘dual nature’ (C.B. Martin, ‘Final Replies to Place and Armstrong’, J. Heil, Philosophy of Mind (London, 1998), p. 182), in other words that a property ‘endows its possessor with both a particular disposition or ‘‘causal power’’ and a particular quality’ (J. Heil, Philosophy of Mind, p. 181). However, rather than offering a solution, this is just a way of stating the problem that these two apparently incompatible ‘aspects’ do nevertheless coexist. My thesis that the dispositional-categorical distinction is of a conceptual and semantic nature solves the problem by offering a hypothesis for analysing the distinction, in terms of the difference between the analytic or a posteriori nature of the entailment, by the attribution of the property to an object, of counterfactual conditionals linking the instantiation of the property to characteristic manifestations, this difference being in turn explained in terms of the meaning of the predicates with which we make reference to the property. 34 F. Jackson, From Metaphysics to Ethics: A Defence of Conceptual Analysis (Oxford, 1998), p. 101. 35 Elsewhere (M. Kistler, ‘The Causal Criterion of Reality and the Necessity of Laws of Nature’) I have argued at length for the opposite thesis that all laws in which a given property takes part are essential to it. 36 I do not exactly follow Mumford’s way of spelling out the distinction, insofar as I put it exclusively in terms of the a priori/a posteriori distinction, whereas Mumford (as Jackson, 1998) sometimes expresses it by saying that categorical properties are contingently related to their nomological consequences.

120

Dispositions and Causal Powers

of the counterfactual conditional is not grounded exclusively on the meaning of the predicate ‘is made of thin terracotta’, but on laws, which are known a posteriori.37 The laws in which a natural property takes part guarantee the existence of such counterfactuals, but their knowledge is not always included in the meaning of the predicates with which we make reference to these properties. Hence this difference allows us to ground the distinction between dispositional and categorical predicates: the attribution of dispositional predicate entails a counterfactual linking a test situation to a manifestation in an analytical and therefore a priori manner, whereas that entailment is a posteriori in the case of categorical predicates.38 Two Ways of Conceiving of the ‘Categorical Basis’ of Dispositions David Armstrong’s view of dispositions is compatible with the thesis that dispositionality and categoricality are two ways of conceiving of properties which are, in themselves, neither dispositional nor categorical. Armstrong cites as an example, the occurrent state of a brittle glass. This state could be considered causally efficacious in that it contributes to the glass breaking when it falls on the kitchen floor. The properties that are causally efficacious in such a case are intrinsic properties of the glass. However, we often ignore those intrinsic properties. Therefore, instead of making reference to that occurrent state by naming directly its intrinsic (and potentially causally efficacious) properties, we do it by means of a definite description in terms of the state’s typical effects in certain circumstances. We call it ‘brittleness’, which is a word whose meaning is defined by its typical effects in certain circumstances. However, these are just two ways of referring to one state, one being direct (but inaccessible to us, because we ignore the intrinsic nature of the property), the other indirect in terms of its typical causes and effects. Two ways of referring to one state

37 Shoemaker gives the example of the predicate ‘being made of copper’ which is ‘not dispositional in this sense. There are causal powers associated with being made of copper – for example, being an electrical conductor. But presumably this association is not incorporated into the meaning of the term “copper”. (Shoemaker, Identity, Cause and Mind, p. 210). I would like to prevent a misunderstanding. Someone who knows the relevant laws about copper can infer a priori from the fact that something is copper, the fact that it is a good conductor of electricity. The same is true for someone who knows the lawful properties of terracotta. But the knowledge of those laws is a posteriori. Knowledge of the behaviour of these substances in counterfactual circumstances is not analytically entailed by the mere meaning of the predicate attributing them the properties of being of copper and of being of terracotta. 38 Another misunderstanding must be avoided here: it is of course not a priori whether a given object is fragile or not. But the attribution of fragility, itself made on empirical grounds, to a given object, brings analytically with it the commitment to the counterfactuals that are characteristic of fragility.

The Causal Efficacy of Macroscopic Dispositional Properties

121

do not make it into two states or two properties. This difference is the result of ‘a verbal distinction between the disposition and the state. (A verbal distinction that cuts no ontological ice).’39 According to Armstrong, ‘disposition’ is a concept that corresponds to a certain functional way of referring to properties or states, rather than to a particular type of properties or states. ‘Dispositions are marked off from (many) other states by the way they are identified’.40 So far, Armstrong expresses a position equivalent to our thesis (1). The controversial step in his reasoning is the following. Armstrong thinks that the only way to make sense of the idea that one property can be conceived both as dispositional and as categorical, is by supposing that the property is really (identical with) its microscopic reduction base, which he calls the ‘categorical basis’. However, Armstrong does not justify the implicit premise that only a microscopic property can be categorical and efficacious. ‘What then is the disposition, the brittleness? It is the “categorical basis”, the microstructure, but it is this property of the object picked out not via its intrinsic nature, but rather via its causal role in bringing about the manifestation’.41 Armstrong starts from Quine’s thesis that, by using a dispositional predicate, ‘we can refer to a hypothetical state or mechanism that we do not yet understand’.42 We can substitute a direct way to refer to it as soon as science has discovered the intrinsic nature of that state: for Quine, a disposition is ‘a partially discerned physical property that will be more fully identified, we hope, as science progresses’.43 However, Armstrong does not justify a second thesis implicitly presupposed by his reasoning, according to which the scientific property, discovered as the ‘intrinsic nature’ of the state that was provisionally conceived of in a dispositional way, is necessarily microscopic. In the case of brittleness, it is a property of the chemical bonds between the glass molecules.44 In the case of the disposition to transmit one’s hereditary characteristics, it is microscopic properties of DNA molecules. ‘A good model for the identity of brittleness with a certain microstructure of the brittle thing is the identity of genes with (sections of) DNA molecules. Genes are, by definition, those entities which play the primary causal role in the transmission and reproduction of hereditary characteristics. [...] in fact sections of DNA play that role. So genes are (identical with) sections of DNA.45 In 39 D.M. Armstrong, ‘Beliefs as States’, p. 419. 40 D.M. Armstrong, ibid. 41 D.M. Armstrong, ‘Place’s and Armstrong’s Views Compared and Contrasted’, in D.M. Armstrong, C.B. Martin and U.T. Place, Dispositions: A Debate (London, 1996), p. 39. 42 W.V. Quine, The Roots of Reference, p. 10. 43 Ibid., p.13. 44 Cf. D.M. Armstrong, A Materialist Theory of the Mind, p. 86; A World of States of Affairs, p. 73. 45 D.M. Armstrong, ‘Place’s and Armstrong’s Views Compared and Contrasted’, p. 39. Cf. also D.M. Armstrong, A Materialist Theory of the Mind, p. 90. A complication arises for Armstrong because he holds, on one hand, that the truth-making relation (between a proposition and a state of affairs) is necessary, so that if having a certain molecular structure makes true the attribution of the disposition to be brittle, it is necessary that all things with this molecular structure are brittle. On the other hand, Armstrong holds that laws are contingent, and that

122

Dispositions and Causal Powers

the case of dispositional mental properties, the underlying properties are microscopic properties of the brain.46 Armstrong does not offer any reasons for thinking that the categorical basis of a disposition is necessarily microscopic.47 However, the debate within which Armstrong develops his thesis suggests a hypothesis about its origin. In the text preceding the introduction of his thesis about dispositions, Armstrong criticizes Ryle48 and Price49 for two mistakes: the first is rejecting for a priori philosophical reasons the scientific research for the microscopic grounding of a given dispositional macroscopic property, in particular in the case of mental properties, and therefore to reject the perspective of a discovery that provides the means for constructing a microreduction of that property. The second mistake is the verificationist refusal to accept the existence of theoretical properties whose identity conditions are independent of single verification conditions or procedures. Armstrong’s reasoning would be valid if these two errors were just one. In that case, it would be natural to assume only one move is necessary and sufficient to avoid it; for indeed, the only way to avoid both mistakes in just one move consists in postulating a theoretical property underlying the disposition, which is at the same time the microscopic property providing the basis for its microreduction. Identifying the disposition with a theoretical microscopic property avoids the verificationist mistake because it is a theoretical property whose identity is independent of particular verification procedures; and it avoids the mistake of a priori rejecting the perspective of microreduction. However, there is in fact no reason to think that Ryle and Price commit only one error rather than two independent errors. To overcome the error of verificationism, the relation between having the molecular structure and breaking after falling depends on the laws. (This is one of the reasons why Prior, Pargetter and Jackson (1982) deny the identity of dispositions with their categorical basis. See above.) Therefore, Armstrong (1997) changes his position: now he says that the molecular structure together with the laws makes true the ascription of brittleness. 46 In A Materialist Theory of the Mind, pp. 76-77, Armstrong develops the analogy between those two identifications, with respect to their contingency. However, in A World of States of Affairs, p. 73, he explains that this contingency stems only from the contingency of the laws of nature responsible for the fact that the DNA molecules play, by virtue of their properties, the role of genes, and for the fact that the brain, by virtue of its properties, plays the roles that characterize mental states. Given the laws, the identification is necessary. 47 Armstrong is not alone in assuming this thesis without giving any reasons for it. Mackie (J.L. Mackie, ‘Dispositions, Grounds, and Causes’, Synthese 34 (1977): 361-370) characterizes his own ‘realist view’ of dispositions by saying that ‘there will always be a categorical ground’, and then immediately moving on to state that this categorical ground is necessarily different from the dispositional property itself: ‘This ground will not in itself be specifically dispositional’, his example being the categorical microproperties underlying the macroscopic disposition of solubility. ‘In crystalline sugar the feature causally relevant to its solubility in water will be something about the bonds between the molecules in the crystal structure’ (Mackie, ‘Dispositions, Grounds, and Causes’ p. 365). 48 G. Ryle, The Concept of Mind (London, 1949). 49 H.H. Price, Thinking and Experience (London, 1953).

The Causal Efficacy of Macroscopic Dispositional Properties

123

it is necessary and sufficient to postulate a theoretical property which is occurrent and categorical and which is not identified with a single (test condition, characteristic manifestation) – or (stimulus,response) – pair, but can contribute in many and complex ways to different causal processes. Nothing prevents the property thus postulated from being a macroscopic property belonging to the same object as the disposition: the person – and not some of its neurons or neuron circuits – in the case of mental properties, the body – not the DNA molecules – in the case of the capacity to transmit hereditary characteristics. It is therefore conceivable to remedy the error of verificationism, without at the same time overcoming the other mistake of rejecting a priori the perspective of microreduction. That second mistake can be avoided in a second and independent step. The microreduction of a macroproperty consists in the discovery of a nomological explanation of the possession of the macroproperty by an object, on the basis of the microproperties of its parts and their interactions by virtue of laws. Molecular biology can explain, on the basis of the numerous microscopic components of an animal’s body and their numerous and complex interactions, why that organism has the macroscopic capacity to transmit part of its hereditary characteristics to its offspring. However, this reduction does not justify the identification of that macroscopic capacity with any particular microscopic property. In particular, a number of authors have shown50 that this capacity of the organism cannot be identified with any microscopic property of the DNA molecules contained in its cells.51 50 Cf. P. Kitcher, ‘1953 and All That, A Tale of Two Sciences’, Philosophical Review, 93 (1984): 335-373; A. Rosenberg, The Structure of Biological Science (Cambridge, 1985); M. Morange, La part des gènes (Paris, 1998), trans. M. Cobb, The Misunderstood Gene (Cambridge, 2001). Armstrong is well aware that the identification of the gene with a segment of a DNA molecule is an oversimplification; however, he thinks that it does not threaten the coherence of his position: ‘The statement “The gene is the DNA molecule” is not a very exact one from the biological point of view. But it will prove to be a useful example in the development of the argument, and it is accurate enough for our purposes here.’ (Armstrong, A Materialist Theory of the Mind, p. 77). However, this does not seem to be a matter of neglecting some details: it is a fundamental mistake to take a microscopic property of a part of an organism for the causal basis of its disposition to transmit part of its hereditary characteristics. The causal basis of hereditary transmission consists in a complex mechanism of which DNA is only a part. However, the property of possessing that mechanism can only be attributed to the whole organism, but not to any of its microscopic parts, be it as important as the DNA. 51 P. Menzies, ‘Against Causal Reductionism’, Mind 97 (1988): 551-574. Menzies proposes another argument against the identification of a dispositional macroproperty with the underlying microproperty. Taking an example from David Lewis, he points out that the electrical and thermal conductivities of a metal are two different dispositions grounded in the same set of microproperties: the properties of the ‘free’ electrons of the metal, i.e. the electrons not chemically bound to individual atoms. Given the transitivity of identity, these dispositions cannot both be identical with their common reduction base, without being also identical to each other, which they are not. However, at a closer look, the reduction bases of these two dispositions are not exactly the same. In the model of their reduction proposed by

Dispositions and Causal Powers

124

Once we recognize that a microreduction requires two separate steps: first, the dispositional conception of a macroscopic property and second, the discovery of its microscopic reduction base, it becomes clear that the expressions ‘causal basis’ and ‘categorical basis’ are being used in this debate in two fundamentally different senses: according to the first, they designate the set of categorical properties underlying a disposition that are together causally responsible for its manifestations. Those properties may well be macroscopic in the sense of belonging to the same object as the disposition, rather than to one or some of its parts. The second sense is strongly suggested by the word ‘basis’: when a microreduction of the macrosopic property underlying a disposition has been found, one often calls ‘reduction base’ the set of microscopic properties of the parts of the object, which nomologically determine the macroscopic property. Armstrong’s mistake consists in confusing these two senses of ‘basis’, and to admit without justification that the categorical basis in the first sense of the term must necessarily also be the basis in the second sense, in which it means ‘microreduction basis’. A Solution to the ‘Problem of the Missing Reduction Base’ A positive argument for the existence of at least some categorical bases that are not reduction bases is the following: suppose that the hierarchy of levels of constitution Drude in 1900, the electrical conductivity σ and the thermal conductivity κ are determined by different properties of the free electrons: the electrical conductivity σ is determined by the microscopic properties n (the number of free electrons per cube centimetre), e (the unit electrical charge), τ (the relaxation time or mean free time of the free electrons, i.e. the mean time interval between two collisions) and m (the electron mass), according to the formula

σ=

ne 2 τ m

(N.W. Ashcroft and N.D. Mermin, (1976), Solid State Physics (Philadelphia, 1988), p. 7), whereas the thermal conductivity κ is determined by n, τ, m and T (temperature), according to the formula

κ=

3nτ k B2T 2m

(N.W. Ashcroft and N.D. Mermin, Solid State Physics, p. 23). Block is therefore right to note (in correspondence with Jackson) ‘that cases where different dispositions appear to have the same basis, and, more generally, cases where different functional roles appear to be occupied by the same state, turn out on examination, to involve subtly different bases and states’. F. Jackson, From Metaphysics to Ethics: A Defence of Conceptual Analysis (Oxford, 1998), p. 92, note 3. However, the case is compatible with our thesis that it is different properties of the various microscopic parts of the metal that nomologically determine the different macroproperties of the metal, by virtue of the various interactions between these properties.

The Causal Efficacy of Macroscopic Dispositional Properties

125

of objects is not infinite, but rather contains a level of properties and objects that are absolutely ‘atomic’ in the sense of not being themselves microreducible because its objects have no parts. In that case, the chain of microreductions also stops with the discovery of that fundamental level. A property M belonging to that absolutely fundamental level still gives, at least indirectly, causal powers to its bearers, else there would be no reason to postulate its existence. The powers, property M gives its bearer have a categorical basis: remember that I understand by the ‘categorical basis’ of the power, the property of its bearer that is causally responsible for the manifestations of the power. Now, given that M has no microreduction base, that categorical basis can only be the property M itself or some other property at the same level as M, in other words a property that is ‘macroscopic’ from the point of view of M. Our hypothesis that properties that are conceived in a dispositional way can also be conceived in a categorical way, suggests a simple solution to Molnar’s (1999) ‘problem of the missing reduction base’. Molnar convincingly shows the implausibility of three solutions to the problem that arises if the following two theses are true. 1. Every disposition has a categorical basis, understood as a microscopic reduction base. 2. There are absolutely structureless fundamental particles. The three implausible solutions are: a) to suppose that the causal basis of structureless absolutely atomic particles are global properties of the universe;52 b) to deny the second thesis, and to speculate that there is an infinite hierarchy of structures grounding even properties and objects appearing at some point as structureless and absolutely atomic; c) to suppose that the fundamental properties of the absolutely atomic particles are not dispositional. After having provided strong reasons for rejecting these three proposals, Molnar concludes that, ‘when it comes to the fundamental micro-entities, no suitable properties exist that could serve as a causal base of their dispositions’.53 Similarly, Mumford concludes that, in the case of absolutely fundamental properties for which no microreduction does exist, there is ‘just the one mode of characterizing it available to us: the dispositional’.54 Our distinction between two senses of the word ‘basis’ provides a means for avoiding this conclusion, which is paradoxical insofar as it implies that the manifestations of the powers due to absolutely fundamental properties have no cause whatsoever. This result can be avoided, in the absence of any microreduction, by 52 Harré has proposed the idea of such a grounding from above, or ‘ultra-grounding’. R. Harré, Varieties of Realism (Oxford, 1986), p. 295. 53 G. Molnar, ‘Are Dispositions Reducible?’, p. 17. 54 S. Mumford, Dispositions, p. 169.

126

Dispositions and Causal Powers

the hypothesis that the dispositional property of a fundamental particle is itself the categorical basis that is causally responsible for its manifestations.55 The Example of the Representation of Colours The perception of colours by the human visual system may illustrate the argument developed above. In psychology and psychophysics, representations of colours are conceived as macroscopic dispositional properties of persons: they give them the disposition to make judgements of similarity and distinguishability.56 These judgements constitute observation data whose theoretical explanation requires postulating the existence of a psychological space by means of which the subject represents colours. The logic of theory construction is essentially the same in psychology and in physics: the theory describes non-observable entities whose postulation allows us to explain a certain number of empirical regularities, and ultimately observable facts. Shepard57 has shown that such judgements of similarity58 contain enough constraints to determine,59 for any domain of perceived qualities,

55 It is common to call such hypothetical dispositions which cannot be micro-reduced, ‘ungrounded dispositions’ (e.g. Mumford, Dispositions, p. 167; Molnar, ‘Are Dispositions Reducible?’, p. 4). However, in the sense of the word ‘basis’, in which it designates the property of the bearer of the disposition that is causally responsible for its manifestations, it is clear that every disposition necessarily has a (categorical) basis. Therefore, to say that a disposition has no basis can only mean that it has not been (micro) reduced and hence does not have any basis in the sense of a reduction base. 56 Two colours are said to be distinguishable if a normal subject can systematically distinguish them. 57 R. Shepard, ‘The Analysis of Proximities: Multidimensional Scaling with an Unknown Distance Function’, Psychometrika, 27 (1962): 125-140 and 219-246. 58 Shepard uses only the order of similarity between pairs of stimuli presented to subjects, such as it appears in the judgements of the tested persons, without making use of any quantitative estimates the subjects make about the apparent distances between the stimuli. Apart from direct judgements of similarity, Shepard also uses data obtained by more indirect means that allow us to judge the proximity of the representations of stimuli in psychological space, such as the frequency with which the subjects confuse different stimuli, the delay required for discriminating two stimuli, or (for young children and animals) the size of the orientation reflex, when the first stimulus is replaced by the second. 59 Mathematically, Shepard’s algorithm uses two a priori constraints: 1) The function relating the apparent similarity to the proximity in the representation space is supposed to be monotonous. The monotony of the function guarantees that if two colours A and B are judged to be more similar than the colours C and D, the distance in psychological space, between the representations R(A) and R(B) of the former is smaller than the distance between the representations R(C) and R(D) of the latter. In the same systematic way, apparent similarity depends on the distances between the representations. In particular, the representations of the stimuli judged to be most similar must be separated by the smallest distance. 2) The psychological space has the smallest dimension that allows us to construct a monotonous and

The Causal Efficacy of Macroscopic Dispositional Properties

127

1. the minimal number of dimensions that the psychological space must have, in order to give the subject the means to represent the domain of stimuli under consideration, 2. a representation of the location of each perceived stimulus within a space isomorphic to the psychological space; this representation contains the coordinates of the different stimuli within the representation space. The algorithm developed by Shepard allows him to construct ‘maps’ of a certain number of psychological spaces corresponding to different domains of stimuli: the ‘proximity structure’ by means of which a subject represents, among others, the different facial expressions of its co-specifics,60 colours,61 the consonants of her native language, musical intervals, or familiar animals.62 In the case of colours, the first result that Shepard’s algorithm provides is the fact that two is the minimal number of dimensions that the psychological space must have in order to be able to represent the colours visible by the human perceptual system (abstracting away from the dimension of saturation). Any simpler psychological structure would be incompatible with the experimental findings. It is in particular impossible to account for the judgements of similarity between (perceptions of) colours on the basis of a psychological space of colour representation that has only one dimension: taking into account the judgements of the similarity of red and yellow, yellow and green, green and blue, and blue and violet, one might try to situate the representations of these colours in only one dimension, following the order of the rainbow. However, such a one-dimensional representation would not allow to account for the perceptual similarity between red and violet. If the representations of the colours were structured in the order of the rainbow along one dimension, red and violet would have to be the most dissimilar colours of all, whereas in fact they resemble each other much more than each of them resembles, e.g., green or yellow. Shepard’s second result is that there is only one topological structure or ‘map’ of represented colours that is related by a unique monotonic function to the judgments of similarity. On this map, the representations of the different visible colours are situated on a circle. Insofar as representations of colours are conceived as intermediaries between stimuli and judgements of similarity, they are dispositional properties. However, once we have overcome the verificationist prohibition that prevented Ryle from conceiving representations as entities independent of any particular manifestation, unique function relating apparent similarities and distances between representations in the psychological space. 60 R. Shepard, ‘The Analysis of Proximities: Multidimensional Scaling with an Unknown Distance Function’. 61 Ibid., and R. Shepard, ‘Approximation to Uniform Gradients of Generalization by Monotone Transformations of Scale’, in D.I. Mostofsky (ed.), Stimulus Generalization (Stanford, 1965), pp. 95-110. 62 R. Shepard, ‘Representation of Structure in Similarity Data: Problems and Prospects’, Psychometrika, 39 (1974): 373-421.

128

Dispositions and Causal Powers

we may consider representations of colours as theoretical entities that allow us to give causal explanations of similarity judgments. As theoretical entities that are not directly observable but are postulated in order to explain observable phenomena, they belong to the same category of entities as protons and neutrons whose postulation allows to construct causal explanations of, e.g. the phenomena observed when elementary particles interact in particle accelerators. The hypothesis of the existence of colour representations and of the structure of these representations in a psychological space is independent of the discovery of a reduction of these representations and of this space, to neurophysiological objects and properties. In an analogous way, the legitimacy of the postulate of the existence of protons and neutrons does not depend in any way on their eventual microreduction (to ‘quarks’), which is rather the object of independent research. The neurophysiological reduction of the representation of colours is a topic of intense research. According to a promising hypothesis, colour representations are determined by the simultaneous activation of neurons within a precise region of the visual cortex, an area called V1. Such reductions have already been accomplished in the case of a certain number of other psychological spaces, relative to the representation of certain perceptual domains in certain animals: the neural structure by means of which the barn owl (Tyto alba) represents the location of a sound source has been identified in the superior layer of the optic tectum of the brain of that bird.63 A neural structure situated in the auditory cortex of the moustache bat has been found to allow the animal to represent the position and velocity of objects in its surroundings by means of the mechanism of echolocation.64 Each time, the subjective sensation, which causes action or judgement, results from, or is nomologically determined by, the simultaneous activation of a large number of neurons situated in the region corresponding to the cognitive map of the relevant perceptual domain. It is conjectured that the representation produces its effects by means of a mechanism equivalent to the vectorial calculation of the mean activation, taken over all neurons of the map.65 The representation of a colour, say yellow, is determined by the neural activity distributed over the whole area. Certain directions on the cortical surface correspond to certain dimensions of the psychological space it determines. In spite of the distributed character of the activation underlying the representation of a given colour, it is possible to attribute a specific content to individual neurons: the neuron corresponding to a given shade of yellow is that whose isolated activation would produce the same sensation of yellow as the distributed activation grouped around it produces if the vector sum of the latter is equivalent to the vector corresponding to the activation of the first.66 63 Cf. C.R. Gallistel, The Organization of Learning (Cambridge, 1990), p. 478ff. 64 Ibid., p. 492ff. 65 C.R. Gallistel, The Organization of Learning, pp. 489, and 515; P.S. Churchland and T.J. Sejnowski, The Computational Brain (Cambridge, 1992), pp. 233-237. 66 It is well known that a large number of global distributions of neural activity can produce the same (type of) sensation: the stimuli producing such indistinguishable sensations

The Causal Efficacy of Macroscopic Dispositional Properties

129

The lesson I propose to draw from this sketch of results of psychological and neurophysiological research about colour representation is that it is coherent and conceptually possible to conceive of the categorical psychological property that produces the manifestations of a mental disposition (its ‘categorical basis’ in the first sense of the term) as a macroscopic theoretical property: as a property belonging to a person rather that to its microscopical parts, its neurons. The research for the microscopical, i.e. neurophysiological, basis of that macroscopic psychological property, which allows us to construct a reductive explanation of it, is a second and independent step. Two Objections Before concluding, we must consider two objections against our thesis that it is coherent and at least in the case of psychological properties also plausible, to suppose that the categorical basis that causes the manifestations of a dispositional property is a macroscopic property rather than its microscopic reduction base. The first objection challenges us to justify the crucial thesis that the categorical macroproperty is not identical to any microproperty. Maybe it is too simple to point out that the manifestations of a disposition of object O must be caused by properties of O, not by properties of O’s parts. Kim has proposed an interesting alternative to our conception according to which a given macroscopic property of a person, say of representing yellow (as seen at a certain point in the visual field), is determined by the interactions of the neural parts of that person’s brain, by virtue of laws of nature. According to Kim, there is a neural property that is in a sense ‘microscopic’ and that does all the ‘causal work’ we attribute to the psychological property. This property belongs to the category of what he calls ‘micro-based properties’, which I shall call ‘micro-based macroproperties’ (in brief, MMP), defined as ‘the property of being completely decomposable into nonoverlapping proper parts, a1, a2, …, an, such that P1(a1), P2(a2),…, Pn(an), and R(a1, a2,…,an)’.67 This concept allows us to attribute to the whole the property of having parts, each of which has its own properties and relations to the others. Thus it points towards the elaboration of the concept of a causally efficacious macroproperty. However, the conditions imposed by Kim on the concept of an MMP are too weak to guarantee that a given property is efficacious at all. It is easy to find MMP that obey Kim’s constraints but have no causal efficacy whatsoever. Take an arbitrary mereological sum whose elements do not physically interact in any way. The mereological sum whose elements are the electrons of billiard ball A and the atomic nuclei of billiard ball B does not have the causal powers of a billiard ball. The existence of an MMP as defined by Kim is a logical consequence of the existence of the ‘parts’ (of the mereological whole to which the MMP is attributed), whereas the existence of a whole possessing its own causal powers depends on the existence of appropriate interactions between those parts. However, are called ‘metamers’. 67 J. Kim, Mind in a Physical World, p. 84.

130

Dispositions and Causal Powers

Kim’s condition does not impose any constraint on the relations R between the parts; in particular, it does not require that they correspond to any physical interactions. Here is another way of showing that the conditions Kim imposes on MMP are not sufficient to guarantee that each so conceived MMP is causally efficacious. A given whole has a different MMP for each of its possible decompositions in parts. However, an indeterminate and perhaps infinite number of such decompositions does not give rise to an equivalent number of different causally efficacious properties. Inversely, an object that can only be decomposed in one natural way can have different causal powers, by virtue of different interactions between the various properties of its parts. A hydrogen molecule H2 whose only natural decomposition consists in splitting it into two H atoms nevertheless possesses several different causal powers, such as its magnetic momentum and its base frequency of oscillation. A person who sees a yellow spot at a certain point in her visual field possesses an MMP by virtue of her neurons and their states of activation. However, even if these neurons with their respective activities are in their ‘normal’ spatial relations (i.e. in conditions appropriate for bringing about the mental state of perceiving a yellow spot at that point), if you prevent them from interacting, every mental property will disappear. This shows that it is necessary to enrich Kim’s concept by taking into account the various nomological interactions between the parts: it is these interactions that determine the existence of a real whole and its various causal powers. According to the second objection, the thesis of the causal efficacy of the categorical macroproperty is refuted by the ‘principle of causal-explanatory exclusion’ according to which ‘two or more complete and independent explanations of the same event or phenomenon cannot coexist’.68 This refutation is based on the hypothesis we have taken for granted all along, according to which it is always in principle possible to discover a microreduction of a given macroproperty. But then, the microproperties of which the reduction shows that they are underlying the macroproperties, seem to monopolize causal efficacy and to condemn the macroproperty itself to the shadowy existence of a mere epiphenomenon. Prior, Pargetter and Jackson who understand ‘causal basis’ as meaning ‘reduction base’, use the exclusion principle to show that the macroproperty itself (also conceived as dispositional) is epiphenomenal: ‘Any disposition (and thus fragility) must have a causal basis. This causal basis is a sufficient causal explanation of the breaking as far as the properties of the object are concerned. But then there is nothing left for any other properties of the object to do. By the Distinctness thesis, the disposition is one of these other properties, ergo the disposition does nothing’.69 The principle of causal-explanatory exclusion is plausible insofar as it is equivalent to the supposition that there is no general overdetermination of a given type of effect by two independent causes. However, in the form Kim gives it, the principle does not exclude the possibility that two sets of properties, which are instantiated at the same time and place but which are not independent because one set of properties nomologically determines the other, are both sufficient for the same 68 J. Kim, Supervenience and Mind (Cambridge, 1993), p. 250. 69 Prior, Pargetter and Jackson, ‘Three Theses about Dispositions’, p. 255, italics theirs.

The Causal Efficacy of Macroscopic Dispositional Properties

131

effect. For this reason, there is no valid argument based on the principle of causalexplanatory exclusion that allows for the conclusion that a macroproperty P of an object O is epiphenomenal, simply because there is a microreduction explaining the presence of P from the properties P1, P2, … of the microscopic parts of O and their interactions. The result of a microreduction is precisely to establish that and to show how the microproperties nomologically determine the macroproperty. Consider the power (or the disposition) of person A to transmit part of her hereditary characteristics to her children. Suppose the molecular mechanism underlying this transmission has been completely discovered and has led to the construction of a reductive explanation of this disposition in terms of microproperties. Does this mean that the only properties that are causally efficacious in the transmission of the hereditary characteristics of A are the properties of her microscopic parts, in particular of the DNA molecules contained in her germinal cells? No, for the simple reason that A’s DNA molecules are not directly causally responsible for the manifestation of the disposition: the apparition of some of A’s phenotypic traits in her children. This effect is only brought about by the mediation of a complex set of interactions between the DNA molecules and a large number of other parts of the organism. No property of any part of A can be said to be in itself the basis of the transmission, insofar as the causal basis is an intrinsic property of the object possessing the disposition, which is, in favourable circumstances, sufficient to produce the manifestation. We have already seen that the logical conjunction of the set of microproperties (the MMP) that intervene in the determination of the power is not the causal basis either: it is only by virtue of their nomological interactions that the set of microproperties of its parts gives A’s organism the disposition to transmit her hereditary characteristics. The micro- and macroproperties do not compete for being the property that is causally responsible for the manifestation of the disposition. Both participate in bringing it about, though in quite different ways: the microproperties of A’s parts provide A, by virtue of a relation of nomological determination itself grounded on the nomological interactions among the parts, with a global property that is first conceived of as the disposition to transmit hereditary characteristics. That very global macroproperty can also be conceived of as categorical insofar as it is directly causally responsible for the transmission. Conclusion I have tried to show that macroscopic properties such as a vase’s property of being fragile or my current property of having the intention to write the word ‘disposition’ can be causally efficacious in bringing about their manifestations although they can also be conceived of as dispositional properties. The defence of this thesis requires arguing that the dispositional-categorical distinction applies first to predicates and concepts expressed by these predicates and only indirectly to the properties to which the predicates refer and which are conceived by these concepts. If this is correct then a dispositional and a categorical predicate can designate one property. We have

132

Dispositions and Causal Powers

seen that this conception provides the means for refuting a number of traditional objections against the efficacy of dispositional properties and for escaping what we have called the epiphenomenalist trilemma with respect to macroscopic properties. According to the major theories defended at present, such properties are either epiphenomenal and thus causally inert or efficacious only by being identical with microscopic properties, which also constitute their reduction base. I have shown that it is coherent and plausible to consider the dispositional macroscopic property itself as causally responsible for its manifestation, thereby showing that this property is also capable of a categorical conceptualization. Its reduction is the object of an independent enterprise; however, the construction of such a reduction does not justify the identification of the reduced property with the reducing property. I have also shown that acknowledging the causal efficacy of macroproperties does not lead to an unacceptable overdetermination of their effects: the microproperties in the reduction base cause these effects only indirectly, by lawfully determining the existence of the macroscopic property. This provides a justification for following the intuition that my present act of typing the word ‘disposition’ has been caused by my decision to do so, this decision being a macroscopic mental property that is not identical with any microproperty of my brain.70

70 I thank my auditors in Lund and London, and Anouk Barberousse, Tim Crane, Joan Cullen, Mauro Dorato, Ludger Jansen, Michael Martin and Jürgen Schröder for helpful comments on earlier versions of this chapter.

Chapter 5

Opium’s Virtus Dormitiva Cyrille Michon

In his famous mixed Latin and French verses, Molière once and for all discredited the medieval notion of a causal power. To the Doctor asking causam et rationem quare/Opium facit dormire, the Bachilierus answers: A quoi respondeo,/Quia est in eo/Virtus dormitiva,/Cujus est natura/Sensus assoupire.1 It became frequent at the time to ironize on the aperitive power of a key, or on the attractive power of a hook (Pascal). The basic sense of the critique is clear: to ascribe a causal power, described only by its results, is of no worth, and the claim to give thereby a scientific explanation seems somewhat obscurantist. Such a vis is hidden, and you can go easily from there to its illusory nature. We face two criticisms, to which a defender of virtues or powers has to address. She should show firstly that there is no illusion in the reference to powers, and then how useful it can be. A third element might be present to a greater or lesser degree in both criticisms: the accusation of anthropomorphism, later generalized to all causal vocabulary.2 It is often said that the origin of the notion is psychological, and/or moral, like the Greek word aitia which has been extended from the domain of legal responsibility to the idea of natural lawfulness. The concept of power would come from introspection, especially from the experience of trials and efforts. It would then have been projected from the mind onto the things in the world, giving way to illusory or even superstitious discourse about active powers. A well done defence of the notions of action, power and cause should also address the accusation of anthropomorphism. The Molieresque objection is relevant: nothing has been explained when it has been stated that the opium smoker fell asleep because of its dormitive power. At least, if we understand that this meant: opium’s dormitive power is the reason why opium made him fall asleep. That would be viciously circular. But it would be so, 1 Malade imaginaire, III intermède: ‘Primus doctor: Si mihi licenciam dat Dominus Praeses/Et tanti docti Doctores,/Et assistantes illustres,/Très sçavanti Bacheliero,/Quem estimo et honoro,/Domandabo causam et rationem quare/Opium facit dormire. Bachelierus: Mihi a docto Doctore/Domandatur causam et rationem quare/Opium facit dormire:/A quoi respondeo,/Quia est in eo/Virtus dormitiva,/Cujus est natura/Sensus assoupire.’ 2 This is an old criticism, taken up by Russell in his famous paper ‘On the Notion of Cause’. B. Russell, ‘On the Notion of Cause’, Proceedings of the Aristotelian Society, 13 (1912): 1-26; repr. Mysticism and Logic (Harmondsworth, 1953), pp. 171-196.

134

Dispositions and Causal Powers

because saying that opium makes one sleep is already to ascribe a dormitive power. Such an attribution is not circular if it means that the reason why the opium smoker fell asleep is to be sought in opium, and not in any other substance he has absorbed, or in any other circumstances. It is not yet an explanation, or it might be a vacuous one, but this does not show that the notion of a causal power is vacuous. If it were, we would have eliminated it from our everyday discourse. But we have not, and we find it useful to indicate that some substances (medicines, food) are ‘somniferous’ which means nothing else. Such a term cannot be explanatory if the question to be answered is ‘why does this make one sleep?’, but it could be, if the question were: ‘why is it dangerous to take opium before driving?’–‘because it can make you fall asleep = because it is a somniferous substance = because it has a dormitive virtue’.3 We asked the question ‘Why?’. But it could be said that this is not the kind of question we ask about a natural phenomenon. What is behind this remark is the idea that scientific explanation proceeds by discovering the laws by which the particular phenomenon to be explained can be covered, or/and predicted. To explain it would be to show that it is a particular case of one or more laws, given some initial conditions.4 Here, I must answer that the feeling of dissatisfaction felt by the critics of dormitive virtues has changed sides: how the very fact that this phenomenon is a particular case of such laws can be an explanation of it? That seems strange. Nonetheless, this is the common view, which I will first address. I then intend to show that a scientific explanation of nature requires the notion of causal power, and finally suggest some elements for the analysis of the notion.

3 See here, the introduction by Gnassounou and Kistler. S. Mumford, Dispositions (Oxford, 1998), sec. 6.6, pp. 136-138, and for an historical perspective, K. Hutchison, ‘Dormitive Virtues, Scholastic Qualities, and the New Philosophies’, History of Science, vol. 29, no. 85 (1991): 245-278. 4 I take it to be the Humean view, eventually ameliorated by the ‘nomological-deductive’ model of Hempel, or the theory of counterfactuals. It is often called the regularist or the MillRamsey-Lewis view (see J. Earman, ‘In Defence of Laws: Reflections on Bas van Fraassen’s Laws and Symmetries’, Philosophy and Phenomenological Research, 53 (1993): 413-419), and is opposed to a realist conception of law of nature, grounded on universals (Dretske, Tooley, Armstrong), or on transfers of energy and individual properties or tropes, see M. Kistler, Causalité et lois de la nature (Paris, 1999). I cannot take a comprehensive view, and refer to M. Kistler, ‘Causation in contemporary analytical philosophy’, Quaestio, 2 (2002): 635-668, J. Schaffer, ‘The Metaphysics of Causation’, Stanford Encyclopedia of Philosophy (2003), (http://plato.stanford.edu/entries/causation-metaphysics/) and E. Sosa and M. Tooley (eds), Causation (Oxford, 1993). Regularists do not admit real differences between accidental regularities and nomological ones: nomological regularities are only those that can be more strictly formulated (by way of counterfactuals) and/or deduced from a theoretical and well confirmed background.

Opium’s Virtus Dormitiva

135

Laws and Explanation From Regularity to Law Since Hume, at least, it has been assumed by many philosophers that we cannot ascribe any active power to natural bodies, nor observe or infer any connection between them or between natural changes. There is no contradiction in principle if a very well confirmed regularity, as that of sunrises and sunsets, breaks down some day. At the same time we take those regularities for granted, and must do so in order to build up physical astronomy. They can be very precisely described, and even deduced, by mathematical procedures, and seem to be the basis of science. Science formulates laws, and laws explain natural phenomena. But different conceptions of law-like regularities might be at work. a. It would not be helpful to stick only to the regularities observed until now: they allow us only to say what has happened, but we cannot infer anything concerning the future, and we have no assurance that they are not mere coincidences. We would like to distinguish between the regularity observed in the motion of planets, and others like the recurrence of certain sport results. b. We might wish to extend the regularities to unknown time and space, to make them bear on what happened, is happening, and will happen. But how can we know the future, and who will say that a misleading coincidence will not still be at work? Sure, the statement of a law can be taken only as an hypothesis, submited to tests, but, if the moas are all dead before reaching the age of 50, someone could have stated the regularity as an absolute one before the extinction, without reaching thereby a biological law. We need more than an omnitemporal truth, which can attract our attention, but is not, as such, a law of nature. c. A law says what must be, sciences give us nomological statements, whether we believe in real laws or not: it is a law that planets follow a (more or less) elliptical trajectory, and we can predict, on the basis of such a law – and others to which one will add conditions on positions, mass, velocity – the position of each one in every moment in the future.5 True, we know a lot of exceptions, for many well accepted laws. Not every body falls down, drawn by the earth’s gravity: other laws, obstacles, can intervene. We should mention the absence of every opposing circumstance: each time I turn the key the door opens, unless it is locked from inside. Contrary to good logical modus ponens, the causal syllogism is defeasible, it conforms to nonmonotonic logic: the addition of a premiss can invalidate the consequence. In pure deductive logic, once a set of premisses leads to a certain conclusion, no addition can change the validity of the inference. In order to reach such a high standard with law-like statements, we 5

F. Dretske (1977), ‘Laws of Nature’, Philosophy of Science, 44 (1977): 248-268.

136

Dispositions and Causal Powers

would have to mention that nothing impedes the consequence otherwise admitted as a particular case of a general rule. But this task – the enumeration of all possible obstacles – is simply unrealizable.6 Prediction and Explanation I have no objection to laws, nor to law-like statements produced by science: they are useful, allow predictions, measure, and most of our technical instruments. But we have to distinguish prediction and explanation: Kepler’s laws describe the motions of the planets, which are in turn (partly) explained by the theory of gravitation. Florence fountainers remarked and stated that water could not rise more than 10,33 metres in a vacuum pump, but this law was explained, and with it the particular cases, when Torricelli discovered atmospheric pressure. Why should the presentation of a phenomenon as a particular case of a law explain this very phenomenon? And why should we think that the notion of a cause could be assimilated to that of the ultimate condition, all others being given, considered as the antecedent of a particular application of the nomological conditional? It is more often when the law is not satisfied that we look for the cause: we look for an explanation of the irregularity.7 Instead of saying that the physical statement ‘the melting point of W is …’ is equivalent to a nomological statement of the form ‘each time a sample of X is raised to the temperature of …, it melts’, we should say that it is equivalent to ‘if a sample of X does not melt at temperature …, there must be a cause’. Why does the car not start when all conditions are fulfilled: fuel, unlocking, ignition, and when it has always been the case before? This example is not meant to show that there is no covering law (maybe is it the cold weather), but that it is a debatable point that the notion of a cause should derive from that of a universal (causal) law. We can question the view on two grounds: a. on the direction of derivation, from law to cause, when it might be the case that we go from particular causal connections to the statement of a general law;8 b. on the universal possibility of associating a law to a singular causal connection.9 All this, even though a regularity may often be the sign of a causal, regular, lawful, connection.

6 See G.E.M. Anscombe, ‘Causality and Determination’, in E. Sosa, and M. Tooley (eds), Causation (Oxford, 1993), pp. 88-104, and more recently N. Cartwright, How the Laws of Physics Lie? (Oxford, 1983), M. Kistler, ‘Laws of Nature, Exceptions and Tropes’, Philosophia scientiae, 7 (2003): 189-219. 7 Anscombe, ibid. 8 This non-Humean view was first formulated explicitly by C.J. Ducasse, ‘On the Nature and Observability of the Causal Relation’, in E. Sosa, and M. Tooley (eds), Causation (Oxford, 1993), pp. 125-136, often quoted for this point. 9 This has an important bearing on the question of determinism, as Anscombe wanted to show. But I cannot pursue the question here.

Opium’s Virtus Dormitiva

137

In any case, to derive a particular case by deduction from, or subsumption under, a law is not to explain it. We want to know why and how the effect happened, what is the connection, the mechanism driving it, and responsible for it. Sure, these are moral and anthropomorphic terms, but my point is that it has not been explained why Tintin fell asleep after taking opium, when one has said that it is a law that each time one takes opium, one falls asleep. This would have been sufficient to predict Tintin’s sleep, and one would not have been surprised by that fact. Empiricist epistemology tells us that scientific explanation consists in giving a law or laws linking the explanandum and some other fact(s). But those facts along with those laws do not constitute an explanans: the explanandum is still explanandum (to be explained). Nomological covering of causes and effects allows us to expect the prediction, or to test an explanation, if one has been given, and allows such a nomological statement. Boyle-Mariotte’s law does not explain why a volume of gas reaches a certain temperature when submitted to a certain pressure: it only links the three parameters, and affords the predictibity of one when two are given. If snow melting in Spring is followed by rivers flooding, the Mississipi will rise. But no explanation of the phenomenon has been given (although a very simple one is almost evident: the snow melting adds water to the river …). If eating deteriorated food is followed in 70 per cent of cases by illness, we could say that my friend’s illness was unfortunately predictable and even probable. But this is no explanation. One should not be surprised, and would be even able to predict, that each time Ali Baba utters ‘Open Sesame’ the rock-door opens: there is a law, but we do not understand. In order to understand this strange phenomenon, as any magical trick (the rabbit out of the hat), one would like to know how it was done: what is the mechanism at work. To say that opium is somniferous drives us towards an answer, it selects a substance and says that this substance has a causal power identified by such a result. We are ready to analyse the substance and to understand how it interacts with the nervous system in such a way that sleep is produced. To ask only for laws as a substitute for explanation is to accept a magical conception of nature. Where is obscurantism? We could understand the resistance of the D-N model (for deductive nomological, after Hempel), because of the widely accepted idea that an explanation in natural sciences, is not an explanation in moral or human sciences. We do not explain human actions as we do for natural processes. If we kept the word ‘explanation’ for the human realm, we should say that we do something else in natural sciences. This has been the motive for the distinction between ‘erklären’ and ‘verstehen’. It is assumed that way that JFK’s assassination is less explained by subsumption under a law associating to each US president in such and such circumstances such and such a death, than by the reasons that have led someone or some ones to act that way. Psychological explanation is not to be extended to nature, where we look only for laws. Hermeneutics seems to accept the epistemology of empiricism for natural sciences (but refuses to extend it, as did

138

Dispositions and Causal Powers

many empiricists, to human sciences).10 But to refuse the extension of the ideas of ‘motives’, ‘responsibility’, ‘intention’ to natural processes, is not to refuse the idea of natural agents. One could accept, without any hint of psychologism that it is not a law that explains the Mississipi flood, but the fact that snow melting adds water to the river. This is the mechanism, summarized by the action verb, nomological or not, and there is no psychology attributed to snow. Three Conceptual Obstacles In order to give a precise diagnosis, it seems to be necessary to take into account three features of the incriminated doctrine of explanation by laws. They are linked together though each one can be accepted independently of the rest. Atomistic Ontology of Events The identification of causes by way of laws, even in its more realistic versions, opts in general for an ontology of events.11 This political event has led to this electoral result. This throwing of a stone is the cause of this window breaking, etc. It seems amply justified to ground the causal explanation between facts on the causal relation between events: ‘p because of q’. We could have not moved so easily from objects to propositions. Propositions describe events, adding circumstances which make us able to select such and such a law or laws, linking them together: John’s fall caused John’s death. This will be easily covered by a law, if it is added that John fell from the 50th storey. But whatever the description, it is always the event A that caused the event B.12 I am not eager to oppose an ontology of events and an ontology of objects: one can consider the world as full (or not so) of objects, which exist, and constituted of events, which happen. Objects take part in events. Events presuppose objects. This being so, we should note that the preference accorded to the ontology of events is not neutral for the analysis of causality. Our language uses both ontologies, and we are sensible to grammatical differences: an object has properties (colour, smell, age, physical constitution) that cannot be ascribed to an event, it has specific

10 This is well shown by V. Descombes, La denrée mentale (Paris, 1995) trans. The Mind’s Provisions (Princeton, 2001), pp. 61-64. 11 Davidson has forcefully argued for this, which is now not a common but very widespread view. See E. Sosa and M. Tooley (eds), Causation (Oxford, 1993). As Max Kistler pointed out to me, if causes are identified from laws, they must be facts (it is a law that if it is a fact that p, then it is a fact that q), which are general and have only the properties signified by the propositions that express, whereas events are particulars (events tokens) with an indefinite number of properties. But the classical view does not make the distinction, and it is true that the ontology of events is well associated with that of causality with lawfulness: an event C is the cause of an event E, if there is a law linking them under some descriptions (between the facts so expressed). 12 I am following Davidson in conceiving the causal relation between events as extensional, and the causal statement, linking descriptions of events, as intensional.

Opium’s Virtus Dormitiva

139

dimensions whereas the link between the event and space is looser. The ontology of events allows for the expression of logical relations between the propositions describing them (negation, exclusion, entailment), which we could not find for objects. It also allows for the expression of spatio-temporal relations (before, after, simultaneously, somewhere else) distinct from those connecting objects. But there is certainly no causal power, nor action, that can be ascribed to an event. Last, this ontology of events is best seen as a kind of atomism. Two events are in principle independent, because of the non-contradiction of the occurrence of one without the other, and because one can think of time as discrete, each instant being logically independent on the immediately preceding and following ones. This is a cinematographic conception of the world, which is largely responsible for the dissociation of events and objects, when an ontology of quadridimensional objects has not been preferred to spatio-temporal and logical relations. This atomism seems to be a kind of logicist prejudice (the only connections are logical ones), pretending to build the world with independent bricks and no other cement than causal regularities. Impressionist Epistemology The associated epistemology has been, with Hume, that of a flux of sensible impressions at the basis of the formation of all ideas, so making unintelligible the idea that we could perceive the causal influence of the flame on the finger, or that of a ball in motion on another. Thus limiting visual perception to instantaneous impressions, one should also question the idea that we really perceive a flame or a finger or a ball. Impressions are of colour patches, phases, whose designation would be impossible: our language is not made for them. This epistemology has failed, and it would be better to say that we observe what we can describe by way of terms whose learning has been directly linked to observation, as it is the case for ‘finger’, ‘flame’, ‘ball’: we say that we see them, that we see (hear, smell) what we can so describe. But the case is similar with different verbs such as ‘to drink’, ‘to blow’, ‘to cut’, describing certain events or actions, causal processes. This is not to say that all causal efficacy is observable: the action of a virus is hypothesized, inferred, tested. But other kinds of natural action are literally seen: the tempest bends the trees, the butcher cuts the meat, and so on. We have learnt to so describe what we observe.13 The objection coming from the unobservability of causality is grounded on disputable prejudices.14

13 C.J. Ducasse, ‘On the Nature and Observability of the Causal Relation’, was also one of the first to criticize this fundamental thesis of empiricism. The argument presented here has roots in Wittgenstein, and is made in different ways by Anscombe, Strawson and Austin. 14 R. Harré, The Principles of Scientific Thinking (London, 1970); R. Harré and E.H. Madden, Causal Powers (Oxford: 1975) considered as three myths the two preceding theses: atomism, empiricim with extensionalism of the causality relation. This last one has been subjected to doubt above, I think it is more implicit than the two others.

140

Dispositions and Causal Powers

One can agree that events are observable, but it is not so clear that a causal relation between two events is observable. Event-causation is not what we observe and describe by our observation causal statements. In other terms, if the logical form of a causal statement is ‘event A caused event B’, this statement is not an observation one. On cannot observe that Brutus’ stab (and others if it was not sufficient) caused Caesar’s death. The Logical Form of a Causal Statement But not every causal statement has the previous logical form, even beneath its superficial grammatical structure. Once again the philosopher imposes the label ‘causality’ as a genus, a single essence, in order to criticize or to defend it. Books, chapters, manuals, encyclopaedias and the greatest achievements in the history of philosophy have made us used to it. It is a ‘term of the art’. But, even if the substantive ‘cause’ and the corresponding verb do have a current use, they do not introduce, as it is the case with the abstract noun ‘causality’, the idea of a common characteristic, an essence of the causal connection. Causal expressions are omnipresent in our descriptions of the world: we eat French fries, cut papers, shut the door, start the car, break glasses, and so on. Not every transitive verb expresses an action, but every causal verb is transitive, there are many of them, and they are not limited to human activities. Can we do without saying that the wind is blowing, that the cat is hunting the mouse, or that the falling box crushed the fly?15 The ontology of events constrains radically our means of expression, and does not give an account of causal verbs as they are applied to natural individual agents. The notion of cause, expressed by a great variety of verbs, is not univocal, rather one would like to say it is analogical, somewhat like a family resemblance term. Again: the choice of an ontology of events forbids, for grammatical reasons, the use of capacity expressions.16 This is without doubt excluded when natural actions are. Once the idea of natural action has been eliminated, those expressions are pointless. Agency is thus limited to human beings, if not purely eliminated. The prejudice here is that nature is the world of what happens – only human beings really act. But we do speak of the action of the wind, of an acid, of a washing machine, of an explosive or of a fertilizer. As with causal vocabulary, action vocabulary is much richer than the word itself. There certainly is a specificity of human actions, but one should not turn the species into a genus, and contradict so heavily our ordinary language and the classical distinction between the actio hominis, the action of a

15 G.E.M. Anscombe, ‘Causality and Determination’, in E. Sosa, and M. Tooley (eds), Causation (Oxford, 1993), pp. 88-104, and P.F. Strawson, ‘Causality and Explanation’, in B. Vermazen and M. Hintikka (eds), Essays on Davidson: Actions and Events (Oxford, 1985), pp. 115-136 have underlined the fact that many verbs and notions are unified under the unique label ‘cause’, ‘causality’, ‘causation’, and that an undifferentiated criticism of it may be misleading. 16 On the notions of ‘object’ and ‘event’, see P.M.S. Hacker, ‘Events and Objects in Space and Time’, Mind, 91 (1982): 1-19.

Opium’s Virtus Dormitiva

141

human being considered only as any other natural agent, and the actus humanus, for which one is taken as (morally) responsible.17 Natural Action and Scientific Explanation The Idea of Natural Action The idea of an action includes that of a connection: the logical connection between action and passion. To say that Brutus killed Cesar is to say that his stabbing and Caesar’s death or mortal injury are one and the same event. What is missing in the event’s description is the link that affords understanding: Brutus’s stab is a stab in Caesar’s body. We can describe this event as an action by Brutus or as a passion of Caesar. Speaking of action I am coming to the pars construens: the virtus dormitiva is nothing other than the power to act, and action is what reveals the power. What happens in nature is not the same thing as what we can describe. The term ‘fact’ is ambiguous, and can serve to indicate both. It is a fact that Bush is called ‘Bush’, that he is president, not Chinese, lives in the White House, has children, etc., but these facts do not happen, are not produced, as does rain falling in Dublin, or Italy’s last victory in the World Cup, which can also be called facts, although ‘event’ has been the term chosen until now. But ‘event’ is less precise. We should distinguish between events and the special kind of operations. Still, not every operation has an agent, and some are expressed by impersonal verbs: it is raining, snowing, breezing. But generally, the distinction between what I do, or what X does, and what happens to X, is one of our basic ways of describing the world. Many events are actions, in the sense in which one can ascribe them to one or more agents, as to one or more patients, so that the same operation can be described in the passive voice: X cuts the bread – the bread is cut by X, the fire burns the finger – the finger is burnt by the fire, etc. Those expressions are correlative, they are two descriptions of the same reality, as the road from Athens to Thebes and that from Thebes to Athens is the same one, described both ways. Burning is an action when described in relation to the fire, a passion when described in relation to the finger. Fire and finger are distinct realities, but the action-passion is one and the same. As a symptom of this identity: there is one localization, usually that of the patient: the burning is in the hand, as the wound is in Caesar.18 Let us now have a look on actions which are undoubtedly natural. They are those events, changes, that relate to one or more individuals, because of a change produced in them, internally, and which are so ascribed because of their structure and of the exterior circumstances that prompted them. The quest for an explanation

17 See V. Descombes, ‘L’action’, in D. Kambouchner, Notions de Philosophie, II (Paris, 1996). 18 The remark comes from Aristotle, but it had also been taken up by Descartes, On passions, art. 1.

142

Dispositions and Causal Powers

of the observed phenomenon begins by recognizing an action, leading to the agent or the agents, and to the study of their internal structure. Some mechanisms are directly observable and do not require any particular research: the blowing of leaves by the wind, the pulling of an object by a rope, the breaking of a pane of glass by a ball, and so on. There is no inference from the action to the power and the agent, and then to the mechanism of its interaction with the world: everything is given, unless one has only observed a later effect of the action, and wants to know the cause. The action of a virus on an organism, on the contrary, is inferred by way of hypothesis, analogically with natural actions which are better known. There is a level of analysis in theoretical physics where mathematical equations do not let any more room for questions about causes: but this does not mean that the level reached is ultimate, nor devoid of causal powers. It shows that mathematical representation has no real explanatory force. Physics works with mechanical analogies which often allow, besides the mathematical apparatus, a representation of the phenomena to which causal vocabulary applies. It has to be questioned whether theoretical physics, or celestial mechanics, has to be taken as the archetype of scientific knowledge. Bacteriology, when it isolates a new virus, even if it formulates less strict laws than physics, is more explanatory than are Kepler’s laws, or Maxwell’s equations. Physics is less explanatory when formulating a law than when it discovers a new entity, planet, galaxy or particle, which are identified by their causal powers. The Object of Scientific Research I would like to distinguish two points. Firstly, scientific activity is focused on the finding not so much of laws than of entities and powers. Secondly, observed regularities and law-like statements are often indicators of such entities and powers. Both points are linked together. In his philosophy of science textbook, Hempel told the story of Ignac Semmelweis, in Vienna, 1844.19 As an obstetrician, he observed in one of the two maternity units of the same hospital a mortality rate much higher than the standard one. Many explanatory hypotheses were put to test, until a correlation was made between the unit with a high mortality rate and the previous examination, in this same unit, of dead bodies by medical students who were then going to women in child-birth, without washing their hands, and so bringing with them cadaveric matter. Hygienic measures showed immediately a correlation in a reduction of the mortality rate. The discovery of the cause led to that of asepsis. The cause was isolated by finding a correlation. But the mechanism of infection had to be understood as the introduction of cadaveric matter on the hands of the surgeons into the bodies of the women, and a further step was understanding infection inside bodies, and so on. Initial regularity oriented the search for a cause, and after many unfruitful hypotheses, led to a causal power in cadaveric matter, which had still to be studied.

19 C.G. Hempel, Philosophy of Natural Science (Englewood Cliffs, 1966).

Opium’s Virtus Dormitiva

143

The identification of a responsible agent is associated to that of power: we observe an action, an operation, we look for an explanation, what produces this effect, the cause. This is hidden by an analysis in terms of event causation. In their book Causal Powers, to which this description owes a lot, Harré and Madden discuss an example taken from Davidson.20 A girl Flora drying herself after a swim in the sea, discovers ugly red patches on her skin. We can specify that the event took place on a Sunday, that the weather was fine, that the towel was rough, but the event-cause is still the same. One could say only ‘the cause of the appearance of red patches caused the red patches’. But it is evident that the description provides some irrelevant information and one clue to why Flora’s drying herself has had this effect: this element is the causal power of the roughness of the towel. When the relevant property has been stated, the causal power is indicated, and we understand why and how A is the cause of B. Until this occurs, the effect and its regularity can attract our attention, but it does not allow identification of a cause. To say that the towel was rough is to indicate the causal power of an agent, acting only in presence of a passive power of irritation, irritability being a property of human skin. We need an ontology of properties and powers, and not only of events. In order to understand how the roughness of the towel provoked the irritation of the skin, we need to study the structure of both, and the mechanism of irritation due to both. The path from the identification of a power to the search for its mechanism is essential to scientific research. To ascribe a vis dormitiva to opium, a causal power to an individual or to a kind, is not to explain an observed change, but it is to indicate that the change is due to such and such an agent, of such and such a kind, with such and such property, and that the explanation is to be found in the nature of the kind or of the property. This discovery of the virtue allows the selection, the classification and the identification of natures. It does not substitute itself for the analysis of the nature, and of the mechanism involved in producing the change. The discovery and presentation of the mechanism is what really does the job of explanation, even if it is only hypothetical. But both steps, classification by causal powers, explanation by analysis of the mechanism, are well illustrated by the combination of pharmacology and chemistry. The first discipline identifies natures by their causal powers, the second one studies those natures and explains their powers by the underlying mechanisms. Scientific research is in large measure a search for natures from which causal powers flow. The Agent’s Point of View Instead of looking at laws in order to cover an observed phenomenon, and thereby explain it, scientific research is more fundamentally oriented towards the discovery of entities, after identifying their powers. The link between natural power and regularity is direct: a natural power always actualizes if there is no obstacle. It is open to generality: to observe a regularity often leads to a causal power. This can 20 D. Davidson, ‘Causal Relations’, in Essays on Actions and Events (Oxford, 1980).

144

Dispositions and Causal Powers

be tested in an experiment. That is why the privileged point of view is that of the researcher in the laboratory, more than that of the theoretician. Experimentation is by way of intervention in the course of nature: placing an obstacle, isolating conditions, interfering with one power in order to highlight it. A theory like Millikan’s hypothesizes entities, electrons, but it is only through an experiment that a real scientific result has been obtained.21 Millikan’s hypothesis of the electron was only an explanatory hypothesis, and it is only when electrons were manipulated, first as bundles, and then individually, that the reality of the electron imposed itself on the scientific community. Even if that is still a disputed question, there is a big difference between a hypothesized entity and its experimentation in a laboratory with manipulations of that entity. One can speak about what does not exist, but one cannot do anything with it. The agent’s point of view is not the projection of an internal feeling (of effort), nor an analogy of natural action with human action, as exercising pressure, or passion, as the received pressure, nor with an emotion, even though these analogies are not without value (the notion of a magnet is still in use in order to point at the agent of a magnetic field). It is only the point of view, not often given by nature itself in a satisfying way for the researcher, that allows one to isolate the powers at work in natural changes. One can study them one by one, try to explain the measurements obtained, and the regularities that have been observed. Only when nature isolates by itself the powers at work (celestial mechanics) can we get predictability immediately, and only for the phenomena where we can easily isolate such powers (physics, chemistry), can we get it indirectly in a large and precise measure. In both cases explanation and prediction go together, mostly when the laws have been deduced from the theoretical apparatus and then only confirmed by experience. Without such contributions of nature or of experimentation, one can know which entities have which powers, but because of the interaction of too many of them, of the multiplicity of necessary conditions, predictability is almost beyond reach (metereology in the long run, seismology, etc.). Defeasibility of causal inference is omnipresent, and we understand why causal laws are always ceteris paribus, and can be trusted only when contrary powers are not in play, so that many natural changes are still beyond our power of prediction.22 It does not mean that they cannot be explained: we have satisfying theories of earthquakes and volcano eruptions, but our capacity of prediction is not at the same level as our capacity of explanation.

21 I am advocating for a realism of entities rather than of theories. It is with entities that we, the agents, can interact. 22 Among Aristotelicians, one will say that causes produce their effect ut frequenter, hôs epi to polu, most often, but that they may be impeded by a default of the active virtue, or by a bad disposition in the patient, or by another agent, etc.

Opium’s Virtus Dormitiva

145

A Metaphysics of Powers Categorial Base and Irreducibility of Powers I would now like to go to a more metaphysical analysis of the notion of a causal power. There is an often made objection to it, which does not deny its usefulness, as a concept, but which says that in reality, all there is are non dispositional properties. The concept of a causal power is just the concept of a supervenient disposition on these, and supervenient items are not real, at least not causally effective: the basic non dispositional properties are really what is doing the job of causation.23 We can speak of solubility (of sugar) or fragility (of the glass), and offer counterfactual conditionals which express those dispositions (if x is a piece of sugar and were put in water – in appropriate conditions – , it would dissolve). But those dispositions, and the truth conditions of the counterfactuals, must be analysed in such a way that they disappear in the end. We may still use the vocabulary of dispositions, but only as a useful tool. The categorical basis, made of non dispositional properties, is what really exists. A first answer must say that there would be a gross confusion in reifying the disposition, and consider it as a quality among others. This gives rise to the notion of a mysterious entity, a hidden quality, the virtus dormitiva mocked by Molière. But the idea of an active power is not that of a hidden quality, because it is not the idea of a quality at all.24 The criticism aims at a straw man. It treats disposition as a third thing apart from the agent and the patient, which would go from the first to the second.25 But there is no such third entity, just the actualization by the agent of a (potential) quality in the patient. To ascribe a power to an agent is not to say what there is (quality), nor what must be (law), but what would be there and now, if… This conditional is not made to express a law, but the (ontological) status of the causal power. The question raised of the categorical basis of dispositions comes as a reaction to their reification. But it comes down to their identification with, or reduction to, their vehicles, the structure of the agent which explains the natural action. The dormitive virtue would be only an epiphenomenal property, the real and basic ones being those at work in the mechanism of sleep. This is confusion here as well, because many mechanisms can have the same power, and a single mechanism may have many 23 See the Introduction, for bibliographical precisions. After nominalist criticisms of dispositions, by Goodman and Quine, there have been realist conceptions of universals and properties, still wanting to reduce dispositions to categorical properties. Others, like Shoemaker and Molnar, take the opposite view of reducing the categorical to the dispositional. I am afraid one will lose the usefulness of the concept if the duality categorical/dispositional is not maintained, at least as having distinct conditions of application. See S. Mumford, Dispositions (Oxford, 1998), ch. 4. 24 See R. Harré and E.H. Madden, Causal Powers (Totowa, 1975), p. 85. 25 This is a criticism made by Aquinas against some Muslim theologians quoted by Maimonides, cf. Summa contra Gentiles III, 69, n. 28.

146

Dispositions and Causal Powers

powers. Liquid water has not the same powers as frozen water, though they have the same (molecular) vehicle, but it has many powers in common with other liquids. An artefact, such as a car, receives some causal powers, like the capacity of going against, and overcoming, the force of gravity, from its components parts. But the sum of those parts, even poorly assembled, would not confer this power. In general a power is not identified by its vehicle, but by its exercise and its object. Moreover, the discovery of a mechanism does not yield a description lacking an active power. Rather, one may travel the road from the causal power to the underlying mechanism until a satisfactory explanation is found. An explanation (theory) can be considered as ultimate (Newton’s gravitation law). But the hypothesis of a mechanism explaining this last level is always conceivable. And some (considered as) ultimate explanations use dispositional notions in order to describe the basic level of reality, as it is the case with the notion of field of forces.26 Power and Potentiality The notion of disposition does not make any distinction between a disposition to act and a disposition to be acted upon. If there is a sense in distinguishing action and passion, although they are a single event, because agent and patient are two distinct entities (they could be one and the same accidentally, as a physician curing himself), this distinction is the basis for distinguishing active and passive power, which I will now call power and potentiality. A piece of butter has a passive power, the potentiality, to melt, but not the active power. Powers are tendencies or go along with tendencies. And it is because of tendencies and powers that changes occur: potentialities are necessary but not sufficient. If the tendencies were only potentialities, what really occurs would be the result of a number of things which would occur if…27 And that way, nothing would occur effectively. Tendency and power are actual, even if not actualized, they are not simply a kind of ‘this would occur if’.28 This distinction is linked to another between internal and external characteristics of a substance. A substance acts when a change occurs that depends on its internal properties, and it is passive, when it is involved in a change depending on external features, features of the environment. The notion of potentiality may seem to be too 26 I follow Mumford, Dispositions, who considers dispositions and categorical properties as two possible descriptions (relational and not relational) of the same ontological feature. It might be that, at the core of physical reality, or at the core of the ultimate version of physical science, basic features could be described only in terms of dispositions. 27 See P.T. Geach and G.E.M. Anscombe, Three Philosophers: Aristotle, Aquinas, Frege (Oxford, 1963), p. 104. 28 This is not to say that powers could not be ultimate constituents of reality (see above, note 26), because nothing would be actual in that case, as D. Armstrong, ‘The Causal Theory of Properties: Properties according to Shoemaker, Ellis and others’, Philosophical Topics, 26 (1999): 25-37; also published in Metaphysica, 1 (2000): 5-20; argues against Shoemaker. On the contrary, this is meant to indicate that a power has a certain actuality, so that it is real even when not actualized in its manifestation.

Opium’s Virtus Dormitiva

147

strict, and we would like to consider as complementary to the active power the sum of necessary conditions for its exercise, the opportunities, or the occasions. This distinction is not far from the Machiavellian doctrine of the ‘great man’, so that we could speak of virtue and fortune: no great man without a singular virtù (military genious, will, insensitivity of Napoleon…), but no great man, either, without a singular fortuna, (no Empire without a Revolution).29 The power is internal to the agent, the fortune is exterior. A tennis champion, because of a disease, may be defeated by an adversary he is otherwise capable of beating. The disease may seem to be interior, but it belongs to the exterior factors (the cause of the disease being an accident, a virus, and so on). These are anthropological examples, but they are given in order to make the distinction clear. We could distinguish as well the virtue of the fertilizer and the fortune constituted, in part, by the nature of the soil, which receives it, and allows its activity, the exercise of its virtue. Some logical features might help in characterizing this opposition. Without making an exhaustive list of the different meanings and uses of ‘can’, ‘power’, ‘possible’, we can remark the distinction between the ‘can’ of the virtus and the ‘can’ of the fortune with respect to the future.30 The opportunity (fortune) may be expressed in the present tense (I can speak English tonight, as we have invited a couple of Americans), and it may also be expressed in the future tense (I will be able to speak English tonight…). In the case of a future capacity (virtue), it must be expressed in the future tense (it would be incorrect to say ‘I can speak English next year, because I am learning it intensively’, we must say ‘I will be able to speak English…’). The reason is that the power is the principal explanation of an operation. It might have been acquired after a long effort, but, as long as it is not present, the operation remains impossible, though it will perhaps come to be possible. Whereas the fortune may be only the absence of a negative factor, an obstacle, and we consider that the operation (eating, opening the door) remains possible, even when the fortune is unfavourable, for a while. If the situation were to last a long time, it could become impossible (the spider cannot eat the fly any more, because of a bad fortune: another spider has eaten it). Physical Modality These logico-temporal considerations are not definitive: that would be a sin of logicism.31 But they indicate or confirm the intuition of a strong distinction between 29 See Machiavelli, The Prince, ch. VI, in which he cites Moses, Cyrus, Romulus and Thesaeus: ‘without the occasion, their talent and courage would have been useless, and without their personal qualities, the occasion would have occurred in vain’. 30 A.J.P. Kenny, Will, Freedom and Power (Oxford, 1975), p. 132. 31 Others could be added: Kenny remarks that capacities as well as occasions are not always present: they occur or appear and disappear. Other logical distinctions follow: for a capacity, the modal expression is the bearer of tense, not the action: I cannot, now, run 100 metres in 11 seconds, but, with some training, I will be able to do it (so I hope). The reason for this asymmetry is the generality of the capacity: it is not universal, but never limited to

148

Dispositions and Causal Powers

natural powers and the occasions of their exercise, which does not only depend on our conceptions of human situations. Exercise reveals power and power reveals nature. That is to say: a given nature possesses necessarily certain powers, and natural powers actualize themselves necessarily in appropriate conditions (no obstacle, no modification of the nature). Opium in good state, consumed by a man in good shape, will produce sleep, necessarily. This necessity is not logical: the form of the statement does not warrant its truth. We could speak of a conceptual link, the concept of the nature being an empirical one, acquired by experience, a posteriori. ‘Opium makes one sleep necessarily’ means: a. the dormitive virtue of opium is conceptually linked to the nature of opium, and if it vanished that would indicate a corruption of its nature; b. the exercise of the virtue is conceptually linked to the virtue itself, so that the virtue would not exist if it did not actualize when not impeded. The necessity is conditional in the following sense: the existence of opium is not necessary, but if opium exists, in appropriate circumstances, it necessarily makes one sleep. Such a necessity bears upon the actual and the possible and can be expressed by the use of a conditional counterfactual: if X had taken opium, in appropriate circumstances, he would have fallen asleep; if Y were opium, it would have a dormitive virtue. The idea of conceptual necessity has the merit of pointing to an internal connection, but it has the fault of being useful only for someone with an adequate concept of the nature of opium. Whereas one can know the dormitive causal power without knowing the real nature of opium. It seems to be better to accept a particular modality: that of physical necessity. One often opposes such a necessity, because it admits of exceptions. All such exceptions are evidently possible, because we omit, inevitably, mention of all the circumstances that could obstruct the exercise of the power. The agents’ specific natures impose limits, induce impossibilities, whatever the circumstances, and, in appropriate ones, they also determine natural operations. And here we find another type of exception: powers can go against each other, because they are limited (it is said that the cyanide swallowed by Rasputin was ineffectual because of the sugar he ate just before, and sugar annihilates the cyanide’s power). They can also add to each other in different ways, as we see with the addition of mechanical forces. Modal logic is not soft enough to accommodate all those characteristics.

only one exercise. Occasions on the contrary may be unique: the fortune is capricious. As a consequence, the indication of tense can be attached (1) to the modal operator: I am now able to speak with you, but was unable to do so one hour ago (I was anaesthetized), (2) to the action itself: I can speak with you now, but not tomorrow (you will be away), or (3) to both the action and its possibility: yesterday I still could cancel my conference, but it has now become impossible. And, in the last case, the temporal modifications are not the same.

Opium’s Virtus Dormitiva

149

Hierarchy of Powers We speak of a link between power and nature, though we ascribe powers to individuals because of their membership of a species, even though no exercise of this power has been observed (nor has occurred) in this individual: we ascribe the power to bark to a silent dog we see for the first time, as we ascribe the power to speak to any person we come across. Powers are generic, and many of them extend to the whole species. It is known because it has been actualized, even if it was by another member of the species. But this is not always the case, as it is not always the case that the exercise reveals the capacity: not all exercise implies the corresponding power, not all power reveals the corresponding nature. Usually, in order to ascribe a power to an individual, it should be the case that the internal power has had external manifestations. If I told you ‘when I am in good shape (conditions), I have the power to beat anybody, including the ATP No.1, but, here is the problem: I am never in good shape’, this would seem more like a joke, a nonsense, than a manifestation of pride. But the reality of one operation does not always allow the ascription of the corresponding power (to realize such operations) to the agent. It is not sufficient to score a basketball free shot to be credited with the power to do so. Many happy circumstances can explain what would seem to be a miracle. It was possible, in a sense, but I did not have the capacity. I certainly had an active power, without which nothing would have happened, but it was the power to throw the ball in the direction of the basket, not too high, not too fast, but fast enough. But in order for this shot to reveal the power not only to shoot but to score, one would certainly demand that I succeed in a good number and in a good proportion. Chance is to be distinguished from capacity. On the other side, if an action is very complex (as answering a question in a foreign language), one operation may suffice to reveal the corresponding capacity.32 We have to recover the distinction between first level and second level powers, above that of active and passive powers. Some natural agents are such that their natural power actualizes itself when the opportunity occurs. They should nevertheless still be considered as active powers, because we consider their subjects as agents: powder or fertilizer are the agents in the destruction of the building or the fertilization of the soil (though we also say that the building can collapse, or that the soil can be fertilized). It remains so when the operation occurs only by removing an obstacle (removens prohibens). What seems to be required is an active power or virtue and an (actual) fortuna, the importance of each being able to vary. At one extreme, the fortune, the opportunity, seems sufficient to trigger the exercise of the power: a spark triggers the explosion of powder. At the other extreme, the idea of omnipotence would be that of a power able to act without any opportunity, or even with the most contrary fortune.

32 There are still exceptions: Kenny gives the counter-example of Pope Pius XII who gave discourses in English though he did not master the language.

150

Dispositions and Causal Powers

A single natural reality can vary from passive to active power: a stone has the passive power of acquiring kinetic energy, which is an active power, explaining a change such as the breaking of a window. But it cannot acquire by itself such a power. On the contrary, I could speak English from my younger age, because I had the capacity of learning languages; and I can speak English now, if I want to, or if I am asked a question in English. Those are not only anthropomorphic examples: they are given to illustrate the notion of acquisition of a power. A dog learns to move and then to hunt: it has the active capacity – internal to it – of acquiring powers. The acquired one is a second power (or first activity), which can actualize itself in the exercise of this power (second activity). We can select among beings those which are able to acquire capacities. Among them, superior animals manifest powers both ways – ad utrumque. The dog can go and fetch the ball or not, at will. If we wanted to distinguish animal and human powers, we should do so considering human actions. But that is another topic. I just have endeavoured to defend the relevance for our ordinary linguistic uses, but also for a good understanding of scientific work, of the notion of a causal power, and the associated notions of nature and natural necessity. That they do not always appear as such is to be explained by the richness of our means of expression. Even if opium’s dormitive virtue is not to be taken as the explanation of sleep, and even though many circumstances are required (beginning with the absorption of opium), without this virtue why would opium consumers fall asleep? To refuse causal powers would be to leave all the job to the fortune, and to consider natural changes as a fairy tale.33

33 Thanks to Pascal Engel and Thomas Benatouïl for inviting me to present earlier versions of this chapter (at the university of Caen and at the Ecole Normale Superieure) and making remarks on them. I am specially grateful to Max Kistler, who read the last one and made useful criticisms.

Chapter 6

Conditional Possibility Bruno Gnassounou

Suppose that an event β is of such a type that it occurs only if another event of type α occurs. We can say that a particular event of the last type is a necessary causal condition of a particular event of the first one. It is plausible to contend that a statement that a certain condition α is a necessary causal condition of β conceptually implies that if this condition obtains (and perhaps so do other conditions), β will occur, and that if α doesn’t obtain, β will not occur. But it appears that it implies more than these simple conditionals. It has a modal import as well, or so it seems. It implies also that if α does obtain, β can occur and that if α doesn’t obtain, β cannot occur. Certainly, if heating the water is a necessary causal condition of its boiling, then if the water is not heated, not only will the water not boil, but it is not possible for the water to boil. Let’s call this possibility a conditional possibility. But which kind of possibility or impossibility is it? There is the straightforward regularist answer: the possibility is a logical one. To say that a particular event β is necessary is simply to say that it is a logical truth that if there is a general law according to which if any event of type α occurs, an event of type β occurs and if a particular event of type α occurs, then a particular event of type β occurs. The necessity attached to the particular event of type β is therefore reducible to the logical necessity attached to the inferential link between premises and conclusion of a specific deduction whose propositions are not themselves modalized. To think otherwise would come down to committing the gross fallacy of confusing a necessitas consequentiae with a necessitas consequentis. Accordingly, to say that a seemingly particular β-event is impossible is to say that there is a law from which we could logically infer that this β-event doesn’t occur, and finally the possibility of a particular event of type β means that there is no law such that we could logically infer that this β-event does not occur. I don’t want to discuss here the regularist account of causal modalities and I shall address those who don’t take this explanation of Humean flavour for granted. I shall rather focus on the attempt to reduce this kind of conditional possibility to a conditional proposition, because I am interested in the events involved in the manifestation of a disposition and because many people think, in a no less Humean way, that the possibility of this manifestation must and can be reduced to such a conditional connecting two events. I argue that this reduction faces difficulties, to say the least.

152

Dispositions and Causal Powers

Some decades ago, Keith Lehrer devised a subtle argument to disprove the analysis of such a possibility in terms of a conditional.1 Lehrer specifically attacked the various attempts of analysing the ‘can’ of freedom in terms of the famous Moorean conditional: ‘I will, if I choose’, but his argument has a more general import; at least some people at that time thought it had. Suppose we have the following triad: 1. If causal condition C obtains, A will φ. 2. If the causal condition C doesn’t obtain, A can’t φ. 3. C does not obtain. These three propositions are clearly consistent. The first two are classically future indicative analogues of subjunctive conditionals and may be regarded as counterfactuals (in the broad sense of the term which doesn’t imply the falsity of the antecedent) and the third simply states the fact that the condition is not fulfilled. Now, from (2) and (3), one infers, by modus ponens, that A can’t φ. But this conclusion must be compatible with (1), since the three propositions are consistent. They are not, therefore, mutually contradictory and the conditional expressed by (1) cannot mean that A can φ. Correlatively, the impossibility given in the consequent of (2) can’t be reduced to a conditional. Alvin Goldman thought the argument too good to be true. He claimed that if the argument were valid, the classical analysis of dispositions in terms of conditionals would have to be discarded, which is absurd. ‘A is soluble in water’, which certainly says that A can dissolve in water, does mean that if A is immersed in water, it will dissolve. Now let A be a piece of sugar and imagine the following situation: if A is immersed in water it dissolves ((1) is true), but when A is not immersed in water, a magician changes its molecular structure so as to make it insoluble ((2) is true). Now suppose that A is not immersed in water ((3) is true). In such a situation, the three propositions are true, implying that we could not analyse the solubility of the sugar by a conditional, which, according to Goldman, is nonsense. So the argument must have a flaw.2 Anscombe, on whom I heavily relied in the preceding lines, immediately notes that Goldman’s objection assumes that the absence of a certain molecular structure is sufficient to make the piece of sugar insoluble.3 It means that the presence of this very molecular structure is at least necessary to make it soluble. But if possessing this molecular structure is necessary for it to be soluble, one cannot claim that the simple conditionnal: if A is immersed in water, it will dissolve supplies an analysis of water-solubility. Anscombe justly concludes that a decision has to be made: either 1 K. Lehrer, ‘An Empirical Disproof of Determinism’, in K. Lehrer (ed.), Freedom and Determinism (New York, 1966), p. 196. 2 A. Goldman, A Theory of Human Action (New Jersey, 1970), note 2, pp. 199-200. 3 G.E.M. Anscombe, ‘Soft Determinism’ in G.E.M. Anscombe, Collected Philosophical Papers (Oxford, 1981), vol. II, p. 164.

Conditional Possibility

153

you claim that the triad of propositions is consistent and then you have to admit that the analysis of dispositions in terms of a simple conditional is erroneous, or you claim that the conditional analysis of disposition is right, but then you have to acknowledge that the triad is inconsistent. You cannot have, as Goldman wants, both the consistence of the triad and the conditional analysis of the disposition. Now the main elements of the argument that we find in Goldman’s objection are familiar to us and have been made popular by C.B. Martin.4 The disposition to dissolve of the magician’s piece of sugar is a ‘finkish’ one. An object’s disposition is finkish if the disposition vanishes when the object is put to the test. As it is well known, however, the argument is directed at any attempt to give a conditional analysis of dispositions. So why did Goldman think we could use a similar argument to establish the validity of the conditional analysis of dispositions? Perhaps for the following reason. A defender of Goldman’s objection may think it possible to avoid the conclusion by granting that the simple conditional analysis of dispositions is wrong and by saying that one of the conditions of the dissolving is the possessing of a certain molecular structure, a ‘causal base’, as it is sometimes called. Then he could have concluded that the original triad with the simple conditional is consistent, but that if you add to the conditions the one which states that the piece of sugar must possess such and such a causal base and so make the conditional more complex, it is contradictory. But this contention is unwarranted. If the possessing of a causal base, call it C1, is now required to ascribe a disposition to φ to an object, we can apply again Lehrer’s argument to the more complex conditional and reach the same conclusion: a. If C and C1 obtain, A will φ. b. If it is not the case that C and C1 obtain, A cannot φ. c. It is not the case that both C and C1 obtain. This new triad seems as consistent as the original one and we must conclude that the ‘can’ is not equivalent to the conditional (a). If the piece of sugar is not immersed in water or if it doesn’t possess the right molecular structure necessary for dissolving, it is not possible for it dissolve. But this is perfectly compatible with the fact that if the piece of sugar is immersed and if it possesses this molecular structure, it will dissolve, even if it is in fact not the case that these two conditions are both fulfilled. D. Lewis has defended the conditional analysis in adding to the set of conditions referred to in a counterfactual that the object put to the test retains its causal base, for example that the non-soluble piece of sugar retains its molecular structure when placed in water,5 which is not the case in Goldman’s or Martin’s cases. Call this retention of the causal base, C2. But we can, then, repeat the argument:

4 1-8. 5

C.B. Martin, ‘Dispositions and Conditionals’, Philosophical Quarterly, 44 (1994): D. Lewis, ‘Finkish Dispositions’, Philosophical Quarterly, 47 (1997): 157.

154

Dispositions and Causal Powers

a’. If C, C1 and C2 obtain, A will φ. b’. If it is not the case that C, C1 and C2 obtain, A cannot φ. c’. It is not the case that C, C1 and C2 obtain. In fact, for every condition, C0… Cn, it is always logically possible that one of these conditions doesn’t obtain, so that the object cannot manifest the disposition. Such is the apparent power of Lehrer’s argument. But, having reached this point, a doubt is likely to assail us. Is it not obvious on reflection that the possibility expressed in the consequent of the second premise of the argument is not the possibility of a disposition? When we say that the piece of sugar cannot dissolve if not placed in water, we don’t deny the piece of sugar the general capacity to dissolve, but only the circumstantial possibility of dissolving: the circumstances are such that it is not possible for the piece of sugar to dissolve, even though, in these circumstances, the piece of sugar retains the general power to dissolve. Manifestly, that is where the flaw is in Goldman’s applying the argument to dispositions, as Anscombe points out:6 even though the conditional is true, the failure now of one of the conditions to occur makes it impossible for the sugar to dissolve right now. It does not suppress its power to dissolve in water. The argument, as such, says nothing against the conditional analysis of dispositions. In brief, we have to distinguish between two concepts of possibility or impossibility, as some good authors have made clear.7 The distinction is plain in the case of human voluntary powers, but it extends to non-voluntary substances as well. You can, at t, have the general power to φ. Besides, the circumstances may give you the opportunity to exercise this power. But then, they don’t give you a new power, a power-to-φ-at-t, which would require itself circumstances giving you the possibility to exercise it. An opportunity is not exercised by the agent; it only allows this agent to exercise his capacity. A general power has no temporal indexation in itself even though you can acquire a power, have it or lose it at a given time. Having, at t, the power to φ, however, is not having the power-to-φ-at-t. So I can certainly have the general capacity to read this book, without having the opportunity to do so. The wise Davidson who seemed to think at one time that Goldman is perhaps right,8 eventually acknowledges in a note that ‘the correct response to Lehrer is simply that if one analyses solubility by a causal conditional, one can’t consistently allow that what causes dissolving is also a necessary condition of solubility, since in that case the only soluble things would be

6 G.E.M. Anscombe, Collected Philosophical Papers (Oxford, 1981), vol. II, p. 165. 7 A.J.P. Kenny, Will, Freedom and Power (Oxford, 1975), p. 131 sq. 8 D. Davidson, ‘Freedom to act’, in D. Davidson, Essays in Action and Events (Oxford, 1980), p. 69: ‘Although Lehrer does not seem to have noticed it, his argument has nothing in particular to do with action. If his line of reasoning is sound, it shows that no attribution of a power or disposition is ever equivalent to a conditional […] Thus on Lehrer’s argument, to say something is water-soluble can’t mean it dissolves if placed in water, since it may not be soluble and yet placing it in water would make it dissolve’ (my emphasis).

Conditional Possibility

155

9

dissolved’. What Davidson should have said is that what causes dissolving is also a necessary condition of the circumstantial possibility for the sugar to dissolve. Having this in mind, the original triad, applied to sugar, must therefore read now, where i is a tiny time interval:10 1*. If A is immersed in water at time t and left in it, A will start dissolving within t+i. 2*. If A is not immersed in water at t, it cannot dissolve, i.e. it has not the circumstantial possibility of dissolving within t+i. 3*. A is not immersed in water at t. Now, if Lehrer’s argument does work, we can draw the conclusion that a statement of circumstantial possibility is no more reducible to a conditional than was the ascription of a disposition (if this ascription made any sense). The triad seems consistent and since, from (2*) and (3*), we can infer that A hasn’t the circumstantial possibility of dissolving in water at t, (1*) cannot give an analysis of this concept of possibility. But, here, the conditionalist may have a radical reply, which Anscombe suggests,11 and which is worth considering because its final dismissal compels us to be more precise on the distinction between the circumstantial and the general concept of possibility. Let us suppose that the conditionalist, in the end, simply defines the possibility as what is given by the conditional (1) and accordingly proceeds to substitute in (2) the negation of this conditional for the ‘cannot’. We then have the following triad: 1. If C does obtain, A will φ. 2’. If C does not obtain, then not 1. 3. C does not obtain. One infers from (2’) and (3) that (1) is false. And, of course, this conclusion is this time contradictory to the proposition (1). Note, however, that, taken in itself, the argument is most odd. For the truth of (2’) means that if the piece of sugar is not immersed in water, then it might be the case that if this piece of sugar is immersed in water, it will not dissolve.12 How is it possible that a piece of sugar, in virtue of the simple fact that it isn’t placed in water, is such that it might not dissolve if placed in water? Note also that you accept (1), 9 D. Davidson, Essays in Action and Events, p. 70, note 7. Davidson says in this note that he is grateful to Anscombe for correcting his interpretation of Lehrer. Anscombe herself in her paper (pp. 163-164) quotes Davidson critically for what he wrote in an apparently earlier version of his paper. 10 See G.E.M. Anscombe, Collected Philosophical Papers, p. 165. 11 G.E.M. Anscombe, Collected Philosophical Papers, p. 168, but she herself does not apply it to natural substances. See below. 12 I use the conventional interdefinability of the would- and the might-conditionals: ψ□→ϕ =df ¬(ψ ◊→¬ϕ)

156

Dispositions and Causal Powers

(2’) and the plausible (4): if C does not obtain, A will not φ, you can prove the logical equivalence of the conditional and the statement that A φs, so that if the conditional is true, that is, according to the conditionalist, A can φ, then A will φ, and that if A will not φ then the conditional is false, i.e. A cannot φ.13 This quasi-megarian reduction of possibility to actuality is questionable. In order to make the argument sensible, we have to figure out a specific context for it. It is not difficult to picture the natural candidate for the required script: imagine that C doesn’t obtain and that it is the case that if C doesn’t obtain, there is an obstacle which prevents A from φ-ing. In that case, one must say that (1) is false. Suppose that if the piece of sugar A is not immersed in water, there is a chemical additive present in the water preventing A from dissolving if it were placed in it. In those circumstances, of course, it is not the case that if A is immersed in water, it will dissolve. Or imagine a ball lying on the ground and touching a wall. If the ball receives an impulsion in the direction of the wall, it will not move, even though the ball has the general capacity to move when pushed. We may say that the falsity of the conditional means the presence of an obstacle and its truth means the absence of any obstacle to the exercise of the power. One might object that the incompatibility of the three propositions hasn’t really been proved for the following reason. Suppose that if C is not fulfilled, then there is indeed an obstacle which prevents A from φ-ing, but that if C is fulfilled, the obstacle vanishes and so A φs; and take it that it is the case that C is not fulfilled. These three facts are perfectly compatible. For instance, suppose that if a ball (which has the general capacity to move when pushed) is not pushed, a wall is there which prevents it from moving. But that if the ball is pushed ahead, the wall disappears and the ball moves away. So, if the ball is not pushed, it can’t move ((2) is true), but if the ball is pushed, it will move ((1) is true as well). There is no contradiction, which means that the possibility can’t be given a conditional analysis.14 13 We need (4) if we want to pass from ¬(C does obtain □→A ϕs) and ¬ (C does obtain) to ¬(A ϕs). 14 Note that the situation looks like the ones put forward recently by A. Bird against the conditional analysis of dispositions: ‘Another way of protecting the glass once it is struck is to find an antidote to striking. The sorcerer, being a brilliant physicist, may be able to administer shock waves to the struck glass which precisely cancel out the shock of the original striking, hence saving the glass from destruction’ (A. Bird, ‘Dispositions and Antidotes’, Philosophical Quarterly, 48 (1998): 228). The first way alluded to by Bird is to imagine a sorcerer à la Lewis capable of changing the molecular structure of the glass when it is struck; the fragility of the glass would then be simply finkish. In our case, the situation is the reverse of the one just described by Bird: imagine that if the glass is not struck, a vigilant sorcerer à la Bird is present, and when the glass is struck, the sorcerer disappears. It would apparently be true that if the glass is not struck, the conditional is not true, but when the glass is struck, the conditional is true. But note that here the argument can be assumed by the conditionalist because it is supposed to be directed against the non-analysability of the concept of circumstantial possibility in terms of a conditional, not against the non-analysability of the concept of disposition.

Conditional Possibility

157

Anscombe uses an argument of this sort to demonstrate that the power of acting freely of a voluntary agent is not reducible to a conditional of the form: if A chooses to φ, A will φ. To take one of her examples,15 A will go through a marriage ceremony with B, if she chooses to ((1) is true). But in fact, she doesn’t choose to go through the ceremony ((3) is true). If the conditional analysis were true, we would have to conclude that it is not logically possible that anyone (for example friends) would prevent A from going through the ceremony if she doesn’t choose to! But certainly this situation is quite possible (imagine that someone forces the reluctant A to go the marriage; then A’s friends stand in the way of the ceremony; they let her go only if she does it voluntarily). Her objection is, in my opinion, perfectly correct when it concerns the operations of a voluntary agent. It works because, for such an agent, even if an item is not physically blocking the action, it functions as an obstacle to the action. A person in jail has no possibility of escaping, even if she is not trying to escape and is not being hindered to do it by the wall: the presence of the wall is already an obstacle for her future action. A reasonable man, deliberating about what to do, if he knows that the wall will prevent him from getting out, ought to decide that he must not try now to get out, and the reason he may advance in favour of this decision is that the walls will be physical barriers to his action. By being regarded as such in the practical reflection of the agent, they are already, so to say, rational, but nevertheless quite real obstacles to it.16 But what is true of voluntary agents is not true of agents deprived of voluntary powers, and especially of natural substances. There is no future obstacle having present consequences for such a substance. For this kind of substance, only present obstacles have present effects. If at t, a ball is moving ahead and if there will be at t+i an obstacle, for example a wall that lies a little ahead of the ball’s probable path, this wall doesn’t at t make the ball stop moving. The wall has at t the general capacity of being an obstacle, but it is not then exercising this power: it is a potential obstacle, not an actual one. If the wall is to function as a real obstacle, the ball must have received an impulsion (the condition C must be fulfilled) at a given time and then the wall must prevent its moving at this point of time. The situation is not different if we consider a motionless ball touching a wall. If the ball doesn’t receive an impulsion, the wall is no obstacle to any movement of the ball yet. Perhaps, it will be, but, in any case, it is not one now. So the conditionalist must say that even though (1) is true (if the ball is pushed, the wall will disappear), and (3) is true (the ball is not pushed), (2) cannot be true: it is not the case, that if the ball is not pushed, there is an obstacle: the wall. If the ball is not pushed there is no obstacle to its action and the ball can move. 15 G.E.M. Anscombe, Collected Philosophical Papers, pp. 168-169. This is not exactly the way Anscombe uses this example but it is structurally similar to her use of other examples given in her paper. 16 And what happens if the prisoner knows that if he chooses to leave, the walls vanish? I prefer to keep away from the muddy waters of what seems dangerously similar to the untractable Newcomb’s paradox.

158

Dispositions and Causal Powers

The truth of a conditional like (1) means that there is no present obstacle to the operation of the substance. It is, of course, not incompatible with the fact that the antecedent of the conditional is not true, i.e. that C doesn’t obtain. But it is incompatible with anything which implies that there is an obstacle to A’s φ-ing. In particular, the fact that C doesn’t obtain must not imply that there is an obstacle. Of course, this argument supposes that A cannot φ means it is not the case that if C does obtain, A φs, and that A can φ means if C does obtain, A φs. And this is made plausible by the internal relationship between the conditional and the absence of obstacle and correlatively the negation of the conditional and the presence of an obstacle. So if natural substances are concerned and if the ‘can’ expresses the absence of obstacles, then the conditional analysis is right. This last consideration, however, testifies to the limited range of applications of the conditionalist’s strategy and indeed to its harmlessness for the original consistency of our triad. In effect, it is logically admissible that 1 there is no obstacle to the dissolving of the piece of sugar, that is if the piece of sugar is immersed in water, it will dissolve, that (2) if the sugar is not immersed in water, it cannot dissolve and (3) that the sugar is not immersed in water. Although absolutely nothing prevents the water from dissolving the sugar, it remains true that if the water and the piece of sugar don’t come into contact, the sugar cannot dissolve in water. The ‘can’ here cannot be analysed by the conditional. What all this amounts to is that it is not possible to assimilate the absence of a causal condition (to be pushed, to be immersed in water) for the exercise of a power to the presence of an obstacle to this exercise. To repeat, in the situation, nothing is preventing the water from dissolving. If a ball doesn’t move because it is not pushed, it would be meaningless to say that the absence of a stroke is an obstacle to its moving. Some may be inclined to say that if a natural substance doesn’t exercise its power, an obstacle must be lurking somewhere. I disagree. If a ball is placed in a portion of space where it cannot be attracted by a massive body because the intensity of the gravitational field is too low at this place, although it has the general capacity to be so attracted (which implies that if it were closer to the massive body, it would be attracted17), this impossibility is not due to any obstacle. And the spatial distance between the two bodies is not such an obstacle: space doesn’t act, no more than time does. It would be no less absurd to contend that space hampers the contact between the two bodies than to say that the earth two million years ago was prevented from attracting the moon now owing to the time interval between the two events. That is the reason why reducing the distance between the two massive bodies doesn’t amount to the suppression of a hindrance. It is not as if the ball was very close to the massive body but was hindered from moving because of the presence of a table. The 17 I agree with S. Mumford that the attribution of a disposition conceptually implies the conditional even though, for reasons excellently put forward by him, the conditional doesn’t supply any analysis of the ascription of a power. See S. Mumford, Dispositions (Oxford, 1998), ch. 4.

Conditional Possibility

159

table is, indeed, an obstacle to the fall and if you remove the table, the ball will be attracted. It is really a case of causa removens prohibens. It would be pleasant if, contrary to the Aristotelian tradition, all this came down to holding that general capacities can be analysed by conditionals and that circumstantial possibilities cannot. It would be pleasant, but it would be false. If the argument says nothing against the conditional analysis of dispositions, it says nothing for it either. And it remains true that many circumstances, if not most of them, are cases of obstacle and can be given a conditional analysis. So, after all, our conclusions are compatible with this tradition, a filiation which, in my opinion, is far from being infamous.

This page intentionally left blank

Chapter 7

On Ascribing Dispositions Ludger Jansen

They Do It Teachers do it. Engineers do it. Scientists do it. Insurance agents do it. Both plain and sophisticated folks do it. What do they do? They ascribe dispositions. -

-

-

Teachers do it in class. They call some students ‘witty’ and some ‘lazy’. They call some ‘intelligent’ and some – though only in private – ‘stupid’. Engineers do it on the job. They build bridges which they call ‘stable’. They invent textiles which they call ‘water-resistant’. And for their work they use materials that are, for example, called ‘non-flammable’ or ‘flexible’. Scientists do it in the lab. They utter sentences like ‘Aspirin relieves headaches’, ‘Sugar is water-soluble’ or ‘Porcelain is an insulator’. Or such sentences like ‘Mass-points attract each other’, ‘Neutrinos have a half-life period of so-andso long’. And many of these sentences find their way into scientific articles, manuals and textbooks. Insurance companies do it for money. They classify smokers as being more liable to have health problems than non-smoking people. They expect women to cause higher expenses in hospital than men, who still cannot give birth to children. And they expect (at least in some countries) civil servants to cause fewer car accidents than the average driver.

In this paper I want to exploit the fact that ascriptions of dispositions are actions. So far, I listed some examples of disposition ascriptions, which give us an impression about who performs such actions and about the variety of different kinds of disposition ascriptions. In what follows I want to answer the questions why, how and when we ascribe dispositions on pp. 161-3, and 172-6, respectively. And, as ascribing always is ascribing something to something, I will ask on pp. 163-5: What do we ascribe when we ascribe dispositions? And on pp. 166-72: To whom do we ascribe dispositions? Functions: Why Ascribe Dispositions? In the examples given, the ascriptions of dispositions serve quite different functions. One of the main motivations for ascribing dispositions is the wish to explain.

162

Dispositions and Causal Powers

Teachers want to explain why some students learn quicker than others. Next comes the necessity to choose certain actions to react to a given situation. The reaction will depend on the dispositions ascribed. E.g. the pedagogical reaction to a student’s failure in an exam depends, among other factors, on the intelligence ascribed to the student by the teacher: A teacher will treat a less intelligent student differently than an intelligent but under-achieving student. In the case of engineers and scientists, these two functions of disposition ascriptions are roughly divided between the two professions. While scientists primarily want to explain what happens in the world and thus ascribe dispositions to the things involved, engineers take into account these dispositions and choose materials and constructions accordingly. Insurance companies are a similar case. They react to known or assumed dispositions, taking them into account in the statistical calculus used to calculate the cost of a policy. Presumably, they do this without any wish to explain anything. Pragmatics: How to Ascribe Dispositions? Obviously, disposition ascriptions like ‘Jennifer is intelligent’ or ‘This bridge is stable’ are predications. In such predications, certain predicates are ascribed to certain individuals. Other dispositions seem to be more tricky, like ‘Aspirin relieves headaches’ or ‘Mass-points attract each other’. Later, I will discuss these more sophisticated kinds of disposition ascriptions in detail (pp. 166-72). Here it will suffice to stick to plain predications to get my points across. Let’s suppose that I want to cross a river, and there is only a tiny little bridge. Someone tells me: ‘This bridge is stable.’ If I trust him, I will trust the bridge. I will suppose that the bridge will securely support me while walking over it, or the weight of my car while driving over it. Will I suppose that the bridge will support any weight? No, I won’t. I will suppose that the bridge will crash down if the weight exceeds a certain limit. And if I know a bit of physics I may also know that the bridge might be destroyed by much smaller forces, if these are applied rhythmically, for example by a group of soldiers marching in step. Will I then think that my informant was wrong, that the bridge in reality is not stable at all? Again: No, I won’t. I will understand the sentence ‘This bridge is stable’ as being short for: ‘This bridge is so stable as to support any weight which will be put on it in the course of any action which I suppose you will undertake involving the bridge.’ Thus this disposition ascription is elliptical. Nevertheless we achieve communicational success, because in most cases we have shared assumptions about the action that walkers or car drivers do with bridges: they cross them. Often it is said that such disposition ascriptions refer to something like ‘normal’ conditions.1 But this misses the point. Because often it is not the normal condition but an extreme condition which is in question. If I am an army lieutenant, commanding 1 Cf., e.g., Spohn, who nevertheless gets interesting results from this assumption. W. Spohn, ‘Begründungen a priori-oder: ein frischer Blick auf Dispositionsprädikate’, in W.

On Ascribing Dispositions

163

a company of infantry, I expect a different piece of information about the bridge than the normal walker or car driver. I have to know how many people and cars may be on the bridge simultaneously, and whether or not the infantry may march in step. Now, the sentence ‘The bridge is stable’ may be short for ‘The bridge is so stable as to support a company of infantry when marching in step.’ And this is quite a different disposition than that asked for by normal car drivers.2 If such disposition ascriptions are elliptical, how is successful communication about dispositions possible? It is possible, because, like in many other kinds of discourse, additional information is provided by the context of the utterance or by tacit mutual assumptions of the speaker and the hearer. This is enough to give action-related information involving dispositions. But is it enough for scientific descriptions? This is not so clear, because there is no related action for science as such – only the desire to explain. I do not want to discuss this question further here. Instead, I will next discuss in which ways a disposition ascription can be elliptical. For even if the context should also turn out to be helpful in the case of explanation, science will be better off trying to fill these gaps. For the aim of science is to make things explicit. As it is not implicit but explicit knowledge science aims at, science should strive to find explicit formulations of its findings. Syntax: What is Ascribed? We ascribe dispositions by ascribing predicates. So much seems to be clear. Thus the surface grammar of a disposition ascription is plain and simple predication, consisting of a disposition predicate (‘D’) and a singular term (‘x’) as a subject: D(x) Now a disposition is always a disposition for some kind of action, its specific manifestation M. Thus, somehow this manifestation predicate must be hidden within the disposition predicate. My suggestion is to express the relation between a manifestation predicate and the related disposition predicate by means of a predicate modifier. A predicate modifier is an operator that takes a predicate and yields a new

Lenzen (ed.), Das weite Spektrum der analytischen Philosophie. Festschrift für Franz von Kutschera (Berlin, 1997), pp. 323-345. 2 Similarly, S. Mumford, Dispositions (Oxford, 1998), p. 90 speaks about ‘the context relativity of relevant ideal conditions’. However, Mumford misses the ambiguity of manifestation terms, which is in question in the bridge example. See the discussion in the next section. In addition it is not clear how Mumford thinks that the context contributes to the meaning of disposition ascriptions.

Dispositions and Causal Powers

164 3

predicate. Here I use the modifier ‘dyn’, so-called in reminiscence of Aristotle’s concept of dynamis.4 Thus, hidden behind the surface is something like: (dyn M)(x) Now, dispositions can differ, because they allow for different manifestations. Say, supporting a certain mass m1, or supporting a larger mass m2 etc. Furthermore dispositions can differ, because they allow for the same manifestation under different conditions, say in different temperatures. A bridge built from ice may have the disposition to support a car under a temperature of -20°C, but not the disposition to support a car under a temperature of 10°C. Thus we have to account for the possibility of such different conditions in disposition ascriptions. The best way to do this is to recognise that the manifestation predicate M can form a complex conjunctive predicate together with a condition predicate C, e.g. ‘... has a temperature of 10°C’. Thus we now arrive at the following structure: (dyn (M.C))(x) The predicate conjunction ‘.’ is defined as follows: If A and B are predicates, A.B is also a predicate. (A.B)(x) is true, if and only if both A(x) and B(x) are true.5 It might seem odd that – according to this conjunctive analysis – the manifestation predicate and the manifestation conditions are on equal footing. Some may want to object that in this case the manifestation predicate and the condition predicate have symmetrical roles and that they may even switch places. This is true, but it should not bring us any trouble. If a bridge has the disposition D1 to support a car while having a temperature of 10°C, it also has – indeed as a logical consequence – the disposition D2 to have a temperature of 10°C while supporting a car. Bridges without this disposition would just cease to exist when reaching a temperature of 10°C while supporting a car. There is nothing wrong with this. The reason, why the ascription of D2 sounds odd but not the ascription of D1 is that bridges are built to support cars; that they can stand certain temperatures is only of instrumental value for this end. However, everything that possesses D1 also possesses D2, and vice versa. This is why the conjunctive analysis is appropriate for manifestation conditions. Of course, the predicate C describing the manifestation conditions can itself be a highly complex predicate. The task of science is to learn as much as possible about the predicates M and C involved, and thus to accumulate as much explicit knowledge

3 Cf. R. Clark, ‘Concerning the Logic of Predicate Modifiers’, Noûs, 4 (1970): 311335. For discussion and further references cf. L. Jansen, Tun und Können. Ein systematischer Kommentar zu Aristoteles’ Theorie der Vermögen im neunten Buch der Metaphysik (Frankfurt, 2002), ch. 1. 4 Cf. L. Jansen, Tun und Können. 5 Cf. L. Jansen, Tun und Können, p. 183.

On Ascribing Dispositions

165

about the dispositions of the things we find in the world as possible. I discuss how knowledge about dispositions can be acquired on pp. 172-6. Another objection may arise:6 Disposition ascriptions, or so the objection starts, should allow us to predict future outcomes. But if we ascribe dispositions, as stated, always for very special manifestations in very special circumstances, dispositions are of no use for predictions, because we do not know whether these very special circumstances will obtain in the first place. We may, for example, ascribe to a bridge the disposition to support our weight at a temperature of 20°C and an atmospheric pressure of 1013 hPa. But then it is quite unlikely that the weather produces these very conditions, and thus we cannot predict any outcome of our walking across the bridge. So much for the objection, now my answer. While I agree to the objection’s premises, I do not subscribe to its conclusion. It is true, I think, that we cannot predict much if we only know a thing’s disposition to react in a certain way in one set of circumstances. But what we usually ascribe to things is a whole family of disposition predicates, describing the reaction in a whole range of circumstances. And in many cases it will be fairly certain that circumstances from this range will obtain in a future situation. For example, we may ascribe to the bridge the disposition to support our weight at any temperature from, say -60°C to 60°C, thus covering virtually any temperature conceivable in earthly climates. Whatever the temperature will be like in a future situation, the bridge will support our weight because of its dispositional properties. Thus, even if disposition ascriptions are elliptical in the way described in this section, they still allow for prediction if they cover an appropriate range of circumstances. Someone who prefers sparse ontologies may want to object here that this suggestion multiplies the number of dispositions: the disposition to support at 1°C, 2°C, 3°C etc., and of course also the temperatures in between, say, at 1.27°C or 37.93°C. But what is being multiplied here is the number of disposition predicates that can be ascribed to the bridge, the number of disposition ascriptions. Thus it is a multiplication at the level of description. Whether this procedure brings with it a multiplication at the level of being, at the ontological level, depends on one’s theory of token identity of dispositions. That all these disposition ascriptions are true of a thing might be accounted for by one and the same property token. Thus there may be different disposition tokens that make this wide range of disposition ascriptions true, but this is not necessarily so: Maybe there is only one property token that makes all these disposition ascriptions true.

6

This objection has been raised by Nancy Cartwright in discussion.

Dispositions and Causal Powers

166

Ontology: To Whom are Dispositions Ascribed? Four Candidates The question to which I want to turn now is: To whom are dispositions ascribed? As an initial approach to answering this question we can look for the grammatical subjects of disposition ascriptions. Browsing through the examples I mentioned at the beginning, we find things like: -

John and Jennifer, two students: John is witty and Jennifer is intelligent. The old Forth Bridge and the new Forth Bridge, two constructions by civil engineers: The new Forth Bridge is stable, the old one wasn’t. Gum and Gore-Tex, two materials: Gum is flexible, Gore-Tex is waterresistant. Mass-points and neutrinos, theoretical entities of quite different standing. Smokers and civil servants: Smokers are more liable to have health problems, civil servants less liable to cause accidents than the average driver.

In reflecting on this list, different categories of entities come to mind: Living versus non-living entities. Theoretical versus non-theoretical entities. Different things versus different kinds of stuff. And, more important, individuals versus kinds. Consider the following sentences: S. Zucky (= my favourite chunk of sugar) is water-soluble. K. Sugar is water-soluble. Of course, in this case K implies S: If sugar is water-soluble, Zucky is water-soluble, too. But this is not valid the other way round: In a world with a different set of laws of nature, sugar might not have the disposition to dissolve in water, while Zucky might still be water-soluble for other reasons than being sugar. Aside from things and kinds, a further candidate for the ascriptions of dispositions has been suggested by Nancy Cartwright: properties.7 As Cartwright puts it – inspired by the second-order-property-view of dispositions:8 PC. The molecular structure of sugar brings with it water-solubility. In (PC), one function of the phrase ‘the molecular structure of sugar’ is to pick out what we are talking about: about the stuff having the molecular structure of sugar. Were this all there is to (PC), we could render this statement as (P*): 7 Cf. N. Cartwright, Nature’s Capacities and their Measurement (Oxford, 1989) and N. Cartwright, The Dappled World. A Study of the Boundaries of Science (Cambridge, 1999). 8 For references for and criticism of this view of dispositions cf. L. Jansen ‘Dispositionen und ihre Realität’, in C. Halbig and C. Suhm (eds), Was ist wirklich? Neuere Beiträge zu Realismusdebatten in der Philosophie (Frankfurt, 2004).

On Ascribing Dispositions

167

P*. Things having the molecular structure of sugar are water-soluble. But in (P*) an important aspect is missing. According to the property-view, the property predicate in question has not only the function of picking out those things that are supposed to have a certain disposition, like in (P*), but the property is thought also to have a causal or explanatory function: those things posses this disposition, because they have that property. This explanatory function can be expressed in reduplicative sentences, i.e. in sentences containing phrases like ‘qua’ or ‘in virtue of’, like (PR): PR. Things having the molecular structure of sugar are water-soluble in virtue of their having the molecular structure of sugar. What Cartwright has in mind are types of properties. We can add tokens of properties, which are often called ‘tropes’ or ‘abstract particulars’9 as a further candidate: TC. This token of the molecular structure of sugar brings with it a token of watersolubility. TR.This token of the molecular structure of sugar brings with it a token of watersolubility in virtue of its being a token of the molecular structure of sugar. Thus we exhaust the four-fold ontology Aristotle aims at in the second chapter of his ‘Categories’:10 PARTICULAR

UNIVERSAL

CONCRETE

individual substances (this particular man)

universal kinds of substances (mankind)

ABSTRACT

individual property tokens (this particular red)

universal property types (redness)

I will argue that disposition ascriptions to individual substances are the most basic kind. I want to reach this conclusion in four steps: 1. I will argue that the ascription of dispositions to individual substances is prior 9 Discussion of and arguments for trope ontology can be found in, e.g., C. Macdonald, ‘Tropes and Other Things’, in S. Laurence and C. Macdonald (eds), Contemporary Readings in the Foundations of Metaphysics (Oxford, 1998), pp. 329-350. 10 For a modern defence of such a fourfold ontology cf. E.J. Lowe, ‘A Defence of the Four-Category Ontology’, in C. Moulines and K.-G. Niebergall (eds), Argument und Analyse. Akten des vierten Internationalen Kongresses der Gesellschaft für Analytische Philosophie (Paderborn, 2002).

168

Dispositions and Causal Powers

to the ascription of dispositions to kinds of substances pp. 168-9, 2. that disposition ascriptions to individual substances are prior to disposition ascriptions to individual property tokens pp. 169-71, and 3. that disposition ascriptions to universal kinds of properties are not prior to disposition ascriptions to individual property tokens pp. 171-2. 4. From this I will conclude, by transitivity of priority, that ascriptions of dispositions to individual substances are prior to ascriptions to any of the other ontological categories. Individual Substances Versus Kinds of Substances When I talk about a disposition ascription to a universal, I do not mean that there is some universal or platonic idea that has a capacity to be changed or to change itself. Rather, the ascription of a disposition D to a universal U means that anything that exemplifies U also possesses that disposition D. Not only does Zucky have the disposition to dissolve in water, but also sugar in general. Not only does Black Beauty have the disposition to digest hay, but horses in general. Now, which comes first, ascription to individual substances or disposition ascriptions to kinds of substances? I want to show that disposition ascriptions to individual substances are more basic in (1) metaphysical, (2) logical and (3) epistemological respect:11 1). Metaphysical priority is not only an old,12 but also a very useful concept, that can help to detect ontological dependencies. It can be tested by a simple contrafactual criterion: A is metaphysically prior to B, if A can be without B, but B cannot be without A. As I already mentioned, the truth of disposition ascriptions to individual substances is independent from the truth of the disposition ascription to the corresponding kinds. ‘Zucky is water-soluble’ can still be true, even if ‘Sugar is water-soluble’ were false, whereas it is not possible that ‘Zucky is water-soluble’ is false while ‘Sugar is water-soluble’ is true. 2). The logical structure of a disposition ascription to individual substances is just the application of a predicate to a singular term, say ‘Wz’ (with z = Zucky and ‘W’ for ‘is water-soluble’). This was the kind of disposition ascriptions discussed on pp. 163-5. The logical structure of a disposition ascription to kinds of substances is more sophisticated. It might be considered to be a universal

11 In showing this I give also reasons against the theory of disposition predicates proposed by E.J. Lowe, ‘Dispositions and Laws’, Metaphysica, 2 (2001): 5-23, who tries to reduce disposition ascriptions to concrete particulars to ascriptions of manifestation predicates to universals. In doing this, however, Lowe conflates the dispositional/categorical distinction with the distinction between predications to universals and predications to individuals. 12 Cf. Aristotle, Categories 12, where Aristotle also mentions the contrafactual criterion that I employ for metaphysical priority. Aristotle tells us in Metaphysics V 11 that this criterion has already been used by Plato.

On Ascribing Dispositions

169

quantification over a conditional (with ‘S’ for ‘is a chunk of sugar’): (∀x) (Sx ⊃ Wx) The point is that the more simple structure of the disposition ascription to the concrete particular is embedded in the more complex structure of the universal disposition ascription. Thus you can explain a particular disposition ascription without reference to a universal disposition ascription, but not the other way round. Every universal disposition ascription contains the predication of dispositions to particulars. 3. Disposition ascriptions to individual substances are also epistemologically prior to disposition ascriptions to kinds of substances. We could know a lot about the dispositions of individual substances without knowing the truthvalue of any disposition ascription to a kind of substances. I may know that Enrico Caruso had the disposition to sing beautifully. I do not have to know any disposition ascription to kinds in order to have this piece of information about Caruso. This will not work the other way round. In order to know the truth-value of a universal disposition ascription to a kind of substance, we have to know a lot about the dispositions of individuals. In order to discover that humans can think, I first have to have acquaintance with several individual humans and learn about the thinking disposition of these individuals. Thus disposition ascriptions to individuals are epistemologically prior to disposition ascriptions to universals. We can conclude that disposition ascriptions to individual substances are prior metaphysically, logically and epistemologically to disposition ascriptions to kinds of substances.13 Substances Versus Property Tokens 1. What, now, about property tokens or tropes in comparison to individual substances? Well, recall TC: TC This token of the molecular structure of sugar brings with it a token of water-solubility. TC cannot be true if my most favourite chunk of sugar is not water-soluble. For, if the molecular structure of sugar brings with it water-solubility, all 13 There is, however, an interesting kind of counterexamples, pointed out to me in discussion by Markus Schrenk. Of course, three-litre-cars have the disposition to drive 100 km by 3 litre petrol, and unicorns are said to have the disposition only to be caught by virgins. At present, there are no three-litre-cars, and there never were or will be unicorns. It seems as if we know these disposition ascriptions to concrete universals without knowing anything about concrete particulars. In the examples given (three-litre-cars, unicorns) this might be because some kind of a priori-knowledge is involved. However, these examples may also hint at the deeper insight that attributions to universals are not sufficiently represented by universally

Dispositions and Causal Powers

170

bearers of this structure, i.e. all chunks of sugar, are water-soluble, a fortiori also Zucky.14 But of course, Zucky may be water-soluble even if TC is false, because the water-solubility may arise through some other property of Zucky. Thus it might be the case that the laws of nature change in such a way that (a) it is no longer true that Zucky’s molecular structure brings with it watersolubility, but that (b) some other property of Zucky brings with it that disposition. Hence, disposition ascriptions to substances are metaphysically prior to those to tropes. 2. The argument concerning logical priority recurs to the logical form of the propositions. But what, exactly, is the logical form of disposition ascription to tropes? If we consider TC, such an ascription poses a certain relation B (‘brings with it’) between two tropes, between a certain structure trope m and a solubility trope w: B(m, w) Hence, the logical structure of (TC) seems to be totally different from the logical structure of (S): Here we have a two-placed relation applied to two tropes, there we have a one-placed predicate applied to a substance. It is only when we consider the semantics of B that we see an intimate connection. For ‘m brings with it w’ obviously is meant to imply that m and w are tropes that belong to the same substance. Thus, ‘m brings with it w’ means, roughly, that the thing that has m also has w out of, presumably, physical necessity. This phrasing, of course, features the predication of the disposition w to some substance in a prominent place. This points to the logical priority of disposition ascriptions to individual substances. 3. Finally, disposition ascriptions to substances are also epistemically prior to disposition ascriptions to tropes because we can know a lot about the dispositions of individual substances without knowing anything about the tropes responsible for these dispositions. On the other hand, if we know that a certain trope carries with it a certain dispositional trope, then we already know that the substance to which the first trope belongs also has the disposition in question. Hence, disposition ascriptions to concrete particulars are prior to disposition ascriptions to abstract particulars, as well metaphysically as logically and epistemologically.

quantified conditionals. Cf. L. Jansen and N. Strobach, ‘The so-called materially valid inferences and the logic of concepts’, in Foundations of the Formal Sciences II. Applications of Mathematical Logic in Philosophy and Linguistics (Dordrecht, 2003), pp. 113-118. In this paper, we hint at a possible alternative that construes ascriptions of predicates to universal as relations between concepts that are logically independent from ascriptions to individuals. I hope to work this out in more detail in a future paper. 14 I assume, of course, that being a chunk of sugar is an essential property of Zucky, i.e. that Zucky is either a chunk of sugar or non-existent in any possible world.

On Ascribing Dispositions

171

Property Types Versus Property Tokens The third pair I will compare with respect to priority are property types and property tokens. 1. To begin with, what about metaphysical priority? Can a token m of a certain molecular structure bring with it a token w of a certain disposition when the respective disposition ascription to the structure type is false? It cannot, if the ‘brings with it’ relation is meant to be stronger than a mere accidental cooccurrence in the same substance. Indeed, it is meant to express some stronger relation, as can be seen in the reduplicative phrasing: An M-thing is thought to be a W-thing in virtue of being an M-thing. I argued elsewhere that for a reduplicative sentence, i.e. a sentence containing phrases like ‘in virtue of” or ‘qua’, three things have to be fulfilled:15 x is F qua G is true if and only if (1) x is F, and (2) x is G, and (3) there is the correct conceptual relation R between F and G. Now that conceptual relation – whatever it is supposed to be: conceptual implication, mutual implication, causal relevance etc. – has been shown not to be a statement about certain particulars alone, but a statement involving some kind of generality, either of a necessary, a factual or a statistical kind (‘for the most part’). Thus such a statement about tropes already presupposes the respective statement about property types, because the statement is not meant to express a mere accidental relation. What about the other way round? Can the dispositional statement about the property type be true if the respective statement about a trope is false? If the generality involved is of the ‘for the most part’ kind, i.e. a mere statistical concept, this seems to be a metaphysical possibility. However, in an important respect this would be a quite unattractive position, as this leaves us in a position in which we cannot account for the fact that some tropes of that kind bring with them that disposition, but others do not. This would just be inexplicable. If, on the other hand, the generality involved is a genuine one, no exceptions are allowed, and thus the truth of the statement about the trope is implied by the statement about the abstract universal. To sum up: The ascription of dispositions to property types and tropes are stronger intertwined than the other cases discussed. According to the contrafactual criterion used so far, neither is metaphysically prior to the other. 2. What about logical priority? In section 5.3, I represented the logical structure of the disposition ascription to tropes as claiming a certain relation B (‘brings with it’) between two tropes, i.e. as ‘B(m, w)’. Accordingly, the logical

15 Cf. L. Jansen, Tun und Können, p. 43.

172

Dispositions and Causal Powers

structure of the ascription to property types might be something like the following: (∀x)(Mx ⊃ (∃y)(Wy & B(x, y))) Here, the disposition ascription to tropes (i.e. ‘B(x, y)’) is a central feature of the logical structure of the disposition ascription to property-types. Hence, if this analysis is correct, disposition ascriptions to abstract particulars are logically prior to the disposition ascriptions to abstract universals. 3. What about epistemological priority? Here again, it is important that the ‘brings with it’ relation is thought to be stronger than a mere accidental concurrence. Therefore, as soon as we know that a certain property token m as such is responsible for a certain disposition token w, we know that also the respective property type to which m belongs is responsible for the property type to which w belongs. And, of course, because of the involved generality of the disposition ascription to the property type, we know that the respective disposition ascription also applies to the tokens of that type. Hence, tropes and property types are on equal footing with respect to epistemological priority. Substances Versus Property Types Now disposition ascriptions to tropes are logically prior to disposition ascriptions to property types and on equal footing with them with respect to both metaphysical and epistemological priority. Disposition ascriptions to substances, in turn, are prior to disposition ascriptions to tropes in all three respects, as I have argued before. Hence, by transitivity of priority, disposition ascriptions to substances are prior to disposition ascriptions to property types in all three respects and they are also prior to disposition ascriptions to kinds of substances. Thus the most basic kind of disposition ascriptions – metaphysically, logically, and epistemologically – are dispositions ascriptions to individual substances. Evidence: When Do We Ascribe Dispositions? Ascription of Surefire Dispositions to Individuals Finally, I want to discuss the question of evidence: When do we ascribe dispositions? I will first discuss the predication of surefire dispositions to individuals. This case is simple in two respects: First, it leaves out probabilistic dispositions and deals with surefire dispositions only, i.e. such dispositions that lead invariably to their manifestations in given circumstances. Second, it takes only predication to individuals into account, and leaves out predication to universals. Thus, we have two twofold distinctions yielding four cases that are to be distinguished: (a) ascriptions of surefire dispositions to individuals, (b) of surefire dispositions to universals, (c) ascriptions

On Ascribing Dispositions

173

of probabilistic dispositions to universals, and (d) of probabilistic dispositions to individuals. I will first discuss the simple case of ascribing surefire dispositions to individuals and turn to the other cases in the following subsections. I will base my discussion on the following plausible assumption on the relation between surefire dispositions and their manifestations, the principle of manifestation.16 Indeed, fulfilling this principle is just what it is to be a surefire-disposition: If x has the surefire disposition D to display-M-in-circumstances-C and x is situated in circumstances C, then x will display M. With this principle and the assumption that a certain disposition D to display M in circumstances C is indeed a surefire disposition, the following three propositions are not jointly compatible: D1. Dx (x has disposition D to display M in circumstances C.) D2. Cx (x is situated in circumstances C.) D3. ¬Mx (x does not display M.) This observation gives rise to two different research strategies for disposition ascriptions: 1. The first strategy takes C to be a known finite string of conditions, fixed either by observation or by definition. This strategy will then try to realise this set of conditions C, i.e. D2 is made true. In that case, it is not possible for both (D1) and D3 to be true. Thus, if the manifestation does not happen, i.e. if (D3) is true, then D1 cannot be true, i.e. x cannot have the disposition D to display M in circumstances C. If we have evidence of such a kind as produced by this research strategy, we cannot rationally ascribe disposition D to x. Thus, this strategy is a method to falsify disposition ascriptions. 2. The second strategy starts out agnostic about D2, but is dogmatic about the truth of D1: We assume that D is just that disposition, that is being possessed by x, and the task is, to find out more about the circumstances for which D allows x to display M. If we follow this strategy, we realise some set C* of conditions. If x does not display M, then D3 is true. Thus, D2 must be false, which in turn implies that, in fact, x is not yet situated in circumstances C, i.e. that C* is not C. This research strategy is no strategy for a straightforward falsification of a disposition ascription. What is being falsified here are

16 In fact, this assumption has already been formulated by Aristotle in Metaphysics IX 7. A discussion of this passage can be found in L. Jansen, Tun und Können, ch. 5. In what follows I suppress time-indices; for a discussion of the temporal relation of dispositions to their manifestations cf. also L. Jansen, Tun und Können, ch. 5.

174

Dispositions and Causal Powers

hypotheses about the circumstances in which D will manifest itself. There is another plausible principle about dispositions, which yields a third research strategy. Whereas the other two research strategies are strategies to falsify disposition ascriptions, this third strategy allows the verification of certain disposition ascriptions. The underlying principle gives us positive reasons to ascribe dispositions. I call it the principle of enabling: If x displays M in circumstances C, x must have had the disposition D, which allows for a manifestation M in circumstances C. Thus, 3, if we have empirical evidence that x is both in C and is displaying M, we can infer that x has the disposition D to display M in circumstances C. This is how we can verify a disposition ascription. Of course, the evidence for D can be only as strong as the evidence we have for C and M. But this is a problem of empirical reasoning in general, not a problem special to dispositions. If Zucky dissolves in water, it must have had the disposition to dissolve in that liquid in the first place. If a compass needle orients itself to the north, it must have the disposition to do so. Sometimes, however, we will have difficulties to find out, whose disposition is at work. If John swallows a pill, and his headache pain is relieved after that: was it really a disposition of the pill that did the work? Maybe the pill was of no causal relevance at all, and it was the water with which he took the pill, or the swallowing, or the beautiful sunshine, or some other feature of this situation. As a rule, this problem occurs with all active dispositions, i.e. with dispositions of things to cause a change or a process to occur in other things. With passive dispositions this problem does not occur. For passive dispositions are dispositions of things to suffer changes. Thus by definition the passive dispositions always are in the things that undergo a change or a process. The bearer of the active disposition, on the other hand, is not so easily known. This is something that has to be detected by empirical science. Ascription of Surefire Dispositions to Universals Science is in general not so much concerned with disposition ascriptions to individuals but with generalised statements, with disposition ascriptions to universals. Scientists do not so much want to know whether Zucky is water-soluble or not, but whether sugar is. They do not care whether Black Beauty can digest hay, but whether horses have this disposition. Of course, the evidence used by scientists may consist in disposition ascriptions to individuals. But what is printed in the textbooks are generalised statements, or disposition ascriptions to universals. That Black Beauty can digest hay will not appear in any biology book. What we learn in biology is that all horses digest hay – or, in short, that horses digest hay, and in chemistry, that (all) sugar dissolves in water. It is an old and still intriguing problem how universal sentences like these can be justified empirically. But this is a general problem of all empirical research, and not a problem that is particular to disposition ascriptions.

On Ascribing Dispositions

175

Ascriptions of Probabilistic Dispositions to Universals and Individuals So far I discussed surefire dispositions only, because the principle of manifestation, by definition, is only valid for these dispositions. Of course, there may also be other kinds of dispositions. I will call such non-surefire dispositions ‘probabilistic dispositions’, because they bring about a certain probability for a manifestation. At times, they are also called ‘propensities’.17 Probabilistic dispositions bring it about that their bearer will react with such-and-such a probability under such-and-such conditions in such-and-such a way. As the manifestation is only probable, a missing manifestation M in a given situation is no evidence for the lack of a propensity to manifest M in this situation. On the contrary, there may be a high probability for M, although by chance no manifestation will arise. A probabilistic variant of the principle of enabling is still valid for propensities, though: If x has the propensity DP to display M in circumstances C with probability P, and if x is situated in circumstances C, then x will display M with probability P. Of course, this variant is much more undetermined. Any non-manifestation of M might be compatible with the ascription of a propensity DP for M, for 0 < P < 1. And a single manifestation of M does not tell us anything about the probability with which M occurs. To test such probabilities, we therefore need numerous series of observations and experiments. For example, we may observe the behaviour of a huge number of atoms of, say, a certain uranium isotope and determine the time it takes till half of these atoms have decayed. In this way we determine the half-life period of this isotope. We have thus established that atoms of this isotope have the disposition to decay with a probability of 50 per cent within this stretch of time. This disposition is, to my knowledge indisputably, not a surefire disposition. It is an example for a genuine probabilistic disposition. There may be other cases, where genuineness may justly be disputed. Tests show that only a certain number, say n out of hundred people, will taste something if a probe of rat-poison is put on their tongue. Now, is there any disposition we can universally ascribe to humans? Do humans have the probabilistic disposition to taste rat-poison with a success rate of n per cent? Such a disposition ascription will not explain why always the same people do or do not taste the rat-poison. What we are likely to say in this case is that n per cent of humans have the surefire disposition to taste rat-poison while the rest of the human population is lacking this disposition. Thus not every statistical result is eligible for postulating a genuine probabilistic disposition. Another example: Most medical tests yield only statistical results. Do these tests establish genuine probabilistic disposition ascriptions? I would say they normally do not. Aspirin, for example, is said to cure headaches. But aspirin cannot in all cases 17 Cf., e.g. K. Popper, A World of Propensities (Bristol, 1990). For a detailed discussion of theories of objective probabilities cf. J. Rosenthal, Wahrscheinlichkeiten als Tendenzen. Eine Untersuchung objektiver Wahrscheinlichkeitsbegriffe, (Paderborn, 2004).

176

Dispositions and Causal Powers

be applied successfully; it fails in, say, m per cent of all cases. We may be tempted to explain these failures by attributing a probabilistic disposition. For the case of aspirin is in an important respect different from the case of rat-poison-tasting humans: While human beings differ quite a lot from each other, aspirin pills are thought to be equal in all causally relevant respects. In this they seem to be like atoms of the same isotope. Thus all aspirins should have the same dispositions, and these seem to be probabilistic ones. But if aspirin had the probabilistic disposition to cure headaches in (100 – m) per cent of all cases, this would not explain why there are patients where aspirin always fails to cure the headaches, while it is almost always successfully applied to others. Let us explore the story further: To cure headaches, aspirin has to connect to certain chemical receptors within the human body. And certain people, it is said, do not have these receptors because they have certain deviant genes. Thus it is nothing within the individual aspirins themselves which makes them work at one time and fail at another. It is just that aspirins do not have the disposition to cure headaches without any further qualifications, but, as we should have expected, only the disposition to cure them under certain circumstances. What aspirins do or do not cause depends on who swallows them. Again, we have a mere statistical result and no genuine probabilistic disposition. Aspirin has the surefire disposition to cure headaches under conditions C; it only happens that conditions C obtain in only (100 – m) per cent of all cases. We see that not every statistical result is evidence of genuine probabilistic dispositions. In many cases we just have individuals with different causally relevant properties. In other cases we have individuals that share all causally relevant properties, but the necessary manifestation conditions do not always obtain. The only case that is of interest for the ascription of genuine probabilistic dispositions is if individuals with the same causally relevant properties behave differently though all necessary manifestation conditions are obtaining. Conclusion By way of conclusion, I will sum up the results reached in this paper. Disposition ascriptions occur both in daily life and in scientific contexts. They can contain action related information and may lead to scientific explanations. Often, disposition ascriptions are elliptical. Communicational success about disposition ascriptions is possible, because the context of utterance supplies the information necessary for disposition ascriptions that are thought to give action-related information. This is not an option for science: Scientists, therefore, should make as explicit as possible what manifestations they are talking about and in which circumstances these manifestations are supposed to be displayed. We can ascribe dispositions both to particulars and to universals, to concrete and to abstract things. However, disposition ascriptions to concrete particulars, i.e. to Aristotelian substances, are the most basic kind: They are prior metaphysically, logically and epistemologically to disposition ascriptions to any of the other three ontological categories. The principle of manifestation and the principle of enabling provide a variety of strategies to verify or to falsify

On Ascribing Dispositions

177

disposition ascriptions, though special precautions have to be observed in the case of probabilistic dispositions.

This page intentionally left blank

PART 2 Dispositions and Causal Powers in Science

This page intentionally left blank

Chapter 8

An Extended Semantic Field of Dispositions and the Grounding Role of Causal Powers Rom Harré

Introduction The importance of dispositions and causal powers in philosophy of science is tied up not only with attempts to recast psychology as the science of active individuals carrying out projects of various kinds but with metaphysics in general. However, they are also essential, so many philosophers now believe, for setting up a defensible version of Scientific Realism. That will be focus of the discussions and analyses set out in this paper. Of course, one hopes that the rather narrow focus of the discussion will not preclude some general metaphysical insights from emerging. The problem situation amounts to this: the triumph of the physical sciences has been achieved by the imaginative step-by-step construction of ever deepening hypotheses about the causal mechanisms underlying observable phenomena. Each phase has involved a step beyond the current domain of what can be observed. Experimental science, particularly in the development of instrumentation, has followed this stepby-step imaginative enlargement of the human umwelt with revelations of new domains of observables. Diseases are the result of the activities of microorganisms now observable in detail thanks to advances in microscopy. Chemical reactions in solution are the result of the redistribution of ions the transport of which can be followed by ingenious equipment, such as the ultramicroscope. However, from the very beginning of the physical sciences imagination has outstripped technology. Outstripping technology it has also outstripped the reach of concepts at home in the world of everyday perception, and even its instrument-driven enlargement. However, Kant1 and Wittgenstein2 have forcefully reminded us that nonsense results from trying to apply empirical concepts beyond the ‘bounds of sense’, that is for the construction of foundations of physics beyond the reach of iconic models based on sources among observable natural kinds. Schrödinger’s cat paradox 1 I. Kant (1788), The Metaphysical Foundations of Natural Science, trans. J. Ellington (Indianapolis, 1970). 2 L. Wittgenstein, Philosophical Investigations (Oxford, 1953).

182

Dispositions and Causal Powers

is a vivid demonstration of the nonsense (contradiction) that results in trying to apply empirical concepts (observables) to the unobservable substrate of quantum phenomena, contrary to Kant’s prohibition. What can we do? a. Perhaps we should subside gracefully into positivism, à la Mach3 or maybe the smart new version proposed by van Fraassen.4 b. Would it be best to follow Einstein in declaring quantum mechanics incomplete? c. The third way is to follow William Gilbert,5 Roger Joseph Boscovich6 and Immanuel Kant7 in developing causal powers concepts to apply beyond the reach of experimental display. Can a domain of nature, not so far beyond experience to be wholly in the Kantian noumenal, nor yet bound into inprinciple observability by a reliance on concepts drawn from how we must describe what can be perceived, be provided with a repertoire of working concepts? If we adopt option c, we encounter a rich collection of philosophical problems, many of which are addressed by the authors of the other chapters in this book. There are several of these topics to which I will give only very brief consideration, before turning to the main issue I wish to discuss: the extensions to our repertoire of concepts necessary to accomplish the task of presenting an adequate philosophy of physics. The Logical Form of Dispositional Attributions The breadth and variety of dispositional concepts will concern us below. For the moment, in setting out some examples, we can think of such dispositions as the tendencies of things to behave in certain ways, and of their powers to bring about effects. Tendencies and powers are ascribed both to particulars and to material substances. For example, ‘My car has a tendency to pull to the right when braking’, ‘Water has a tendency to flow downhill’, ‘The liquid in this canister can (has the power to) etch this glass’, ‘Lemon juice can (has the power to) dissolve egg shells’. Evidently, the analysis of such statements will involve devices for expressing conditionality, under what circumstances the disposition or power is likely to be displayed. It will also involve devices for expressing categoricity, since the disposition or power is supposed to be possessed whether it is manifested or not.

3 E. Mach, The Analysis of Sensations (Chicago, 1914). 4 B. Van Fraassen, The Scientific Image (Oxford, 1980). 5 W. Gilbert, De Magnete (London, 1600). 6 R.J. Boscovich (1763), Theoria, Venice, published as A Theory of Natural Philosophy (Boston, 1966). 7 I. Kant (1788), The Metaphysical Foundations of Natural Science.

An Extended Semantic Field of Dispositions

183

It will also require devices for expressing relations, since powers not only have a source but also a target, that on which they act. Despite many attempts to express these aspects of the statements with which we attribute dispositions and powers to appropriate logical subjects by the use of the established techniques and concepts of Russellian logic none has proved satisfactory. More about this below. The Role of Ceteris Paribus Conditions As Markus Schrenk (this volume) points out the seeming ubiquity with which statements of natural laws include ‘ceteris paribus’ (henceforth abridged ‘CP’) clauses is problematic. Surely, the laws are supposed to describe regularities in the behaviour of material things and substances that are without exceptions. As far as my own uses of the CP qualification are concerned, I must admit to being somewhat careless in how I express what I mean by it. There is some ambiguity in the meaning of the word ‘hold’. If a law is stated to hold ceteris paribus, this ought to mean only that it will be seen to hold with occasional exceptions. To preserve the status of the law, these exceptions ought not to be treated as potential or actual falsifiers of the generality of the holding of the law. There must be an explanation of the seeming failure of the law in this or that particular case. Bhaskar8 has argued that any (and perhaps every) case of failure that can be justifiably dismissed as a refuting instance by the use of a CP qualification can be put down to the joint activity of unpredictable combinations of causal powers brought to bear in some quite specific situation. The action of a causal power can be seen to generate regularity, fit to be expressed in an exceptionless ‘law’, only in the artificial conditions contrived in experiments. Exceptions to a law might also be put down to variations in the conditions under which a causal power is supposed to act. However, as was pointed out in discussion, this kind of exception is really just a case of the more general Bhaskarian point. My introduction of the CP condition on putative laws of nature was intended to refute the Machian or positivistic analysis of law statements. On this account, a law statement is just a mnemonic device for recalling any member of a conjunction of instances of a certain correlation between types of phenomena. An exception to a regularity either refutes the projection of the regularity into the future, rendering it useless for making predictions, or it is not to be entered into the Machian conjunction. However, the rejection of the exception this way would be unmotivated. Natural Kinds There is a connection to be made between ascriptions of dispositions and powers in statements that purport to be law-like and the assignment of the particulars to which 8

R. Bhaskar, A Realist Theory of Science (Brighton, 1978).

184

Dispositions and Causal Powers

they are ascribed to natural kinds. If a certain medicine does not cure pneumonia, that is lacks the power, as expected, it is not sulphonilamide. The distribution of dispositions and powers in distinct clusters is centred on particulars as instances of kinds. The dispositions to be malleable, ductile, and to look yellow and so on are a distinctive cluster of dispositions, uniquely attributable to just one metal, ‘gold’. The typifying cluster of dispositions must be referable to the best current account of the nature of gold as a natural kind. Counterfactual Conditionals A traditional response to the ubiquity of CP qualifications and the categoricity of the temporal persistence of powers and dispositions when undisplayed or unexercised has been to use the notion of a counterfactual conditional to make sense of the conditionality of displays of dispositions. While indeed the display clause of a disposition or power attribution admits of both the indicative and the subjunctive mood for the verbs in antecedent and consequent, the inclusion of a clause referring to the nature of the particular or substance in question grounds the range of moods without requiring the invocation of anything so unsatisfactory as ‘possible worlds’. As Leibniz9 insisted there is only one possible world, this one. There may be no formal analysis of dispositional or powers talk, so that we will have to rest content with various discursive analyses, in a variety of dialects. The vernacular will be adequate for some cases, and the language of physics will be adequate for other cases. There may be no common form to all the cases of interest. There may be in this as in other central topics in philosophy, a field of concepts linked by family resemblances, and no common essence. Dispositions and Powers in Philosophy of Psychology The most famous example of the use of dispositional concepts to analyse the meaning of the psychological terminology of the vernacular is Ryle’s Concept of Mind.10 The preservation of the status of persons as agents is a desideratum for moral philosophy. It is no less important for psychology. There are human powers and there are natural powers. All kinds of questions emerge around the distinction. However, in this chapter, I do not intend to discuss human powers and psychological dispositions.

9 G. Leibniz, Philosophical Papers and Letters, (ed.) L.E. Loemker (Chicago, 1956). 10 G. Ryle, The Concept of Mind (London, 1949).

An Extended Semantic Field of Dispositions

185

The Horrible Character of Statements by which Powers are attributed to Particulars An early proposal (Harré and Madden)11 for an informal analysis of the kind of statement with which power attributions are made runs as follows: ‘X has the power to Y’ can be analysed as ‘If X is in conditions C, then it will (can, could, would etc.) bring about Y by virtue of its intrinsic nature, N, ceteris paribus.’ A glance at this analysans will surely send a shudder through the hardiest philosopher. The words and phrases printed in italics are among the most intractable in the catalogue of philosophically difficult notions. ‘If … then …’ with various moods of the verbs has proved not to be susceptible to analysis in terms of Russellian logic, nor does modal logic seem to do much better, particularly if the relevant verbs are in the subjective mood. ‘Bring about’ challenges the one time hegemony of event causality. What analysis will be adequate for substance or thing causality? How is the notion of ‘natural agency’ to be explicated? Is it best thought of sui generis or as a stripped down variety of human agency? What relation is referred to in the phrase ‘by virtue of’? What role does the intrinsic nature of something play in its efficacy if it is an agent? The hierarchy that seems to open up when the ‘powers/natures’ structure is applied to the physical sciences has some special difficulties to be explored below. Alternatively, should we avoid the material mode and think of the phrase as referring to the discursive relation ‘explains’? What about ‘intrinsic nature’? If there is such a something as an intrinsic nature how is the intrinsic nature of a particular related to the intrinsic nature of a kind, the kind to which that particular belongs? How are the dispositions displayed by an instance of a kind related to what it is to be of that kind? Perhaps these are two sides of just the one coin. As I have already argued, there is a close tie between the role of CP clauses and the alleged intrinsic natures of the particulars or the substances to the behaviour of which the CP qualification is taken to apply. Two Provisional Taxonomies Bearing in mind that I want to avoid slipping into any suggestion that this family of concepts is based on a common essence underlying each and every one there do 11 R. Harré and E.H. Madden, Causal Powers (Oxford, 1975).

Dispositions and Causal Powers

186

seem to be sufficient similarities to warrant attempts at constructing some provisional taxonomies. Starting with the vernacular there seems to be a broad but imperfect distinction between dispositions the display of which implicates the exercise of an active powerful particular and those which implicate a passive liability. Whenever an agent engenders an effect, it must ‘impact’ a patient which suffers it. The Existing Family of Dispositional Concepts Dispositions (Potentialities) / \ Active (Powers) Passive (Liabilities) / \ / \ Capacities

Capabilities

Pronenesses

Tendencies

Sometimes ‘tendency’ is used with a hint of activity about it (Cartwright12). Like most of the concepts in this scheme, not only are they are part of a field of family resemblances but they are clusters of family resemblance themselves. A Taxonomy of Property Types Properties / Occurrent /

\

\ non-Occurrent / \

Observable

Unobservable

Observable

Unobservable

(Perceptible)

(Imperceptible)

(Dispositions 1)

(Dispositions 2)

It is tempting to suppose that there is a common essence to the members of this taxonomic scheme. If these are all properties, must they not all be properties of something, of some substance? Could any sense be made of the practice of indexing a property specification with a latitude and longitude? To soften up our intuitions do we not index an earthquake with an epicentre, at, say 23N and 3W? Of course, in giving a deeper account of earthquakes, tectonic plates, very thing-like beings, are cited as the subjects of predication of the relevant relative motion words, but more of that anon. Is there a generic schema, a kind of template for bringing out the structure of concepts in these taxonomies? It is important to get away from abstract nouns in doing philosophy so the concepts to be studied are expressed as follows: ‘X is

12 N. Cartwright, Natures’s Capacities and their Measurement (Oxford, 1989).

An Extended Semantic Field of Dispositions

187

disposed to …’, ‘X tends to …’, ‘X is capable of …’, ‘X is prone to …’, ‘X has a capacity for …’, ‘X is liable to …’, etc. In this form, it is easy to see that all these concepts are relational. A Working Generic Template The content of the use of a verb such as any of the above includes at least: a. that there is a typical context sensitive display: conditional. b. that there is a persisting property without which the display would not occur: categorical. c. that the display may not occur though the relevant property persists. d. that the persisting property can be either active or passive. Thus, ‘capabilities’ will have typical displays different from ‘tendencies’. The persistent property relevant to ‘proneness’ is passive, while that relevant to ‘capacities’ is active. (My slipping back into abstract nouns is for stylistic convenience only!) Note also that items a and c attract the possibility modality, while items b and d attract the necessity modality, if, as I believe they must, the persisting properties are part definitive of the natural kinds appropriate to a certain scientific discipline. Some of the difficulties that have beset the analysis of this family of concepts have arisen when a philosopher has tried to complete an analysis using either a or b but not both. Despite his invention of the logical type ‘semihypothetical’ Ryle’s strategy is to use only dispositional concepts in the analysis, favouring item a. Russell, suspicious of anything smacking of modality, adopted all sorts of devices to hang on to concepts only conforming to item b. Causal Powers, Dispositions and the Defence of Scientific Realism The basic principle of Scientific Realism is that patterns among observables are to be accounted for by patterns among unobservables. What these patterns are and how the concept of ‘accounted for’ is to be analysed is the topic of what follows. Types of Patterns of Observables Patterns of observables can be analysed as truth functions of more elementary observables, e.g. p and q, p or q, not-p, nor both p and q, if p then q. 1. Most truth functions of observables are not patterns of jointly occurrent properties of material individuals. Only a conjunction of observables has

188

Dispositions and Causal Powers

this character. Perhaps this is why it attracted both Hume13 and Mach as the logical form of regularities. Hume’s account of regularity is not fundamentally conditional. For example, a substance is the mythical support of a joint contemporaneous occurrence of groups of properties at the same place, a view reiterated by Mach. A causal connection, for Hume, is the mythical support for a sequential co-occurrence of pairs of properties, again a view reiterated by Mach. 2. Dispositions are manifested as sequences of observable properties of material individuals, the stability of which is expressed in unrestricted conditionals. In order to justify the ascription of a disposition to a material particular that is not currently displaying the property referred to in the consequent of the conditional attribution, the disposition must be grounded in some occurrent property of that individual, as its nature or part of its nature. In the physical sciences, the natures of powerful particulars are either taken to be microstructures of further particulars or the result of embedment in macrostructures. E.g. the inertial mass of a material particular derives from the rest of the universe. For many dispositional properties, the grounding will call for attributions of unobservable properties to the material individual in question. The rationality of this move requires a commitment to Scientific Realism. Why? The Regresses that Give Birth to Scientific Realism It has recently been suggested that it is helpful to contrast two ways in which regresses can proceed. In heterogeneous regresses, at some level in the regress the ontological type of the supporting level changes from that of the supported levels to something different. For example, the regress of explanations of an agent’s skilled performance in cognitive psychology proceeds a few steps by invoking unobservable cognitive processes, for example Jerome Bruner’s schemata14 such as those we use when judging the value of coins.15 It is not long before neural processes in the brain are introduced to describe the mechanism or tool with which an agent acts skilfully. In a homogeneous regress, the ontological type of the beings in the supporting level remains unchanged however far the regress is pursued. Wittgenstein’s examination of the regress of rules supported by yet more rules, convinced him that this progress could not go on ad infinitum, since otherwise action according to a rule would be paralysed by the need to consult yet more rules. His solution was to declare 13 D. Hume, ‘An Enquiry Concerning Human Understanding’, (1748), in Enquiries Concerning Human Understanding and Concerning the Principles of Morals, L.A. SelbyBigge and P.H. Nidditch (eds), 3rd edn, (Oxford: 1975). 14 J.S. Bruner, In Search of Mind (New York, 1983). 15 A visitor to Australia may find it very difficult to manage a coinage in which the $2 piece is smaller than the $1.

An Extended Semantic Field of Dispositions

189

that forms of life are defined by the normative practices that the local folk carry out without further ado. If asked to justify a certain practice that is foundational for a culture a member can say only ‘This is what I do’. The physical sciences seem to involve initially a homogeneous regress or regresses that are bounded by a descent (to follow the prevailing metaphor) into another homogeneous regress that is heterogeneous with respect to that which depends upon it. Chemical reactions among material substances are explained by reference to the causal powers and mutual relations of their constituent material parts, molecules. These causal powers are explained by reference to the causal powers and mutual relations of their constituent material parts, atoms, and so on? So far the regress has looked homogeneous in that it follows the pattern of grounding the causal powers of material things and substances in their material natures, that is in the causal powers and relations of their constituent parts. However, research in physics has sooner or later come to a boundary between the realm of material beings (Kant’s empirical world) and the realm of charges, poles, fields, potentials and so on. Whatever these are, they are not just more material entities of the everyday sort with versions of everyday properties, say Boyle’s ‘bulk, figure, texture and motion of the insensible parts’. One strand of physics has always been heterogeneous, the physics of Gilbert, Boscovich, Kant and Faraday.16 What Sorts of Unobservables Are There to be Invoked? This leads to a provisional survey of the some of the ontologies that have been put to use in Physics. Since the end of the sixteenth century, there have been two competing ontological traditions. Corpuscularianism, derived from Epicurean atomism, invoked a world of indivisible material particles characterized by their bulk, figure (shape), texture (arrangement into larger wholes) and motion. Corpuscles were unobservable by the human senses, but were of the same ontological category as those beings which could be perceived. Dynamism, derived from neo-Platonism and the Hermetic tradition, invoked a world of interacting patterns of occult forces and powers. Occult powers were quite distinct ontologically from anything that could be perceived by the human senses. 1. Conservative ontology of Newtonian physics restricts unobservables to a subset of observables, the bulk, figure texture and motion of the insensible parts. This is tied in with the methodology of iconic model building, e.g. kinetic theory of gases, crystal lattices etc. use models of currently unobservable states

16 M. Faraday, Experimental Researches in Electricity (London, 1951), vol. 1, § 1: 3.

190

Dispositions and Causal Powers

and processes built exclusively from mechanical, that is observable, concepts. Using these concepts for occurrent grounding of dispositions conserves the substance-attribute metaphysics and the subject-predicate grammar. 2. Radical ontology of classical physics invokes causal powers, e.g. field potentials as observable dispositions with charges as causal powers providing occurrent groundings. Methodologically this requires a move beyond the domain of iconic model building where one is using observable substances, entities, properties and processes as the sources of concepts for constructing models, and so as the means for creating theoretical concepts. Both novel dispositional concepts and causal powers are required. Whatever it is that is constitutive of the unobservable regions of the world cannot be justly captured by the use concepts based on and developed for the description of observable phenomena. Bohr17 seems to have made use of an analogue of Kant’s famous prohibition on using empirical concepts constitutively beyond the bounds of sense. Kant allows that we can use the concept of the Cartesian immaterial soul regulatively, to manage our moral lives. However, we may not use this concept constitutively, as a proper and adequate way of understanding the nature of the noumenal or true agentive self. Bohr’s boundary, which can be transcended only by a regulative use of empirical concepts, is the boundary set by the limits of experimentation. It is convenient to think of the causes of tracks in a cloud chamber as produced by electrons as small, charged material beings, but that is to make a regulative use of the concept of a locatable material thing. We may not, on pain of incoherence, use that concept constitutively, to say what an electron is or indeed what really generates the track. Neither may we use the wave concept constitutively. However, finding a way in which concepts might be found or created with which to transcend the level in physics which mimics the noumenal/phenomenal boundary that Kant identified, is not enough. We are still working with a naïve account of observable dispositions. The taxonomy sketched above, however rich it is, is not rich enough. We need to find a way to incorporate human practices into the conceptual repertoire. The sciences are human practices and what they reveal about the world is the result of the performing of those practices. What those practices allow us to know, though always in the context of an experimental procedure, is something dispositional about the world. J.-M. Monnoyer makes the point that the case of grounding psychological dispositions in occurrent neural states is only analogous to the case of physical dispositions. So we have the key move in physics poised between two analogies. There is the analogy of the physical regress of dispositions and natures to that of cognitive science and the development of neural hypotheses to refer to whatever grounds cognitive skills. This leaves open the possibility of terminating and completing the 17 N. Bohr, Atomic Physics and Human Knowledge (New York, 1958).

An Extended Semantic Field of Dispositions

191

regress by drawing on a heterogeneous ontology for grounding the dispositions displayed in experimentally produced phenomena. There is also the analogy of the noumenal/empirical distinction of Kant’s by which Bohr makes clear the limitation of empirical knowledge to the behaviour of apparatus/world complexes as they are run in certain distinctive human practices. The final level can only be described in terms of qualified concepts of causal powers. Elaborating the Semantic Field of Dispositions: Niels Bohr’s Innovation a. Manifest phenomena, e.g. tracks, are attributes of complex beings, apparatusworld entities, scientifically indissoluble. Wave functions describe these, not a hidden realm of quantum states. To understand Bohr’s philosophy of physics we need a new dispositional concept, the Gibsonian ‘affordance’. An affordance is a disposition, displayed in contexts specified by some human practice. The antecedent of the ‘if … then …’ pair that expresses an affordance involves a specific human practice, fixing the context of display. A floor affords walking; a cloud chamber affords electron-tracks. What walking a floor affords will depend on the project a human being has in engaging with it. To take the example offered by L. Jansen (this volume), a bridge may afford walking to a pedestrian but it may not afford walking to that pedestrian’s elephant. In the case of the bridge, there is a finitely researchable question as to why the bridge has these seemingly contradictory affordances, since the nature of the bridge can be empirically investigated. There is no such resolution as to why the world seems to afford both tracks (electrons) and interference patterns (waves). We will simply be bringing to light the same or some other affordances once again in further experiments since there is no way of studying the world in isolation from the apparatus with which it forms a single ontological unit. b. Bohr’s insight is that electrons are not revealed by the use of an apparatus, say a cloud chamber. To talk this way would be to use the concept ‘electron’ constitutively. Such an apparatus is indissolubly integrated into the world. It affords tracks, and thus track makers, that is electrons. However, to continue the Kantian exegesis this use of ‘electron’ is no more than regulative. c. Affordances are observables and so must conform to the substance/attribute grammar which has evolved for describing the world as we perceive it. (This is Bohr’s Correspondence Principle). d. Affordances are grounded in causal powers of the world as realized in the manifestation of the dispositions of apparatus-world complexes. e. There are practical prohibitions on the simultaneous construction at the same place of certain pairs of apparatus types. Some affordances cannot be

Dispositions and Causal Powers

192

displayed together. (This is Bohr’s Principle of Complementarity.) f. Each persons is, inter alia, a Bohrian apparatus. Mauro Dorato (this volume) cites a proposal by Rovelli that is similar to that of Bohr. However, Rovelli’s ‘relationism’ that links apparatus and world is insufficiently radical. To say that the being which displays quantum phenomena is Apparatus + World suggests that there are two determinate beings welded together in the laboratory. It leaves open the possibility that the apparatus could be ‘transparent’, a window on to the world. Bohr’s conception is more radical. In the deep parts of physics we must abandon that conception completely. Phenomena are properties of apparatus/world complexes and the world has the power to engender them only as far as it is indissolubly melded (not ‘linked’) with apparatus. The Concept of a Causal Power adjusted for Physics. Causal powers are possessed by material particulars continuously, but whether or not they are manifested in occurrent properties of some observable particular, depends on contexts, as, e.g. in Bohr’s philosophy of experimentation above. THUS, CAUSAL POWERS ARE UNOBSERVABLES WHICH ARE MANIFESTED OBSERVABLY

A material particular that possesses a causal power ‘brings it about that …’. The root idea is that of efficacy. Efficacy cannot be analysed in terms of occurrent properties. As Hume long ago pointed out, the results of the activity of a causal power are observable, but not the power itself. As an unobservable Hume casts it out. Yet, whether some material being is an agent or a patient, acting or acted upon, is a distinction we can all make in the observable world. We can certainly tell whether something has a causal power, as we watch it in action, though the ‘power itself’ is displayed only in the phenomena. THUS, CAUSAL POWERS ARE THE INITIATORS OR SOURCES OF CHANGE

The dynamicist thesis is simply this: Chains of events are produced by the activity of particulars endowed with causal powers. Events, as Hume and Berkeley18 rightly pointed out, are radically independent and passive. Stephen Mumford (this volume) is quite correct to say that at this point in the discussion philosophical arguments 18 G. Berkeley, Treatise Concerning the Principles of Human Knowledge (Dublin, 1710).

An Extended Semantic Field of Dispositions

193

run out. Whether we count ourselves as Humeans or Kantians depends on the price each is prepared to pay. For an event ontologist nothing is hidden, but nothing is explained. For a substance ontologist explanations are available, but at the cost of drawing on beings the existence of which cannot be displayed in the phenomena of perception. Types of Causal Powers 1. Emergent causal powers: when the powers of the micro-particulars constitutive of the macro-particular are different from the relevant causal powers of the macro-particular. E.g. solubility is emergent relative to the charge distribution of the electron configurations of the molecules. 2. Aggregate causal powers: when the relevant power of a macro-particular is the result of the aggregation of the same type of causal power of the microparticulars. E.g. charge on a thundercloud is an aggregate of the positive charges of the ionized molecules. 3. Derivative causal powers: when the relevant power of a particular is a consequence of being at a certain location in the gravitational field of a more powerful particular. E.g. the weight of a material body in the gravitational field of a star. These concepts provide us with models for transcendental causal powers. That X is an agent is observable, but X’s agency is not. 4. Original causal powers: these analyses are hierarchical in that the causal powers of particulars at one level are explained by the citation of causal powers of their constituent particulars. This analysis requires that we acknowledge the existence of original causal powers, such as the charge on the electron, the gravitational field of an inertial mass, and so on. That a certain causal power is original is defeasible, should it be shown to be emergent, aggregate or derivative. Characteristics of Original Causal Powers Causal powers are attributes of material particulars, which in everyday cases can be individuated, identified and re-identified by criteria based on occurrent properties, that is properties other than the manifestation of those causal powers. For some original causal powers, the deepest, the criteria of identity and individuation could not depend on the existence of material particulars to which they would be attributed. To introduce such beings would be to violate the Kant/Bohr prohibition on using material concepts constitutively. The criteria must be reduced to continuity of spatio-temporal trajectory and consistency of the display of the same causal power (sameness being qualitatively identical manifestations).

194

Dispositions and Causal Powers

Returning to the horrible schema that seems to be the least we can manage with to explicate the meaning of the attribution of a causal power, it seems that any further attempts to sharpen the components of the schema are hopeless. Here it is again. ‘If X is in conditions C, then it will (can, could, would etc.) bring about Y by virtue of its intrinsic nature, N, ceteris paribus.’ In these circumstances, the only plausible strategy is to follow Wittgenstein’s procedure by assembling examples of the use of the concept of a causal power. Physics, chemistry, pharmacology, engineering, animal husbandry, as well as the practices of everyday life, are rich sources of examples of causal powers and their consequential manifesting dispositions. However, I think we can go a step beyond Mumford’s ‘balance of advantages and disadvantages’ move. In the final section, I would like to summarize some of the arguments that might be assembled to defend the dynamicist point of view. Why Should We Adopt the Dispositions/Powers Ontology? Argument 1: These concepts are already in use in physics and chemistry in wellaccepted forms for the transcendent parts of the sciences. Argument 2: By adopting this ontology the shortfall in the defence of Scientific Realism can be made up. In particular, the Borhian version of a causal powers account of deep physics resolves the paradoxes that result from attempts to extend iconic modelling into the quantum realm. The ‘quasi-noumenal’ world may not be one of substances and properties at all. Argument 3: The standard pattern of scientific explanation by citation of unobservable causal processes is retained throughout the domain of physics and chemistry. Argument 4: Leibniz and Boscovich, and then more systematically, Kant made great efforts to show that the Newtonian mechanical properties, both occurrent like shape, and dispositional like solidity, could be accounted for as displays in the empirical world of the activity of unobservable causal powers. Bohr’s philosophy of physics revives the Leibnizean project for the current state of the development of physics.

Chapter 9

What Makes a Capacity a Disposition? Nancy Cartwright

Introduction Many, if not most, of our highly prized ‘laws’ of physics cannot be adequately rendered as statements of regular association among the values of ‘categorical’ quantities, I have argued.1 This is true even if we do not balk at the concept of natural necessity and are willing to add that the associations hold ‘by law’. They are rather ascriptions of capacities. They tell us what capacities a system will have by virtue of having a given property. The law of gravity is one example. A system of mass M has the capacity of strength GMm/r2 to move another object of mass m a distance r away towards itself. I call this the gravitational capacity. My second thesis is a commonly shared one. Ascriptions of capacities do not reduce to conditionals involving only categorical properties. I shall here discuss two questions about these theses: 1. Why think of capacities as akin to dispositions or powers; and 2. Why allow them in science? Before tackling the first question, I shall first try to figure out what features we expect to be characteristic of dispositions and powers themselves. What Makes a Disposition a Disposition? There are a number of features on account of which we might call something a disposition or a causal power. I shall discuss some that I think are particularly promising; that is, I think they are good starting candidates for characterizing a feature of the world that we have very good evidence for, especially in the overwhelming evidence of everyday experience.2 The list I shall discuss here includes some characteristics that do not figure centrally in current philosophical discussion. That is because I think my irascibility is a paradigm disposition and I 1 Cf. N. Cartwright, Nature’s Capacities and their Measurement (Oxford, 1989). 2 Though of course we should not be wedded to the supposition that there is exactly one such feature we are aiming to characterize.

196

Dispositions and Causal Powers

look for characterizations that are sure to include it. There is of course the alternative paradigm for philosophy of science – water-solubility; we may suppose that we also have good reason to accept dispositions of this kind because of their scientific credentials. My focus here, however, is on what I have called capacities and my aim is to understand why I and others should have assimilated them to dispositions. Are capacities – like the gravitational capacity – genuinely like dispositions in significant ways, or was the assimilation a mistake, generated perhaps by unthinkingly lumping all non-categorical features together? For this enterprise we may miss important aspects of similarities and differences if we focus our account of dispositions too narrowly. Here then is my list of promising candidates for what is special about dispositions: a. Substance Causation The causal relata are not events but rather the cause is an enduring substance and the effect a change in another substance. This was Aristotle’s view; so too Kant’s according to Eric Watkins.3 On account of this Watkins ascribes to Kant a causal power view. But this does not seem to be true of what I have called ‘capacities’. It seems, rather, that it is the having of a mass by the first object at a given time that causes the motion of another at that time; and this looks far more like event causation than like causation by an enduring substance. b. Latency Dispositions are not always on display. If that is so, the capacity of one mass to move another is not a disposition. c. Conditionality Dispositions are generally thought to be closely connected with conditionals. The simplest connection is ‘If C, then systems with disposition D will M’ where C is a description of some specific conditions, including perhaps something that triggers the disposition, and M is a manifestation. Stuart Hampshire, however, has argued the contrary. Hampshire maintains that this connection with conditionals holds for what he calls ‘causal properties’ but not for descriptions of human character and disposition, which have mistakenly been assimilated to causal properties. He argues, ‘Such causal properties of things, as being magnetized and being soluble in aqua regia, manifest themselves, if at all, in specific and definitely statable reactions, which can be produced in specific and statable conditions. The incidents which may count as manifestations of human dispositions – of intelligence, ambition, generosity, and honesty – are essentially various and these words are vague, summary, interpretive and indeterminate’.4 It may seem that the gravitational capacity is a paradigm causal property and hence satisfies

3 E. Watkins, Kant and the Metaphysics of Causality (Cambridge, 2005), ch. 6. 4 S. Hampshire, ‘Dispositions’ in Freedom of Mind and Other Essays (Oxford, 1972), pp. 37-38.

What Makes a Capacity a Disposition?

197

Hampshire’s conditionality condition. I think not and shall return to this topic. d. Malleability There are a number of things that can usually be done to dispositions to affect their manifestations. There are three that seem most central: interfering, triggering, and enhancing or weakening. The gravitational capacity does not need triggering – it seems to operate all the time. Nor, it seems, can it be enhanced or weakened. But it can be interfered with. This is a central feature of all the things I call ‘capacities’. e. Two-sidedness There is a distinction between the occurrence of a disposition and its being manifested. There is no such distinction for categorical properties. Of course to use this as a characterization of a disposition we must at the same time also recognize the concept of a manifestation and manifestations themselves may be difficult to characterize. Consider: having the minimal unit of electric charge is a categorical property of an electron. But it is often described as ‘hidden’. We have to do something very special to see that it is there. One might say that in the right kind of experiment the charge ‘makes itself manifest’ in certain experimental results. That is not the sense of manifestation intended when we contrast dispositional properties, that have manifestations, with categorical properties that do not. Capacities are like dispositions in being two-sided in this way. f. Missing from Logbooks Hampshire says of a statement that refers to a disposition, ‘It could not be entered into a logbook of the day’s events opposite some time of the day, or in the annals of someone’s life opposite some definite date.’5 This I take it is in large part what it means to say that the disposition is not categorical. I suppose it is tied up with the Aristotelian idea that the power is an enduring characteristic of a substance, not one that occurs at specific times. g. Constancy of Tendency This is the reason I introduced the idea of ‘capacities’ into my discussion of scientific laws in the first place. The outcomes that occur when the gravitational capacity operates are indefinitely various, but there is something fixed. The first mass is always trying to bring other masses closer to it; we say that a mass always attracts other masses no matter how the other masses actually move. Capacities are modelled on John Stuart Mill’s tendencies;6 I used the word ‘capacity’ to underline that the tendencies I was discussing are tendencies to cause things to happen, not just more general tendencies to behave in some particular way (e.g. always to move in a straight line). The first question that I am addressing in this paper is why assimilate capacities to dispositions. I certainly did so but I am not alone. Here are just some recent examples from Australia, a hot-bed for dispositional analysis. Alan Chalmers says that the appeal to capacities, which we both endorse, ‘serves to capture what is implicit in common sense usage of utterances such 5 6

Ibid., p. 34. J.S. Mill, A System of Logic (London, 1949), book IV.

198

Dispositions and Causal Powers

as “glass is brittle” or “acorns grow into trees”’.7 A more thorough attempt to give an ontology of powers, dispositions and capacities appears in Brian Ellis’s book, Scientific Essentialism.8 Peter Menzies in a commentary on my views9 offers the same analysis for capacities, in terms of conditionals relating categorical properties, that he uses for dispositions. Still, what have capacities, as I discuss them, to do with dispositions or powers? I suppose the connection has to do with failures not being entirely defeating conditions. One can exercise one’s power even though the result is not achieved. But that seems little connected with the central ideas defining power: control, influence, ascendancy, authority.10 Where is the authority? The sway over others? Nor does it have to do with what I isolated as the central feature of capacities – that the system always tries to do the same thing even if the results differ. This leaves me with a puzzle. Mass indeed has the feature under discussion – constancy of tendency – but I do not any longer see what constancy of tendency has to do with either dispositions or powers. This puzzle is exacerbated by returning to Hampshire’s piece on dispositions. One may not agree with all Hampshire’s requirements but I think he is fairly well attuned to the features that characterize human dispositions and character traits, many of which are not part of our current philosophical discourse – and after all, human dispositions are paradigms of dispositions. Some of these are shared by the gravitational capacity and some are not, leaving a mixed verdict. Here are the other characteristics on his list that I have not already mentioned. (The titles and summary are mine.) h. Non-episodic ‘There are short-term and long-term dispositions, but a disposition cannot come into being, then pass away and then come into being again very rapidly’.11 The gravitational capacity shares this feature with dispositions. i. Necessity of Display ‘A disposition must be manifested and must show itself in actual incidents.’12 Again, the gravitational capacity is like a disposition by this criterion. Indeed recall my earlier worry that it is perhaps disqualified because it overfulfils this norm – it seems always to be on display. j. Need for Scrutiny in Ascription To be confident in ascribing a disposition one must review actual incidents, looking especially for counter indications. One must thus have the opportunity for ‘prolonged and continuous study of the 7 A. Chalmers, ‘True Fundamental Laws in a Dappled, Patchwork World,’ Popper Centenary Conference Lecture, ms, Flinders University: Adelaide, 2002, p. 3. 8 B. Ellis, Scientific Essentialism (Cambridge, 2001). 9 P. Menzies, ‘Capacities, Natures and Pluralism: A New Metaphysics for Science?’, Philosophical Books, 2002. 10 Cf. G. Ostler (ed.), The Little Oxford Dictionary of Current English (Oxford, 1986). 11 S. Hampshire, ‘Dispositions’, in Freedom of Mind and Other Essays (Oxford, 1972), p. 34. 12 Ibid., p. 35.

What Makes a Capacity a Disposition?

199

conduct and calculations of the person in question. When one has surveyed many incidents and found virtually no contrary evidence, one can say, for example, ‘‘He is certainly and indisputably generous.’’ What is claimed as certain and beyond dispute,’ according to Hampshire, ‘is that the word “generous” is so far the right word to summarise the general trend or tendency of his conduct and calculations.’13 The gravitational capacity satisfies this requirement, but the requirement does not really seem relevant to it in the right way. k. Manifestations not Necessarily Behavioural ‘Most ordinary character descriptions refer compendiously to a tendency discernible equally in the behaviour, and in the thought and in the feelings, of the subject.’14 This is clearly a feature relevant for human character traits and not for my capacities. l. Wide-Scope Negation The opposite of S has the disposition to X is not S has the disposition to not-X but rather X does not have the disposition to X. This is true of capacities like the gravitational capacity. m. Possibility of Opposite Behaviour ‘To attribute a disposition to someone is never to preclude that he may on some occasion act…in some way contrary to his general tendency or disposition…It is typical of human behaviour… that it allows of lapses…’15 I take it that Hampshire means we are capable of lapses even without something actively interfering with the disposition in question. Sometimes we simply do not act in character but for no special reason. I am irascible and given to nagging, but I do not always explode about my daughters’ messy rooms – and the days I don’t need not even have been particularly good ones. The gravitational capacity is not like this; it never lapses. Without positive interference the canonical behaviour will always be displayed in the attracted object. Perhaps though it is misleading to focus on the gravitational capacity. An atom in an excited state has the capacity to deexcite but the display is chancy. Should we then count a failure of the atom to de-excite in any particular case as a lapse? Hampshire concludes from his list that statements of human disposition or character are ‘summarizing statements’. I am not sure if I agree. What I conclude from reflecting on my list and my worries about whether capacities are dispositions or causal powers is Not Everything that is not Categorical is the Same This is a familiar kind of lesson, but one I think we need to be reminded of since we have a tendency to look for the account of dispositions. There are character traits, dispositions, habits, capacities, powers; they are in humans or in non-humans; they 13 Ibid. 14 Ibid., p. 36. 15 Ibid.

200

Dispositions and Causal Powers

usually derive from some underlying structure, but some may be fundamental. We could hardly expect that there are very many features that all these share. If we wish to fix on one criterion as central it seems to me from studying my list that what must be common is: Two-Sidedness There is a distinction between the occurrence of a disposition and its being manifested. I take it, though, that this is not the standard choice. The primary focus is generally on conditionality. That I think is a mistake. The reasons are Hampshire’s. Human character traits and dispositions are certainly as central a member of this family as anything else. Yet, as he argues, ‘The incidents which may count as manifestations of human dispositions … are essentially various and these words are vague, summary, interpretive and indeterminate.’ So no conditional (or set of conditionals) will capture the content of a dispositional ascription. How Much Like other Dispositions is a Capacity? If we take two-sidedness as the central criterion, capacities like the gravitational capacity do fall into the same family as dispositions, habits, character traits and the like. But how many of the other features associated with this family do capacities share? I want here to focus on two features from our list. The first is conditionality; this is the usual candidate for being the feature that, after two-sidedness, most firmly connects capacities with the other members of the disposition family. The second is malleability; this is my choice. Conditionality Many take this to be the central feature of the capacities that science studies. Peter Menzies for instance in discussing my notion of capacity offers just such a conditional account.16 It looks as if Hampshire too would count the capacity due to gravity as what he calls a causal property – satisfying conditionality – rather than assimilating it to a disposition. On the other hand, it does not satisfy his stringent criteria for the causal properties. First, the incidents that count as manifestations of the gravitational capacity are indefinitely various. What actually happens to a second object when the capacity operates depends on the setting; the second can move anyway whatsoever. If we follow Gilbert Ryle’s usage, though, this will make them not dispositions as opposed to causal properties, but rather ‘highly generic’ or ‘determinable’ dispositions. Ryle explains that verbs for highly generic dispositions … are apt to differ from the verbs with which we name the dispositions, while the episodic verbs corresponding to the highly specific disposition verbs are apt to be the same. A baker 16 P. Menzies, ‘Capacities, Natures and Pluralism: A New Metaphysics for Science?’, Philosophical Books, 2002.

What Makes a Capacity a Disposition?

201

can be described as baking now, but a grocer is not described as ‘grocing’ now, but only as selling sugar now or weighing tea now, or wrapping up butter now.17

Second, all conditional claims linking what happens to any categorical description of the setting are, I maintain, ceteris paribus laws. For many philosophers this immediately makes them vague. For instance, John Earman, John Roberts and Sheldon Smith maintain that the ceteris paribus clause is vague and cannot be stated in a precise form.18 I claim on the contrary that it can be stated in a precise form: ‘If nothing interferes, then…’ This claim is not vague, I argue, because the antecedent refers to a specific state of affairs that either obtains or does not obtain on a given occasion.19 This brings us, though, to a third possible reason why the gravitational capacity resembles Hampshire’s dispositions more than his causal properties. Though not a vague word, ‘interference’ is an abstract word, like ‘good’ or ‘work’. Whenever it truly describes a situation, some other more concrete descriptions will always apply as well. But what description that is depends entirely on context, and there are no rules that can be stated using entirely categorical expressions for setting which categorical descriptions will do in which contexts. For any context there is a fact of the matter about whether something is an interference or not; but there probably is no list – even for that context – of what are interferences. Interference is not only multiply realizable; as many philosophers stress, the list of realizations is open ended. Conditionality, I conclude, is not then a widespread feature of capacities, at least not in any sense in which conditionality goes beyond two-sidedness. Capacities and their manifestations are different, as two-sidedness demands. But there need be no set of conditionals that connects a capacity with specific manifestations. Malleability The central feature that locates capacities in the family of dispositions, I claim, is not conditionality but malleability. All other members of the family of dispositions, habits and character traits seem to have at least one of the three central features of malleability I mentioned: They need triggering, they can be enhanced or weakened or they may produce different manifestations, or no manifestations at all, if they are interfered with. So too with capacities. They are all malleable in at least the last way. They can all be interfered with, and when they are interfered with, there is no guarantee that the canonical manifestations will occur. What does occur, if anything at all, will depend on the type and method of interference and there may be no system to either the possible interferences or the possible affect they have on the capacity’s manifestations. This 17 G. Ryle. The Concept of Mind (London, 1949), p. 118. 18 J. Earman, J. Roberts and S. Smith, ‘“Ceteris Paribus” Lost’, Erkenntnis, 57, 3 (2002): 281-301. 19 N. Cartwright, ‘In Favor of Laws that Are Not Ceteris Paribus After All’, Erkenntnis, 57, 3 (2002): 425-439.

Dispositions and Causal Powers

202

indeed, as we have seen, is one of the chief reasons that conditionality fails. It is also why, according to many, capacities can have no place in science. I shall return to this claim at the end. For now let us consider capacities vis-à-vis the two other ways in which dispositions are commonly malleable. With respect to the other features of malleability, the gravitational capacity that has been my focus seems to sit at the far end. As I said, it does not seem to need triggering and we have no evidence that it can be strengthened or weakened. Can other capacities? First, we should consider what this means. It does not mean that the mass of an object, which brings with it this capacity, can be increased or decreased but rather that the effect of a given mass can be changed. This kind of change, which may seem strange in fundamental physics, is commonplace to economists. Economic relations are now almost universally expressed in equations and the equations tend to be linear in their variables (though perhaps not in their parameters). For instance, here are some equations used by Kevin Hoover20 to relate government spending at a time t (Gt) to taxes at a time t (Tt) and the rate of interest (R): Gt+1 = γ + δ[Gt - γ] + εt+1 Tt+1 = Tt + [(R – 1)/(R – δ)][Gt+1 – δGt + (δ – 1)γ]. The Greek letters are parameters. Look for instance at δ. It represents the strength of the capacity of Gt to influence Gt+1. It is typically supposed that the size of δ can change. In fact the important point for Hoover of writing this as a parameter rather than as a variable (since he assumes that it can vary) is that we have it in our power to change it. Other economists may not be so sanguine about our powers to affect things, but the idea that parameters shift is universal. What that means is that the capacities involved can be enhanced or weakened for a given value of the quantity that brings the capacity with it. For a less formal example, you might want to think about the familiar ‘crossedsticks’ supply and demand curves. What do economists do with these? Typical questions are like this: What happens to price if the demand curve shifts upward or downward? If the supply curve shifts? Shifts in these curves are just shifts in the slope of the line, which are represented by the parameters in front of the variables representing supply and demand in the quantity equations; that is, they represent enhancements and weakenings of the capacities of demand and supply to affect quantity produced. The natural thought about the difference between the most fundamental capacities studied in physics and the capacities studied in economics is that the economic capacities are derived whereas those of fundamental physics are basic. Economic features have the capacities they do because of some underlying social, institutional, legal and psychological arrangements that give rise to them. So the strengths of 20

K. Hoover, Causality in Macroeconomics (Cambridge, 2001).

What Makes a Capacity a Disposition?

203

economic capacities can be changed, unlike many in physics, because the underlying structures from which they derive can be altered. Surely that is in some sense true. But there is a puzzle – at least if current economics practices are well-grounded. It is common in economics to assume that, for the most part, the parameters are independent of each other. Each can be altered leaving the others fixed. Hoover in fact takes this to be part of the characterization of a parameter. I say ‘for the most part’, but not ‘always’. When two parameters are not independent in this sense, however, there is a tendency (which is explicit in Hoover) to suppose that the two that are not independent can be written as functions of an overlapping set of parameters, each of which is independent of the others in the set. So if θ can not be altered independently of φ then there is a set {α,β,γ,…,χ} such that θ = f (α,β,γ,…,χ) and φ = g(α,β,γ,…,χ), and every member of this set can be altered independently of every other. How is this possible if the alterations of the parameters involve alterations of the underlying structure? Why is it typical rather than untypical, that we should be able to change the parameters one at a time when such changes involve us in mucking about with the underlying structure? We usually know virtually nothing about this structure or how it operates. One would expect our interventions to be more like those with a sledge hammer than like surgical incisions. I shall leave this issue for discussion elsewhere since it seems more happily situated in a volume on philosophy of economics than in one on dispositions and powers. There is one fact, however, that has come to the fore in our brief discussion of the malleability of economic capacities that is characteristic of dispositions in general. We see here in the economics case a threefold distinction that is typical for dispositions once they reside in even slightly complex systems and that is often conflated. There is i) the disposition, ii) the property in virtue of which the system has the disposition and iii) the underlying structure that ensures that that property is associated with that disposition. In Defence of Interference I claim that it is characteristic of the capacities that science studies that they can all be interfered with. More strongly, the only way to state a true conditional with the canonical manifestation as consequent is roughly this:21 If the capacity is triggered properly and is not interfered with, then the canonical manifestation will result. We are often told that in science we are not allowed to use terms like interference. Abstract or umbrella terms of this sort are admissible, but they must be accompanied by bridge principles that link them with more concrete terms which already have 21 My formulation here is similar to that of Geoffrey Joseph in ‘The Many Sciences and the One World’, Journal of Philosophy, 77 (1980): 773-791.

204

Dispositions and Causal Powers

strict standards of application. The so-called ‘special force laws’ are the paradigm of bridge principles. The abstract force function F = GMm/r2 obtains when a mass m is situated a distance r from another mass M; the abstract force function F = εoq1q2/r2 obtains when a charge q1 is located a distance r from a second charge q2; etc. We are also told that claims with the general rider ‘if nothing interferes’ in front are untestable, indeed vacuous – they allow anything. In closing I would like to make a number of remarks about these charges.22 The absence of ‘special interference laws’ is not so epistemically damaging as many suggest. The special force laws do tell us when a particular force function obtains, but only for very specific descriptions – the descriptions that appear in our bridge principles. For other descriptions that may be applied far more immediately, such as a truck passing by or the press of the wind, we are just as much on our own without the help of a system of rules as we are in deciding if we can label the truck passing by as an interference. The lack of systematic rules does not mean that we cannot have knowledge about whether a certain kind of occurrence constitutes an interference. Galileo after all knew to use smooth planes for his rolling-ball experiments because he knew he should eliminate the interference of friction with the pull of the earth. Similarly he knew to drop small compact masses and not feathers from the Leaning Tower. And that was long before he could have had any idea whether friction or the wind exerted a force in the technical Newtonian sense. The fact that we cannot identify what counts as interference with respect to a claim does not mean that we cannot test whether that claim is true or not. Consider Aspirins relieve headaches, if nothing interferes. We regularly test claims like this in randomized treatment/control experiments. Nor does it mean that is it is too easy to dismiss disconfirmations. When the predicted result fails to transpire, one can always say that something interfered. But saying does not make it true. And as epistemologists are always reminding us, saying, even when it is true, does not constitute knowledge, or even reasonable belief. We need a good reason for claiming that something is an interference. When we do not have any idea whether a nominated factor is an interference or not, then we equally have no idea how to classify the case. Our intended test is no test at all. It follows that one needs a great deal of information about what might and might not interfere with a process before we can carry out serious tests on the process and that in turn implies that we need already to have a great deal of information about the process itself. That just means that science is difficult, as we already knew, and that it is hard to get started in a vacuum of knowledge.

22 For more detailed discussions see Cartwright’s Nature’s Capacities and their Measurement (Oxford, 1989) and The Dappled World: A Study of the Boundaries of Science (Cambridge, 1999). Paul Pietroski and Georges Rey make similar defences that ‘interference’ is not vacuous in their ‘When Other Things Aren’t Equal: Saving Ceteris Paribus Laws from Vacuity’, British Journal for the Philosophy of Science, 46 (1995): 81-110.

What Makes a Capacity a Disposition?

205

Conclusion What makes everything in the disposition family belong there is, if anything, two-sidedness. Within this family ‘capacity’ seems especially like a power word. Nevertheless, I think that is the wrong way around to look at it. What marks out all capacities as capacities is not primarily that they enable systems to do things, but rather that they can be stopped – they can be interfered with. That does not rule them out of science, though, whether something is an interference in a given situation is a matter of fact; and it is a fact we can know about – though not in any mechanical way. But it is at any rate a big mistake to think that science could or should be mechanical.

This page intentionally left blank

Chapter 10

Causation, Laws and Dispositions Andreas Hüttemann

In this paper I will take a look at what I believe to be the best argument for dispositions. According to this argument we need dispositions in order to understand certain features of scientific practice (pp. 207-9). I will point out that these dispositions have to be continuously manifestable (pp. 209-11). Furthermore I will argue that dispositions are not the causes of their manifestations (pp. 211-13). Nevertheless, dispositions and causation are closely connected. What it is to be a cause can best be understood in terms of counterfactuals that are based on dispositions (pp. 213-9). Why Do We Need Dispositions? According to Galileo, in a vacuum all bodies fall with the same speed. Here is how Salviati, Galileo’s spokesman, argued for this claim: We have already seen that the difference of speed between bodies of different specific gravities is most marked in those media which are the most resistant: thus, in a medium of quicksilver, gold not merely sinks more rapidly than lead but it is the only substance that will descend at all; all other metals and stones rise to the surface and float. On the other hand the variation of speed in air between balls of gold, lead, copper, porphyry, and other heavy materials is so slight that in a fall of 100 cubits a ball of gold would surely not outstrip one of copper by as much as four fingers. Having observed this I came to the conclusion that in a medium totally devoid of resistance all bodies would fall with the same speed.1

There are at least two noteworthy features in Salviati’s argument.2 Firstly, the law according to which all bodies fall with the same speed concerns a situation that is not realized. Secondly, there appears to be some kind of continuity between the actual and the counterfactual. Salviati is able to provide evidence for what would happen if the vacuum were realized. The historian of ideas Amos Funkenstein has

1 Galileo Galilei, trans. H. Crew and A. de Salvio, Dialogues Concerning Two New Sciences (New York, 1954), pp. 71-72. 2 For an analysis of this passage see A. Bartels: ‘The Idea which we call Power. Naturgesetze und Dispositionen’, Philosophia Naturalis, 37 (2000): 255-268.

Dispositions and Causal Powers

208

emphasized that this commensurability of the actual and the counterfactual was a new and characteristic feature of early modern physics.3 Let me turn to the issue of non-realization first. The law in question concerns the behaviour of a system under very special circumstances. It pertains to bodies falling in a vacuum. The question is not so much whether this situation ever gets realized. The point is rather that for the law to be established as well as for its being explanatorily useful, it does not seem to be necessary that it ever will. This situation is typical of many laws. The law that hydrogen atoms behave according to Schrödinger’s equation with the Coulomb-potential is valid in case there are no magnetic or electric fields. Newton’s law of gravitation describes the behaviour of heavy bodies as long as there are no charges and no electro-magnetic fields. Crystals have a certain specific heat, given there are no disturbing factors such as impurities. Both for the testing of these laws, and for their being explanatorily useful, it is not necessary that the relevant conditions ever obtain. The behaviour that laws of nature attribute to physical or other systems is in general not manifest under all conditions. It is manifest under certain conditions only. As long as these conditions are not realized, laws of nature are counterfactual claims. In physics these ascriptions typically concern systems in isolation, i.e. in the absence of disturbing factors. Another way of putting the same point is the observation that laws are often ceteris paribus laws. If all else is equal, i.e. if the relevant conditions are realized, the regular behaviour in question will be observed. So laws of nature describe how physical systems would behave if the systems were isolated, or if other ideal conditions were to obtain. It is sometimes argued that laws of nature cannot be ceteris paribus laws, because if they were, they would be immune to empirical testing. It would always be possible to claim that the relevant conditions are not realized if an attempt to verify the law in question fails. This objection assumes that laws can only be tested by, and can only be explanatorily relevant for the regularities that are implied by the laws, i.e. it is assumed that laws can only be tested or explanatorily useful if the manifestation conditions are realized. I take it that it is simply a fact about the science we have that, firstly, its laws describe ideal situations and, secondly, these laws can nevertheless be tested in, and are explanatorily relevant to less-than-ideal situations. The challenge is to explain how this is possible. Cartwright’s idea is that the introduction of what she calls ‘capacities’ rationalizes our scientific practice: When […] disturbances are absent the factor manifests its power explicitly in its behaviour. When nothing else is going on, you can see what tendencies a factor has by looking at what it does. This tells you something about what will happen in very different, mixed

3 179.

A. Funkenstein, Theology and the Scientific Imagination (Princeton, 1986), pp. 152-

Causation, Laws and Dispositions

209

circumstances – but only if you assume that the factor has a fixed capacity that it carries with it from situation to situation.4

Something carries over from the ideal to the less-than-ideal. It is present and manifest under ideal circumstances. It is present but fails to be manifest in less-than-ideal circumstances. Properties of systems which become manifest provided certain conditions are given are dispositional properties, as opposed to categorical or occurrent properties which are manifest under all circumstances. Solubility and fragility are dispositional properties. The systems in question have these properties under all circumstances, but the property becomes manifest only when certain manifestation or enabling conditions are given. Thus, if we want to understand why ceteris paribus laws can be tested in, and are explanatorily relevant for non-ideal situations, we have to assume that something carries over. The systems in question have dispositional properties. This is Cartwright’s argument for dispositions (or capacities). Put differently: If one thinks that laws of nature are explanatorily relevant only for those situations in which the regularities implied by the laws actually obtain, the use scientists make of these laws cannot be explained. It is only if what the law says somehow carries over from the ideal to the less-than-ideal that sense can be made of the scientific practice as illustrated by Salviati. The assumption of dispositions is the best explanation of certain features of our scientific practice.5 In what follows I will take a closer look at this argument. I am particularly interested in what we have to assume about these dispositions so that they can do the work they are supposed to do. Continuously Manifestable Dispositions The first thing to note is that not all dispositions will do the work the carryingover-argument supposes them to do. Dispositions ought to explain how laws that describe the behaviour of systems under ideal circumstances can be tested in, and be explanatorily relevant for less-than-ideal situations. However, if a disposition fails to be partially manifest in less-than-ideal situations, then it is of no help. If the disposition’s manifestation is an all or nothing affair, it will be impossible to attain empirical evidence for what will happen under ideal circumstances on the basis of the mere presence of the disposition. Thus, if the disposition is what I call a discontinuously manifestable disposition (DMD) then, even if it is true that it carries over from the ideal to the less-than-ideal, it will explain neither how we can test the ideal in non-ideal situations, nor the explanatory relevance of the ideal for the less4 N. Cartwright, Nature’s Capacities and their Measurement (Oxford, 1989), p. 191. 5 Elsewhere I have argued in more detail that theories of laws which do not assume dispositions cannot explain our explanatory practice. See A. Hüttemann: ‘Laws and Dispositions’, Philosophy of Science, 65 (1998): 121-135.

210

Dispositions and Causal Powers

than-ideal. The reason is that DMDs are empirically inaccessible as long as the ideal circumstances are not realized. Fragility is an example of a DMD. A thing is either broken or it is not; fragility cannot be partially manifest. Not all dispositions are discontinuously manifest. Continuously manifest dispositions (CMDs) allow for partial manifestations. If the partial manifestations are continuously ordered they permit an extrapolation. Here is an example which illustrates how one might get evidence for dispositions even though they are not completely manifest. Lithium fluoride is a crystal. Its specific heat can be expressed as follows: c = (12/5)π4nkB (T/θD L )3 V

where cV is the specific heat, if the volume is kept constant, n is the phonon-density, k the Boltzmann constant, T the temperature and D the Debye-temperature – its B value for Lithium fluoride being θDL = 730 K. The law attributes a behaviour to the crystal in case there are no impurities that would work as disturbing factors. Even if this is a disposition that will never be completely manifest, we might nevertheless get empirical evidence for the disposition’s being present. We may proceed as follows: First collect a few samples of impure lithium fluoride crystals. With the help of spectroscopic investigations and other means we will be able to find out the amount of impurities in the samples. We can therefore order them according to the degree that the manifestation condition for the disposition is realized. The fewer the impurities, the more the relevant condition is realized. If we measure the specific heat of all of these samples, we are able to extrapolate to the behaviour of the pure system as the limiting case. What is essential is that the disposition is partially manifest in the non-ideal situation, and that the transition from the less-than-ideal to the ideal is continuous so as to allow for extrapolation. The falling objects in a vacuum as discussed by Salviati provide another example of a CMD. The thinner the medium, that is, the closer we approach the ideal condition – the more the behaviour in question becomes manifest. Given such a continuity we can accumulate evidence for what would happen if the ideal circumstances were realized even if they actually never are. The partial manifestations of the disposition allow for an extrapolation to the ideal situation under the assumption of a continuity between partial and complete manifestation. If the disposition is supposed to explain how the testing of the behaviour of a system under ideal conditions is possible in less-than-ideal situations, we have to assume that it is a CMD rather than a DMD. CMDs as opposed to DMDs not only allow for extrapolation when it comes to testing. It is only if the disposition in question is a CMD rather than a DMD that the counterfactual or ideal situation is of any explanatory interest for the actual or lessthan-ideal situations. The fact that an object is fragile (DMD) does not tell us anything about the behaviour of an object as long as the manifestation conditions for fragility aren’t realized, even though the disposition carries over from the ideal to the less-than-

Causation, Laws and Dispositions

211

ideal situation. In contrast, what we know about the behaviour of falling bodies in a vacuum does help us, e.g. to explain and to predict the behaviour of falling bodies in a medium. Possessing a disposition implies that something carries over from the ideal to the non-ideal. What is constant across the various conditions is the dispositional property. However, it is only in the case of CMDs, i.e. dispositions that are partially manifest in non-ideal situations, that this fact can be exploited in prediction, explanation etc. In the case of DMDs, such as fragility, the carrying over has no implication for prediction or explanation of the behaviour of the system in less-thanideal situations. Galileo’s law of free-falling bodies describes the behaviour of objects in a vacuum. His argument assumes these objects to have CMDs. On the basis of such a CMD we can explain or predict the behaviour of falling bodies in a medium, given that we are able to calculate how the disturbing factor, the medium, affects the falling body. Let me close this section by some remarks on the notion of ‘continuity’. It is not meant to imply that the manifestation of a disposition is a process that takes some time. If there is such a process, then what I am interested in is manifestation as the product of this process. A CMD is manifest given ideal conditions and it is partially manifest (as opposed to DMDs) in non-ideal situations such that there is some kind of continuity between the ideal and the non-ideal. What kind of continuity is there between the ideal and the non-ideal? In the case Salviati describes, continuity is simply assumed. It is assumed that nothing spectacular is going to happen if the vacuum condition is reached. If we have a complete description of the medium (or other disturbing factors), continuity is not just a mere assumption. In that case we can consider the medium plus the falling body as a compound system, and give a complete account of its behaviour. There are laws of composition that tell us how the medium affects the velocity of the falling body. These laws of composition will also tell us what is going to happen if the medium is replaced by a thinner medium. So if we have a complete description of all the relevant factors, it is the laws of composition on the basis of which we can give a quantitative account of the continuity between the non-ideal and the ideal situation. What Is Explained by the Assumption of CMDs? CMDs have to be assumed in order to rationalize our testing and explanatory procedures. Their assumption is the best explanation of the success of certain features of our scientific practice. In the philosophy of science, dispositions do not have the best reputation. They are often associated with occult qualities of pre-seventeenth-century natural philosophy.

212

Dispositions and Causal Powers

With respect to the explanation of regular behaviour of physical objects a threestep development from the scholastics to the twentieth-century seems to have occurred, and the introduction of dispositions seems to be a fall-back to pre-cartesian times. Roughly speaking the picture is somewhat as follows: Before the seventeenthcentury the manifest behaviour of physical systems was explained with recourse to the essences or natures of the systems in question. These essences or natures had certain capacities or dispositions to bring something about. The dispositions were conceived of as causes, and the explanations in terms of dispositions as causal explanations of the manifest behaviour of physical systems. In the seventeenthcentury this conception was abandoned. Physical systems were no longer thought to have dispositions or capacities that were postulated to explain the observed behaviour. It was rather laws of nature that philosophers appealed to. These laws in turn were explained in terms of God’s activity, or in terms of the activity of created spiritual or mental substances (plastic natures, monads). Today the behaviour of physical systems is still explained in terms of laws of nature but these are no longer thought to be grounded in something else. Ultimately it is a brute fact that certain systems behave according to certain laws. The laws themselves need no explanation in terms of God’s activity or in terms of the capacities and dispositions of the scholastics. The introduction of CMDs appears to be a fall-back to pre-Cartesian times. It looks as though CMDs give a causal explanation of why physical systems manifest a certain behaviour.6 But the lesson of Molière’s well-known little dialogue on the virtus dormitiva seems quite plausible. Molière ridicules these kinds of explanations, because they are trivial. To call the cause of the act of falling asleep the virtus dormitiva is not an informative answer. The important point is that CMDs as introduced by the carrying-over argument are not of the same kind as the virtus dormitiva. The explanandum of a CMD is not the manifestation of the disposition of a physical system. CMDs (as opposed to causal capacities) are not introduced as causes of their manifestations.7 What CMDs explain are features of scientific practice. This is why we had to assume that physical systems have dispositions. They explain why we are allowed to extrapolate from less-than-ideal, to ideal circumstances. They explain why the ideal circumstances are explanatorily relevant for less-than-ideal situations. They explain why in experimentation we are interested in placing physical systems in certain surroundings, rather than others. We assume that they possess dispositions that become manifest under these, and not under other experimental conditions. The disposition that lithium crystals have the specific heat 6 Some of Cartwright’s remarks suggest this view. See N. Cartwright: The Dappled World (Cambridge, 1999), p. 28 and p. 66 where she describes natures and capacities as trying to bring about an effect. 7 In fact, it is often the case that some of these dispositions never get perfectly manifested. In these cases the alleged explanandum does not even exist (because the disposition is not completely manifest).

Causation, Laws and Dispositions

213

c = (12/5)π4nkB (T/θD L )3 V

does not explain why lithium has this specific heat rather than another. Dispositions are not meant to give causal explanations of this specific heat. Why the specific heat is the way it is remains a brute fact (unless it can be micro-explained). Rather, what the CMD explains is why this specific heat can be measured even though it is not (completely) manifest. It explains why we are interested in measuring the behaviour of crystals which are as pure as possible. It explains why the law in question can be used to explain the behaviour of impure lithium crystals. Furthermore, CMDs help to understand causation as I will elaborate in the second part of the chapter. Causation The carrying-over argument requires physical systems to possess dispositions. Bodies have the disposition to fall with a certain speed in a vacuum. According to this dispositional interpretation of laws of nature, these laws tell us what would happen if ideal circumstances were realized. In what follows I will argue that counterfactuals relating to what physical systems would do, if certain (ideal) circumstances were realized, provide the basis for claims about causal dependence. Thus, the introduction of dispositions does not only rationalize our scientific practice, it also allows for an analysis of causation. Lewis’s Account and its Difficulties Laws of nature or theories ascribe dispositions to physical systems, they describe how these systems would behave if they were isolated. Causation comes into play when we describe the behaviour of systems that are disturbed and thus fail to be isolated. The observation that a cause is something that makes a difference – and a disturbance is such a difference – has been the motivation for counterfactual analyses of causation. If the cause had not occurred neither had the effect. I will use David Lewis’s account as my own starting point.8 According to Lewis an event c is a cause of an event e if c is part of a causal chain that leads up to e. A causal chain is a sequence of events c, d, e etc. such that d depends causally on c, e depends causally on d, etc. Finally, causal dependence is spelt out in terms of counterfactual dependence. Lewis defines counterfactual dependence among events as the counterfactual dependence between the corresponding propositions O(c) and O(e), which state that 8 See D. Lewis, ‘Causation’, Journal of Philosophy, 70 (1973): 556-567, reprinted in his Philosophical Papers (Oxford, 1986), vol. 2, pp. 159-172, p. 167.

Dispositions and Causal Powers

214

c and e occur respectively. For e to depend causally on c it has to be true that firstly, if c had occurred e had occurred, and secondly, if c had not occurred e would not have occurred either. 1. O(c) □→O(e) 2. ~ O(c) □→ ~O(e) (the Ludovician counterfactuals) Any two events for which these two counterfactuals are true depend on one another causally. The following example is an illustration: Billiard balls A and B roll towards each other, they shortly interact at time tI and move apart from each other in a different direction and with a different velocity at time t2. We take it that A causes B’s behaviour to change, and similarly, that B causes A’s behaviour to change.
A


A


B t1


B t2

tI

Fig. 1 How does Lewis’s account work in this case? Let us focus on A causing B’s behaviour to change. Two conditions have to be met for causal dependence to obtain. Firstly, if A had collided with B, B would have been deflected, and secondly, if A had not collided with B, B would not have been deflected. (see figure 2).

Causation, Laws and Dispositions

215

b*

b
B t1

tI


B t2

b: B’s actual path b*: the path B would have taken if A had not interfered at time tI Fig. 2 According to Lewis, the truth-conditions of the relevant counterfactuals are defined in terms of possible worlds. A□→C is true (at a world w) iff either (1) there are no possible A-worlds (in which case A□→C is vacuous), or (2) some A-world where C holds, is closer (to w) than is any A-world where C does not hold.9 Closeness is spelt out in terms of similarity, and similarity in turn depends to a large extent on shared laws of nature. At any event, the central idea is that we have to consider possible worlds in which A is true and possible worlds in which C is true. A possible world is a way things could be, ‘at its most inclusive’. The counterfactual A□→C is true if among the A-worlds the one most similar to our actual world turns out to be a C-world as well. In the case of the collision of two billiard balls A and B, A’s collision with B caused B’s deflection. The first counterfactual we need to consider is the claim that if A had collided with B, B would have been deflected. The world closest to our actual world in which the antecedent is true is the actual world itself. The consequent is true in the actual world as well. Thus the first of the Ludovician counterfactuals is true. According to the second counterfactual, if the collision had not occurred, B would not have been deflected. According to Lewis for the second counterfactual we have to consider possible worlds in which no collision occurs. The second counterfactual is true (in our world), says Lewis, if the following holds: The worlds in which B is deflected (even though no collision has occurred) are less similar to our actual world than at least one world in which B is not deflected (and no collision has occurred). Lewis’s account has to face some well-known difficulties. First, there are counterexamples of cases in which the relevant counterfactuals hold but that do not pick out causal relations, such as ‘If my sister had not given birth at t, I would not have become an uncle at t.’ The birth of the child determined my becoming an uncle, but

9

See D. Lewis, ‘Causation’, p. 164.

216

Dispositions and Causal Powers

my becoming an uncle is not a causal effect of the birth.10 Thus Lewis’s account is too broad. A second problem has been posed by Philip Kitcher. It concerns Lewis’s account of the counterfactuals’ truth conditions, i.e. its semantics. Lewis relies on possible worlds. The problem that arises concerns our epistemological access to possible worlds. How do we ever get to know, on this account, that a causal relation actually holds? [T]he best semantic accounts make reference to possible worlds, our best epistemological views make knowledge (and justification) dependent on the presence of natural processes that reliably regulate belief, and it is (to say the least) unobvious how any natural process could reliably regulate our beliefs about possible worlds.11

Causation Based on CMDs I now want to suggest a way for counterfactual analyses to circumvent these difficulties. In particular, I want to make plausible why we have epistemic access to what makes counterfactuals true. The essential point is that the counterfactuals that go into an analysis of causation hold in virtue of dispositions of physical systems. Let me explain. The first point is that causes and effects are actual. When we consider whether event c is a cause of event e, we presuppose that c and e are actual or that O(c) and O(e) are true. The truth of O(c) and O(e) collectively entails the truth of the counterfactual 1. But there is no need to state the counterfactual in the first place. For causal dependence among actual events the Ludovician counterfactuals can be replaced by the following two conditions.12 1*. O(c) O(e) 2. ~ O(c) □→ ~O(e) (as before) So we are left with one counterfactual claim. This claim is meant to capture the intuition that if the cause had not occurred, the effect would not have occurred either. What I maintain is that whether or not 2 is true is entirely a matter of laws of nature, considered as claims about ideal circumstances. I will introduce my proposal by way of discussing the billiard ball example. 10 For a discussion of these counter-examples see J. Kim, ‘Causes and Counterfactuals’, Journal of Philosophy, 70 (1973): 570-572, reprinted in E. Sosa and M. Tooley (eds), Causation (Oxford, 1993), pp. 205-207. 11 P. Kitcher, ‘Explanatory Unification and the Causal Structure of the World’, in P. Kitcher and W. Salmon (eds), Minnesota Studies in the Philosophy of Science, vol. 13 (Minneapolis, 1989), Scientific Explanation, pp. 410-505, p. 473. 12 See D. Lewis, ‘Causation’, p. 167.

Causation, Laws and Dispositions

217

So, as before, let us assume that billiard balls A and B roll towards each other, that they shortly interact at time tI and move apart from each other in a different direction and with a different velocity at time t2. With the exception of a short period of interaction both A and B can be considered to be closed or isolated systems before and after the interaction took place.
A


A


B t1


B t2

tI

Fig. 1 Why do we hold that A is a cause of B’s change in behaviour? If A had not collided with B, B would have taken path b* rather than path b – as it actually has (see figure 2). b*

b
B t1

tI


B t2

b: B’s actual path b*: the path B would have taken if A had not interfered at time tI Fig. 2 The fact that B is disturbed by A makes the difference that is relevant for causation. The central idea of my proposal is simple. As I have elaborated on pp. 207-9, laws of nature describe how systems would behave if they were isolated, i.e. they attribute dispositions to physical systems. Laws of nature usually describe counterfactual situations and should therefore be read as saying, for instance, ‘If the hydrogen atom were isolated it would behave according to the Schrödinger equation with a Coulomb potential.’ According to my proposal, it is exactly these kinds of counterfactuals that make true condition 2. Causal dependence is spelt out in terms of counterfactuals

218

Dispositions and Causal Powers

that are based on dispositions. Laws of nature tell us how a system would behave in ideal circumstances.13 We want to say that A’s collision with B caused B to be deflected. The essential conditions are therefore 1* and 2. 1* is fulfilled because the collision has occurred and B has been deflected. 2 turns out to be true because there is a law that tells us that B would have continued along path b* if it had continued to be isolated. If B were isolated it would behave according to the Hamilton equations with the Hamiltonfunction H=p2/2m. Less pretentiously, it is Newton’s first law that tells us how B will continue in the absence of a collision. The counterfactual 2 is true because there are laws (of succession) about what would happen in the absence of the causeevent. These laws describe the temporal evolution of physical systems under ideal circumstances, i.e. they describe ideal processes. This analysis seems to be adequate in general. 1* simply registers that the causeevent and the effect-event have occurred. 2 is made true by a law that states how a system that goes into the effect-event would have behaved if it had remained isolated. A consequence of this proposal is that we do not need possible world semantics for assessing the truth-values of 1* and 2. The truth of the counterfactual 2 can be established as discussed by Salviati. What we extrapolate to in such a situation is certainly not a possible world that is similar to ours. The counterfactual situation envisaged is a falling body in a vacuum – that’s all there is. One may call this a possible world – or a possible mini-world –, however, these worlds are not the kinds of worlds Lewis deems relevant for our analysis. The worlds he considers to be relevant have to be similar to our actual world, and therefore have to include all kinds of things such as the milky way, earthquakes in Turkey, the complete history of the Roman Empire and other facts, all of which seem to be completely irrelevant for whether or not billiard ball A causes billiard ball B to be deflected. To summarize: According to my proposal, causation is a relation among events such that an event c is a cause of an event e if c is part of a causal chain that leads up to e. A causal chain is a sequence of events c, d, e etc. such that d depends causally on c, e depends causally on d, etc. (as in Lewis). Two (actual) events c and e causally depend on one another if the following conditions hold: 1*. O(c) O(e) 2. ~ O(c) □→ ~O(e) Counterfactual 2 is true in virtue of laws of nature, which describe the temporal evolution of systems under ideal circumstances.

13 For an analysis in a similar spirit see D. Dieks, Studies in the Foundations of Physics, PhD thesis (Utrecht, 1981), chap. III, Dieks describes systems such as the billiard balls as semi-closed systems.

Causation, Laws and Dispositions

219

Let me finally discuss the advantages of this proposal. Firstly, it excludes some of the counter-examples against Lewis’s account such as those depending on linguistic conventions. According to the proposal, condition 2 has to be true in virtue of laws of nature. Thus, counterfactuals that are due conventions have to be excluded. Secondly, if one accepts the account of laws in terms of testable dispositions (pp. 209-11), Kitcher’s problem of the epistemic access is resolved. Causal claims rely essentially on laws of nature. Laws of nature describe dispositions of systems. These dispositions can be measured as exemplified in the case of the specific heat of the lithium crystal and the case discussed by Salviati. They are therefore epistemically accessible. Dispositions allow counterfactuals to be true in virtue of facts in this world only. Thus, assuming that physical systems possess dispositions not only explains various features of scientific practice, it also provides an account of causation.

This page intentionally left blank

Chapter 11

Can Capacities Rescue us from Ceteris Paribus Laws? Markus Schrenk

Many philosophers of science think that most laws of nature (even those of fundamental physics) are so called ceteris paribus laws, i.e. roughly speaking, laws with exceptions. Yet, the ceteris paribus clause of these laws is problematic. Amongst the more infamous difficulties is the danger that ‘For all x: Fx ⊃ Gx, ceteris paribus’ may state no more than a tautology: ‘For all x: Fx ⊃ Gx, unless not’. One of the major attempts to avoid this problem (and others concerning ceteris paribus laws) is to claim that the subject matter of laws are ascriptions of dispositions, powers, capacities etc., and not the regular behaviour we find in nature. That we do not know whether the cetera are paria in a specific situation does not matter to the dispositionalist because the objects have the disposition regardless of the circumstances. The defence of the latter claim is that dispositions can be instantiated without being manifested. Hence, the laws that ascribe dispositions are strict and it looks as if they do not face the above mentioned problems of ceteris paribus laws. In this chapter I attempt to show that these assumptions are wrong. I hope to illustrate that not only does the ceteris paribus clause reoccur inside the dispositions, moreover, there are laws – laws about non-fundamental entities with instable dispositions – which bear a ceteris paribus clause that cannot be hidden in a disposition. All Prevailing Ceteris Paribus Laws Many philosophers of science think that all, most, or many laws of nature are so called ceteris paribus laws. Take the following statements for examples: ‘All laws are ceteris paribus laws. (in Fn.) I even intend to include most so-called fundamental laws of physics.’1 ‘Whatever the law says must happen, hold or obtain, everything else being equal.’2 1 N. Cartwright, ‘Precis of Nature’s Capacities and Their Measurement’, Philosophy and Phenomenological Research, 55, 1 (1995): 155. 2 R. Harré, Laws of Nature (London, 1993), p. 79.

222

Dispositions and Causal Powers ‘The validity of most laws – with the possible exception of fundamental physical laws – is compatible with the existence of exceptional situations.’3

The notion ‘ceteris paribus’ – literally ‘all else being equal’– is usually read in a broad sense; namely, that a ceteris paribus law is a law which sometimes has exceptions. I adopt this broad reading. It should be mentioned, however, that it would be better to speak more generally of proviso laws if we have this broad sense in mind and of ceteris paribus only if we really mean that something, circumstances for example, have to be equal to a certain standard. In any case, and disregarding these verbal issues, my focus will be on the underlying unifying feature of the whole variety of proviso laws: in some cases the proviso law does not hold good,4 it faces prima facie a falsification while nonetheless being a law, i.e. the prima facie falsification is just an exception. The Problems with the Ceteris Paribus Clause Although many think that there are ceteris paribus laws – not only in the provisoridden social sciences but even all the way down to fundamental physics – the ceteris paribus clause in law statements is problematic. Three infamous difficulties are well known. 1. ‘For all x: Fx ⊃ Gx, ceteris paribus’, is in danger of being tautologous or incomplete: tautologous if we specify or define the ceteris paribus clause by saying circularly ‘For all x: Fx ⊃ Gx, except in those cases where Fs are not Gs’; incomplete if the ceteris paribus clause is thought to be cashed out by demanding (in the antecedent of the law) that possible interferences A, B, C, etc. do not occur. The problem with this variant is that we most certainly have to leave a gap in our statement ‘For all x: Fx and ? x ⊃ Gx’ because either we do not know all the interferers, or there are endlessly many interferers, or the class of exceptions is too heterogeneous. In the latter case it is not certain whether the exclusion clause can be formulated in the terms of the science to which the law belongs, that is, the interfering phenomena could possibly not belong to the scope of the science of the law under concern. In ‘Birds can fly, unless they are struck by lightning’, the weather conditions are not a biological phenomenon. Tautologous statements are empirically not very useful, since they are empirically vacuous. Incomplete statements, on the other hand, fail to express

3 M. Kistler, ‘Laws of Nature, Exceptions and Tropes’, Philosophia Scientiae, 7, (2003): 197-198. 4 Better, the core of the proviso law does not hold good; where I define: the core of ‘∀x (Fx ⊃ Gx), ceteris paribus’ is ∀x (Fx ⊃ Gx).

Can Capacities Rescue us from Ceteris Paribus Laws?

223

any well-determinate content and, hence, to pick out a law of nature. 2. Apart from these semantic problems of ceteris paribus law statements, proviso laws face other, epistemic, difficulties: proviso laws can neither be confirmed in any situation, since they do not hold good in any situation, nor can they be refuted easily. The ceteris paribus clause could be misused as an immunization strategy: we could claim that whenever the law does not hold the cetera weren’t paria. A bad result for the sciences if we are keen on demarcation: a non-falsifiable empirical science is in danger of resembling a pseudo-science like astrology. 3. Finally, ceteris paribus laws do not support counterfactuals, and are of little help in predictions, since we do not know whether in those counterfactual and/or future circumstances the cetera are paria. Whilst I am sure there are many more problems with provisos, 1 vacuity because of being either tautologous, or incomplete, 2 being uncomfirmable and not falsifiable, and 3 being unable to support counterfactuals or predictions are certainly amongst the most famous ones. Note finally that the difficulties I have mentioned are not tied to a specific account of lawhood. They arise for anti-Humean theories based on necessary connections between universals and for neo-Humean regularity accounts alike. As we shall see now it is the dispositionalist’s main claim that the latter theories especially, i.e. regularity theories, are threatened by those problems. The Dispositionalist’s Interest in Ceteris Paribus Laws The question of my paper is ‘Can Capacities Rescue us from Ceteris paribus Laws?’ It is noteworthy that this is usually not the question proponents of dispositions like Cartwright, Kistler, Hüttemann, Lipton, Bartels, etc., emphasize.5 That is, the dispositionalists’ motivation for claiming that the subject matter of all laws are dispositions is not primarily that dispositions solve all the problems the ceteris paribus clause poses.

5 See A. Bartels, ‘The Idea which we call Power. Naturgesetze und Dispositionen’, Philosophia Naturalis, 37 (2000): 255-268; Cartwright in several publications, e.g. N. Cartwright, Nature’s Capacities and their Measurement (Oxford, 1989); N. Cartwright, ‘Aristotelian Natures and the Modern Experimental Method’, in J. Earman (ed.), Inference, Explanation, and other Frustrations: Essays in the Philosophy of Science (Berkeley, 1992), pp. 44-71. N. Cartwright, The Dappled World: A Study of the Boundaries of Science (Cambridge, 1999), ch. 4; N. Cartwright, ‘In Favor of Laws that Are Not Ceteris paribus After All’, Erkenntnis, 57, 3 (2002): 425-439. A. Hüttemann, ‘Laws and Dispositions’, Philosophy of Science, 65 (1998): 121-135. P. Lipton, ‘All Else Being Equal’, Philosophy, 74 (1999): 155-168. M. Kistler, ‘Laws of Nature, Exceptions and Tropes’, Philosophia Scientiae, 7, 2 (2003): 189-219.

224

Dispositions and Causal Powers

A proper theory of ceteris paribus laws is actually only a marginal step in their enterprise. Their main goal is to show that the problems of proviso laws are only problems if we subscribe to a regularity theory of lawhood.6 Because of this, ceteris paribus laws are, for the dispositionalist, a welcome battleground against Humeans where best hope of winning seems to be with the one who has got the powers and capacities. The argument is roughly this: 1. There are ceteris paribus laws all the way down the sciences. 2. We cannot make sense of ceteris paribus laws if we adopt a regularity view of laws. 3. But they do make sense if we adopt dispositionalism. Hence, dispositionalism. The offended Humean anti-dispositionalist usually responds to this argument by attacking 1, i.e. by denying the existence of ceteris paribus laws. It is the Humean’s creed that the fundamental laws of physics are strict. Adding to this belief the reductionist claim that every science is reducible to physics they think they can avoid dispositionalism. I myself do not intend to defend the Humeans, that is, I do not intend to argue against dispositions, powers, capacities, natures etc. I try to remain neutral on this matter. I will nevertheless try to cast doubt on dispositionalism (which I understand from this point on as both the thesis that dispositions, capacities, powers, tendencies, etc., exist on top or instead of laws and that they can rescue us from ceteris paribus laws). My plan of attack focuses on 3, not 1. But before I present my central arguments let us (p. 226ff) see how the dispositionalist argues for 1, 2 and 3. The Dispositionalist’s Idea about Ceteris Paribus Laws Why do dispositionalists believe that they can avoid the problems ceteris paribus laws pose? As the quotations from above reveal, the dispositionalist’s presupposition is that all, most, or many laws are ceteris paribus laws. Take for example the prototype of a law, Newton’s law of gravitation. It says that masses m attract other masses M at distance r with the gravitational force FG=GmM/r². Let’s consider the special case of the earth and an arbitrary massive object near its surface, an overhead transparency for example. If I let it drop, will it fall according to the equation for the motion derived directly from the law of gravitation? It won’t. There can be all sorts of interferences: air resistance, the blowing of the overhead projector’s fan, electromagnetic forces due to electrostatic charge of the plastic, etc. So, even the prototype of lawhood – the law of gravitation – is a ceteris paribus law and we have

6 As I discovered later this has also been underlined by Earman, Roberts and Smith in their article ‘“Ceteris paribus” Lost’, Erkenntnis 57, 3 (2002).

Can Capacities Rescue us from Ceteris Paribus Laws?

225

a good reason for 1, i.e. that there are ceteris paribus laws all the way down the sciences. I quote Cartwright: The force of size GMm/r2 does not appear to be there; it is not what standard measurements generally reveal; and the effects we are entitled to expect – principally acceleration in a system of mass m a distance r away of size GM/r2 – are not there either.7

Real objects nearly never display the laws which are supposed to hold for them. If taken as statements about regularities laws are either all false or they are ceteris paribus laws because there are hardly ever regularities. But if the law of gravitation is not a regularity what is it then and what are all the other putative laws? The dispositionalists’ solution is to claim that the law of gravitation and others are not about objects’ behaviour but about their disposition to behave. The objects do in fact have these dispositions, but they might not be displayed or only partially displayed in real situations. The laws we use talk not about what bodies do, but about the powers they possess.8 The idea is that the properties linked by a law are dispositional rather than manifest.9 To say that laws describe how physical systems would behave in specified situations is to say that laws ascribe dispositions to physical systems.10 We don’t know when all things are equal, but the whole point of the dispositional view is in a sense that we do not need to know, since the disposition is present regardless.11

The basis of those thoughts is that dispositions can be present without their manifestation being realized. To put it another way, dispositions seem to be able to play the double role we need for ceteris paribus laws, namely permanence and absence at the very same time. Permanence because even if the circumstances are not fortunate and the lawlike behaviour is not displayed we think of laws as remaining somehow in good order. Absence because it seems that the law has – although still existent – little or no impact or manifestation if things are not equal. So far so good. Laws are about which kinds of objects have which dispositions and not about what objects do. That is great news because we can drop the ceteris paribus clause of the laws. Our newly interpreted laws are strict. Massive objects have the power to attract other masses, negative charges have the power to repel other negative charges, etc., regardless of the actual circumstances or actual behaviour. Therefore we get rid of all the problems ceteris paribus laws pose since our laws are strict now. Or are they? 7 N. Cartwright, ‘In Favor of Laws that Are Not Ceteris paribus After All’, Erkenntnis, 57, 3 (2002): 428. 8 N. Cartwright, How the Laws of Physics Lie (Oxford, 1983), p. 61. 9 M. Kistler, ‘Laws of Nature, Exceptions and Tropes’, Philosophia Scientiae, 7 (2003): 192. 10 A. Hüttemann, ‘Laws and Dispositions’, Philosophy of Science, 65 (1998): 128. 11 P. Lipton, ‘All Else Being Equal’, Philosophy, 74 (1999): 166.

226

Dispositions and Causal Powers

Does Dispositionalism Solve the Difficulties of Ceteris Paribus Laws? I have divided the dispositionalists’ overall argument into three parts: (1) There are ceteris paribus laws all the way down. (2) We cannot make sense of ceteris paribus laws if we adopt a regularity view of laws. (3) But they do make sense if we adopt dispositionalism. I have just presented the dispositionalists’ arguments for (3). To support the first claim (1) I have quoted the law of gravitation as an example. Claim (2) is made plausible by the problems I have listed earlier for ceteris paribus laws and by mentioning that those problems affect every other theory of lawhood but dispositionalism and, hence, a fortiori regularity theories. So, can we conclude that dispositionalism is the way to solve, or better, to avoid the problems of ceteris paribus laws? I think that even if we grant that dispositionalism (as opposed to Humeanism) is the correct metaphysical stance on laws of nature it is still an illusion that we have escaped the ceteris paribus problems. Taking quotes from the works of the most prominent dispositionalist, namely, Nancy Cartwright, I will present two major difficulties:12 First: dispositionalists have so far only gestured at, but never spelled out in detail, what their new law statements – law statements involving dispositional concepts – look like. It will prove more difficult than it might seem to come up with a precise version of, say, Coulomb’s law which will satisfy the dispositionalist. Moreover, while arriving at a dispositionalist formulation it becomes clear that the ceteris paribus clause sneaks back in again (see the next section). Second: the difference between an object whose disposition’s manifestation is masked or counteracted against and one whose disposition is lost (because the basis is lost) has not been properly distinguished. The latter case leads to a sort of ceteris paribus clause which cannot be accommodated by dispositionalism easily (see p. 243ff). How the Ceteris Paribus Clause Re-Enters the Stage First and foremost, note that the alleged capacity concepts of scientific theories have to have an internal mathematical functional structure just like laws do.13 Charges q do not simply have the capacity to attract (repel) other charges Q but, roughly, to attract (repel) other charges Q in distance r with force FC= qQ/(4πε0r2). Dispositionalists agree that otherwise dispositions cannot take over the former laws’ job. Hence, Cartwright writes: 12 I will take Coulomb’s law as an example for my enquiry. Any of my arguments should be applicable to at least all the other force laws. 13 If this does not hold for all of them my critique might only count for those for which it does (although I believe that my critique could then be extended with only a few alterations).

Can Capacities Rescue us from Ceteris Paribus Laws?

227

Coulomb’s capacity differs from everyday ones [in] that it has an exact functional form and precise strength, which are recorded in its own special law. 14

With this important insight I will try to formulate the new kind of law in a trial and error manner. Eventually only the last, the third, attempt will prove to be a good candidate. I start with my first suggestion: charges, so we are told, have Coulomb’s capacity C to attract (repel) other charges: ‘Coulomb’s law tells not what force charged particles experience but rather what it is in their nature, qua charged, to experience.’15 Hence, the following formulation should express the dispositionalist’s law, or so it seems: 1. It is a strict law that for any two objects x and y if they have charge q and Q respectively and are in distance r to each other then x has the capacity C to exert the force FC= qQ/(4πε0r2) on y (and v.v. for y, i.e. y has the capacity to exert the force FC= qQ/(4πε0r2) on x). This, however, cannot be right. In reading 1 the capacity C is the capacity to exert force FC, i.e. an exerted force FC would be the capacity’s manifestation. Yet, the immediate question is: when is this capacity manifested, i.e. what is the trigger for that capacity and/or under which circumstances does it manifest itself? C is the capacity to exert force FC if what?16 That this question needs an answer stems from old empiricist concerns. For the capacity to be a respectable entity we must be able to test it. In fact, the dispositionalists insist themselves that their capacities are empirically testable, even measurable. ‘Causal capacities can be measured as surely – or unsurely – as anything else that science deals with. Sometimes we measure capacities in a physics laboratory.’17 Yet, it is the demand for the possibility of a test or measurement operation which forces us to spell out a condition under which the respective capacity will manifest itself. For how are we to measure the capacity if we do not know how to trigger its manifestation? More forcefully, if there are capacities which are just the capacities to do xyz without there being a specified trigger condition then what is to prevent us from supposing that everything has those trigger-less capacities (no test could ultimately prove that they do not). So what is the trigger in our concrete case? If we answer that it is the location of the two charges in distance r we get an undesirable result because the trigger

14 A puzzle aside the main line of argument is, by the way, the ontological issue whether there is then just one capacity or one for each charges-distance-force triple. N. Cartwright, The Dappled World. A Study of the Boundaries of Science, p. 54. 15 N. Cartwright, ‘Aristotelian Natures and the Modern Experimental Method’, p. 48. 16 Compare: fragility – the capacity to break easily if struck; solubility – the capacity to solve if immersed in water. 17 N. Cartwright, Nature’s Capacities and their Measurement, p. 7.

228

Dispositions and Causal Powers

condition would appear twice: once as antecedent of the law and once as antecedent of the capacity. 1*. It is a strict law that for any two objects x and y if they have charge q and Q respectively and are in distance r to each other then x has the following capacity: to exert the force FC= qQ/(4πε0r2) on y if it has charge q and is in distance r to y with charge Q (and v.v. for y). This redundant claim cannot be the new law formulation the dispositionalist is after. Apart from the redundancy which is already a reason not to accept the statement – it would also make the disposition’s manifestation a sure thing because the disposition’s antecedent is met whenever the law’s antecedent is fulfilled. In other words, the disposition manifests whenever it is instantiated and this means we can get rid of the disposition entirely and just write: 1**. It is a strict law that for any two objects x and y if they have charge q and Q respectively and are in distance r to each other then x exerts the force FC= qQ/(4πε0r2) on y (and v.v. for y). That is, we would be back to the law with which we started. The obvious step to take is to delete some trigger conditions from the antecedent of the law. Hence my second suggestion: 2. It is a strict law that for any two objects x and y if they have the charges q and Q respectively then x has the following capacity: to exert the force FC= qQ/(4πε0r2) on object y if x is in distance r to y (and v.v. for y). I have reduced the antecedent condition of the law to ‘if charges’ (as opposed to the first suggestion: if charges in distance r then capacity), i.e. now having charge is sufficient to have the capacity.18 Suggestion 2 has, however, still a consequence which the dispositionalist cannot welcome. In order to see that, let me repeat the main advantage dispositionalists find in capacities: capacities fulfil a double role, that is, they can be instantiated without being manifested. As we will see now this claim needs qualification, for what cannot be meant by this claim is the triviality that although the capacity can be instantiated its antecedent condition might not be met so that it does not manifest. Compare solubility: sugar is soluble, but sugar does not solve if it is not put in water. Translated into Coulomb’s capacity a single charge has Coulomb’s capacity but trivially it does not manifest if there is no second charge around at distance r.

18 Actually, one has to give some thoughts to whether we should formulate the dispositionalists’ claim for a single object x or as done here for two objects x and y. However, a reformulation just for x would not affect my arguments.

Can Capacities Rescue us from Ceteris Paribus Laws?

229

The crucial feature the dispositionalists are after has to be something deeper. They have to stipulate that a capacity can fail to manifest even though the trigger is pulled, i.e. even though the antecedent of the disposition is met. Here, that a charge q fails to exert the force FC even though another charge Q is present in distance r.19 Indeed, this is what Cartwright says: Consider Coulomb’s law of electrostatic attraction and repulsion. Coulomb’s law says that the force between two objects of charge q1 and q2 is equal to q1q2/r². Yet, this is not the force the bodies experience; they are also subject to the law of gravity. […] Coulomb’s is not the force that actually occurs; rather, it is a hypothetical power hidden away in the actual force.20

However, this demand – that the capacity fails to manifest although its antecedent is met – is not fulfilled by our definition 2 which, so far, stipulates that if the trigger is pulled the capacity must manifest. Yet, the changes we have to make to 2 in order to meet the new requirement are devastating: the capacity has, internally, to bear a proviso: an object x with charge q exerts the force FC= qQ/(4πε0r2) on object y if x is in distance r to y and y has charge Q, ceteris paribus(!) – where ceteris paribus means (amongst other things) that no masses should be around. The amended version of 2 then reads: 2*. It is a strict law that for any two objects x and y if they have the charges q and Q respectively then x has the capacity C which is the capacity to exert the force FC= qQ/(4πε0r2) on object y if x is in distance r to y, ceteris paribus (and v.v. for y). Note that the whole phrase after ‘which is the capacity…’ is the characterization of the capacity C, i.e. the ceteris paribus clause is contained within that characterization and does not contradict the strictness of the overall law which states that charges always have that capacity. In any case, we are back to the ceteris paribus clause and all the problems of ceteris paribus laws as mentioned in §2 are simply shifted one level down. The capacity the dispositionalists have invented turns out to be a veil covering the ceteris paribus clause. But maybe I have completely misunderstood dispositionalism and the whole reconstruction 2* is wrong. Another possible interpretation should be discussed: Proto 3. It is a strict law that for any two objects x and y if they have charge q and Q respectively then x has the Coulomb force capacity FC (and v.v. for y), i.e. the

19 A point which I cannot discuss here is that dispositionalists have to go even one step further: they have to stipulate that once the antecedent conditions are met the disposition is trying, urging, pressing, pushing to manifest but does not succeed. 20 N. Cartwright, ‘Aristotelian Natures and the Modern Experimental Method’, p. 48.

230

Dispositions and Causal Powers

capacity is identified with the force; the force itself is the capacity sought after. The force (seen as a capacity) is there in any case: ‘if charge then force’ is a strict law. It is, however, the force’s manifestation (that is its effect) which might not occur. So far this sounds like a promising claim. Yet, what is a force capacity’s trigger and what is its manifestation? The force capacity is the capacity to do what and when? I take it that it is rather uncontroversial to identify the manifestation with an acceleration – here the acceleration a=qQ/(4πε0r2)/m (with m being the mass of x; mutatis mutandis for y).21 But what are the antecedent conditions for this force disposition? For a start, I take just the presence of the two charges in distance r as the trigger. Hence, the Coulomb force capacity has the following overall structure: two objects x and y each have the force capacity FC ↔ if x is in distance r to y then x is accelerated towards y with a=qQ/(4πε0r2)/m (m being the mass of x)22 (mutatis mutandis for y).23 So the overall law statement of the dispositionalist reads as follows: 3. It is a strict law that for any two objects x and y if they have charge q and Q respectively then x has the Coulomb force capacity FC which is the following capacity: to accelerate towards y with a=qQ/(4πε0r2)/m if x is in distance r to y (and mutatis mutandis for y). Unfortunately, this definition is at odds in two respects with Cartwright’s quote from above. Reconsider: ‘Yet, this is not the force the bodies experience; they are also subject to the law of gravity.’24 So, firstly, where Cartwright says that Coulomb’s force is not there in certain cases 3 claims that Coulomb’s force is always there. What the bodies might indeed not experience (according to 3) is the Coulomb force capacity’s manifestation, i.e. a certain acceleration. If, then, we take 3 for granted we have to change Cartwright’s claim to: ‘Yet, this is not the acceleration(!) the bodies experience; they are also subject to the law of gravity’. The second respect in which 3 does not match up with Cartwright’s quote is the same for which 2 was imperfect and had to be changed to 2*: we have not taken care of the possibility that the manifestation might well not occur although the antecedent condition of the force capacity is met. This is, for example, the case when, to take Cartwright’s example, also masses, and not only charges, are around. Taking care of this fact we arrive, similar to 2*, at 3*: 21 However, this claim will be challenged below. 22 Note that x’s and y’s being charged does not appear anymore in this force capacity’s antecedent. This is acceptable because objects x and y have this force capacity if they are charged. It’s the law (if charge then force-capacity) which takes care of the fact that the particles are charged. (One might even go one step further and claim that the law is a biconditional: charge if and only if Coulomb-force-capacity. But I do not have the space to discuss this, maybe essentialist, possibility here.) 23 y is accelerated towards y with a=qQ/(4πε0r2)/M; M being the mass of y. 24 N. Cartwright, ‘Aristotelian Natures and the Modern Experimental Method’, p. 48.

Can Capacities Rescue us from Ceteris Paribus Laws?

231

3*. It is a strict law that for any two objects x and y if they have charge q and Q respectively then x has the Coulomb force capacity FC which is the following capacity: to accelerate towards y with a=qQ/(4πε0r2)/m if x is in distance r to y, ceteris paribus (and mutatis mutandis for y).25 A brief excursus: in principle, there is another logical possibility to accommodate the needed ceteris paribus caveat in version 3. We could claim that the antecedent of the law (not the capacity) has to bear the ceteris paribus clause which would amount to the claim that not in any circumstances would a charged object x have Coulomb’s force capacity FC: it is a law (not a strict law!) that for any two objects x and y if they have charge q and Q respectively then x has, ceteris paribus, Coulomb’s force capacity FC. This, however, would bring us straight back to the original ceteris paribus law itself (or a very close relative) and it was the dispositionalists’ aim to avoid that. In the light of the last remark it could seem to be a matter of choice where the ceteris paribus proviso appears: if you prefer it inside a capacity choose dispositionalism, if you prefer it at the law-level choose orthodox theories of lawhood. That things turn out to be even more complex will be discussed in §8. For now let’s suppose we have discarded the version with the ceteris paribus clause at the law level. Back to version 3*, I have to ask whether it is legitimate to exchange forces for accelerations in Cartwright’s statement (as done above) and, more importantly, how authentic it is to introduce the ceteris paribus clause into the capacity. A later claim of Cartwright’s can serve as justification for the latter: ‘It is in the nature of a force to produce an acceleration of the requisite size. That means that ceteris paribus, it will produce that acceleration.’26 This fairly recent quote seems to be in full accordance with my analysis 3*. Anyway, the attempt to find an alternative interpretation to 2* in order to avoid a ceteris paribus clause, has been frustrated. Whichever way turns out to be the right one, 2* or 3*, to spell out dispositionalism in detail – it unavoidably contains a ceteris paribus clause. Hence, my primary goal here (§6), to show that it is doubtful whether dispositionalism can rescue us from ceteris paribus clauses in laws has been achieved.27

25 And again, the ceteris paribus clause is inside the force capacity. That a charge has that capacity remains a strict law. 26 N. Cartwright, The Dappled World. A Study of the Boundaries of Science, p. 28, my emphasis. 27 With a different argument Peter Lipton comes to a similar conclusion: ‘[The] problem for the occurrent view of cp laws [i.e. the regularity view dispositionalists fight against; MAS] is that we cannot give cp laws determinate content because we cannot specify the antecedent of the law; the problem for the dispositional view is that we cannot give the dispositional attribution determinate content because we cannot specify the antecedent of the corresponding conditional that would give the dispositional term its meaning. So we have made no advance.’, P. Lipton, ‘All else being equal’, p. 167.

232

Dispositions and Causal Powers

However, I must spend some more time on deciding whether (2*) or (3*) is the correct reading of dispositionalism. Recall that (2*) defines a capacity C which is the capacity to exert a force when circumstances are right, where (3*) says that the capacity is the force FC which, when the circumstances are right, will have an acceleration as its effect. Both C and FC are supposed to be linked strictly to charges, i.e. it is either a strict law that charges have C or it is a strict law that charges have FC depending on which version, (2*) or (3*), turns out to be the right interpretation of dispositionalism. A lot counts for (2*) being the right interpretation. (2*) is suggested by Cartwright’s statement from 1992: Yet, this is not the force the bodies experience; they are also subject to the law of gravity. […] Coulomb’s is not the force that actually occurs; […] Coulomb’s law tells not what force charged particles experience but rather what it is in their nature, qua charged, to experience.28

And also by a statement from 1999: To say that it is in their nature to experience a force of q1q2/4πε0r² is to say at least that they would experience this force if only the right conditions occur for the power to exercise itself ‘on its own’.29

Also, other authors seem to have interpreted dispositionalism in the way of (2*). Summarizing the essence of Cartwright’s theory, Earman and Roberts say: They [laws] are attributions of capacities and tendencies to various kinds of systems; in a nutshell, ‘cp: (x)(Fx→Gx)’ is true just in case all Fs have a capacity or tendency to be G, so that they will be G in (the rare) cases where there are no other capacities or tendencies acting on them.30

Cartwright insists that […] the law attributes to charged bodies the capacity to exert a force on other charged bodies. The exertion of forces, it seems, counts as behaviour, whereas the capacity to exert a force does not – so the latter but not the former is the sort of thing we should expect there to be lawlike regularities about.31 Although, a lot counts apparently for (2*), I will now argue that not (2*) but (3*) should be the correct reading of dispositionalism. For that purpose let me consider a practical example Cartwright originally designed as an attempt to show that virtually

28 N. Cartwright, ‘Aristotelian Natures and the Modern Experimental Method’, p. 48. 29 I take it that C, the capacity from (2*), can be identified with ‘their nature to experience a force of q1q2/4πε0r²’ which is the nature to ‘experience this force if only the right conditions occur’. N. Cartwright, The Dappled World. A Study of the Boundaries of Science, p. 82. 30 J. Earman and J. Roberts, ‘Ceteris paribus, there is no Problem of Provisos’ Synthese, 118 (1999): 455. 31 J. Earman, J. Roberts and S. Smith, ‘“Ceteris paribus” Lost’, p. 288.

Can Capacities Rescue us from Ceteris Paribus Laws?

233

anything can happen due to a capacity of an object and that, hence, the endeavour to press capacities into the definitorial straightjacket of a single trigger condition and a single manifestation linked by a conditional is not tenable. I will return to this specific threat to interpretations like (2*) and (3*) later. Presently, I want to establish that of (2*) and (3*), (3*) is the more accurate reading. Cartwright states that even the opposite of what we intuitively think should happen when a capacity is active can occur: There is no one fact of the matter about what occurs when charges interact. With the right kind of structure we can get virtually any motion at all. We can even create environments in which the Coulomb repulsion between two negatively charged particles cause them to move closer together.32

Then follows a description of an experimental setup in which two electrons at distance r1 repel each other such that the second one of them moves into a magnetic field (the first is supposed to be held stationary). The initial force between the two electrons is stated as FC = (e1e2) / (4πε0r1²).33 This force causes the second electron to enter the magnetic field with the velocity v1. The field is supposed to be set up in strength and direction such that the second electron moves in a circular trajectory back to the first electron. There is an insulation chamber at the side from which it approaches the first electron again. Hence, from this side, the repulsion between the electrons is swallowed by the insulator. All the parameters of the experiment (distances, strength of the magnetic field, etc.) are chosen such that the moving electron comes to a halt at distance r2 to the first electron which is smaller than the initial distance r1. Now, with which interpretation, (2*) or (3*), can this experiment be best reconciled?34 The initial state – two electrons each with unit charge e- in distance r1 – would, according to (2*), lead to them both having capacity C which is the capacity to exert a force on each other when circumstances are right. As the experiment has it, circumstances are right such that the full force causes the second electron to accelerate into the magnetic field. What does (3*) say? Contra (2*) it says that the force is there in any case – whatever the circumstances. Whether the electron does accelerate into the magnetic field or not is, however, dependent on the circumstances because a force is supposed to be the capacity to accelerate only if the right circumstances occur. But again, the

32 N. Cartwright, The Dappled World. A Study of the Boundaries of Science, p. 60. 33 ‘According to the laws of electromagnetism, the force between the two electrons is a repulsive force equal to F = (e1e2) / (4πε0r1²).’ (N. Cartwright, The Dappled World. A Study of the Boundaries of Science.) 34 One might challenge my line of argument here by saying that focusing on a single experiment cannot be decisive as to whether (2*) or (3*) is the correct interpretation. It is quite clear, however, that this experiment is a paradigmatic case for force calculations and additions. Whenever forces are involved we do what we do in this experiment. Hence, this example can well stand for every other experiment.

234

Dispositions and Causal Powers

experiment is engineered such that the situation is fortunate and the electron does indeed enter the field with velocity v1 due to the acceleration. It seems that in order to distinguish a difference between 2* and 3* we have to make the circumstances less fortunate – but nothing is easier than that. We just have to have a closer look at the scenario to realize that electrons also have mass, that is, we just have to take into consideration that there will also be a gravitational force FG = - Gmeme/r² between the two electrons in order to introduce a disturbing factor that makes circumstances less favourable. Now, the story told by 2* is that the electrons still have their capacity C. After all, according to 2*, the connection between charges and capacity C is supported by a strict law. And yet, C does not (or does so only partially) manifest, i.e. the force F = (e1e2) / (4πε0r1²) between the two electrons is not (fully) there. The analogous story in terms of 2* can be told about Newton’s law of gravitation. The gravitational capacity G is there, however, its manifestation, a gravitational force is not (or is so only partially). Yet, which force is there? That there must be one is obvious because the second electron is still accelerating into the field.35 Well, the answer can be given by school physics: the sum of both forces is responsible for the acceleration. However, turning from schoolbook physics to metaphysics we might ask how on earth it is justified to add forces in order to arrive at the resultant force if those component forces are not there. Coulomb’s force (so interpretation 2*) is not there, because masses are around and Newton’s force is not there because charges are also around. Why should the addition of both forces be the resultant force if none of the components is actually there? The best we can say about 2* is that it leads to a very artificial way of speaking: ‘Coulomb’s […] force […] is a hypothetical power hidden away in the actual force.’36 The worst we can say is that it is misleading because it could make one believe that it is not Coulomb’s force FC = (e1e2) / (4πε0r1²) which is there when masses are also around but that it is a different force (say, F = (e1e2)/8πr7). What has 3* to say about the whole affair? According to 3* both Newton’s and Coulomb’s force are there. Yet, they are both not (or not fully) manifested because circumstances are not favourable: the two forces are, so to speak, both in each other’s way.37 To see what the resulting force is we add the two while not changing the calculation for each of them. And why should we, for they, the forces, are wholly present. Earman et al. who, in a quote I have presented above seemed to interpret Cartwright in terms of 2*, give my reading 3* even more credibility when they say in a different passage that successful physical theories apparently quantify over component forces, and there seems to be no natural way of ‘paraphrasing away’ reference to such forces (as there is for, e.g. references to absolute motion in Newtonian mechanics).

35 I suppose that FG and ⎜z-down> in that direction (i.e. suppose that before the measurement the particle has spin in the x-direction and is therefore in a superposition of the two previous states). If we invert the polarity of the measuring apparatus (from P1 to P2) and leave everything else unchanged, we change the measurement result on spin in the z-direction: if we had obtained z-up in the previous measurement, we obtain z-down in the second and vice versa. Given this result, the property of ‘having a definite spin in the z-direction’ is not categorical, intrinsic or faithfully measured: measurements do not in general reveal the preexisting possessed value of spin along the z-direction, for the simple reason that there isn’t any! If by switching the magnet and leaving everything else unchanged we change the experimental outcome from z-up to z-down, there is no definite property of having a spin in the given direction at all. In the terminology introduced before, the property ‘spin along a given direction’ is therefore contextual2, for the simple reason that the experimental outcome causally depends on the measurement context together, of course, with the ‘hidden’, non-contextual2 value of the position of the particle. 24 For a readable account of Bohmian mechanics, see S. Goldstein, ‘Bohmian Mechanics’, Stanford Encyclopedia of Philosophy, http://plato.stanford.edu/entries/qm-bohm (2001). 25 Furthermore, the assumption that the observable A has a definite, possessed value that – in being independent of the measurement context – is non-contextual2, would lead to a contradiction with a well-known ‘no-go theorem’, prohibiting a non-contextual2, simultaneous assignment of a definite value to all the observables of a system whose dimension in Hilbert space in greater than 2. S. Kochen and E. Specker, ‘The Problem of Hidden Variables in Quantum Mechanics’, Journal of Mathematics and Mechanics, 18 (1967): 1015-1021.

Dispositions, Relational Properties and the Quantum World

261

Exactly for this reason, however, in the Bohmian interpretation the dispositional properties of ‘spins along a given direction’– referred to particles that before measurement are in a state of superposition with respect to that direction – are reducible to position and the context of measurement.26 The idea of reducibility is based upon the fact that our hypothetical knowledge of the position of the particle plus knowledge of the orientation of the magnet would allow us to deduce the result of the outcome with certainty. However, for the reasons illustrated above, there is nothing pertaining to ‘spin in the z-direction’ that is categorically, intrinsically and definitely possessed by the particle before measurement, precisely in virtue of the above stipulation that a property is categorical if and only if it has a definite value before measurement and is therefore contextual1. Except for position then, both in Bohm’s and in Heisenberg-Bohr’s interpretations, all properties of micro-systems are contextual2 and therefore dispositional, but in the Copenhagen interpretation, unlike in the Bohmian case, all such dispositions are of an irreducible type. As Clifton and Pagonis correctly remarked, if we assumed that spin-directions and other properties were irreducibly dispositional, we would lose the benefits of the completeness of Bohm’s theory!27 Dispositions in Non-Collapse Views: the Relational Interpretation of QM (Rovelli) Rather than assigning an ontological meaning to the wave function, the relational interpretation of QM focuses on the sequence of actual measurement outcomes q1, q2, qn,… But such outcomes are to be regarded as the result of correlations of quantum systems with particular ‘observing physical systems’ S, and no absolute meaning is attached to the intrinsic properties of an isolated quantum system Q. A quantum system Q can be said to possess a certain property q only relative to a system S; relative to another observing system S’, Q and S may be in an indefinite state, i.e. in a superposition. Relational quantum mechanics28 is therefore a way of reconciling the universality of application of the principle of quantum superposition with the fact that the observed world is characterized by uniquely determined events. In Rovelli’s and Laudisa’s words: ‘there is no meaning in saying that a certain quantum event has happened or that a variable of the system S has taken the value q: rather, there is meaning in saying that the event q has happened or the variable has taken the value q for O, or with respect to O… If I observe an electron at a certain position, I cannot conclude that the electron is there: I can only conclude that the electron as seen by me is there. Quantum events only happen in interactions between systems, and the fact that a 26 R. Clifton, and C. Pagonis, ‘Unremarkable Contextualism: Dispositions in the Bohm theory’, pp. 285-286. 27 Ibid., p. 288. 28 C. Rovelli, ‘Relational quantum mechanics’, International Journal of Theoretical Physics, 35 (1996): 1637-1678.

262

Dispositions and Causal Powers

quantum event has happened is only true with respect to the systems involved in the interaction. The unique account of the state of the world of the classical theory is thus fractured into a multiplicity of accounts, one for each possible “observing” physical system.’29 This interpretation combines features of the Copenhagen view of QM (limitation of what we can say about unmeasured systems) with Everett’s view to be presented later. Since in this interpretation we cannot talk about intrinsic properties of physical systems, for our purpose it is important to stress that according to Carlo Rovelli’s interpretation the quantum world possesses no categorical properties at all. All properties are clearly contextual2 and therefore, in our language, irreducibly dispositional; in this interpretation, the notion of ‘correlation’ is so central that all properties of the quantum universe are not only relational but also dispositional. While we have seen that in general a relational property need not be dispositional, in the philosophy of QM the fact that a property is categorical (contextual1) implies its being intrinsically possessed; therefore the converse implication also holds: relationality implies dispositionality. In Rovelli’s interpretation, it is meaningless to even think of categorical properties that ground the disposition to show certain values of position, spin etc.: while correlations may be ascribed physical reality, the quantities that are the terms of the correlations cannot.30 Considering a view of the world in which only (cor)relations have a determinate reality and their relata don’t, couldn’t we take the correlated systems S+O to be the basic, categorical ontological ingredient and thereby avoid the dispositionality of the quantum system S? This move is unfortunately going to be unsuccessful, since relative to another observer O’ the joint system S+O will appear in a superposed state, and will therefore fail to have a determinate value until O’ performs a measurement by correlating with it. ‘As far as the second observer O’ is concerned, the only relevant element of reality is that a correlation is established between S and O. This correlation will manifest itself only in any further observation that O’ would perform on the S+O system. Up to the time in which it physically interacts with S+O, the system O’ has no access to the actual outcomes of the measurements performed by O on S. This actual outcome is real only with respect to O.31 Before the moment in which O’ correlates with the state S + O, the state will be mathematically well-defined and definite, but the corresponding, joint physical system, being in a superposition, won’t have a definite property at all.32 In conclusion, in Rovelli’s interpretation the strategy of considering the joint system S+O to be a categorical basis fails.

29 C. Rovelli and F. Laudisa, ‘Relational Quantum Mechanics’, Stanford Encyclopedia of Philosophy, http://plato.stanford.edu/entries/qm-relational/ (2002), sect. 2. 30 N.D. Mermin, ‘What is quantum mechanics trying to tell us?’, American Journal of Physics, 66 (1998): 753-767. 31 C. Rovelli, ‘Relational quantum mechanics’, pp. 1650-1652. 32 This remark answers an objection raised by Nancy Cartwright during the discussion.

Dispositions, Relational Properties and the Quantum World

263

Dispositions in Many Worlds/Many Minds Interpretation The main ontological difference between Rovelli’s relational view and Everett’s view in its various versions (many worlds, many minds, etc.) derives from the fact that while Everett takes the ψ-function as the ontological basis of QM, Rovelli regards as (relationally) real only the outcomes of two-systems correlations. As Carlo Rovelli and Federico Laudisa stress, it is one thing to say that a definite value or property or event is relative to a system, as in Rovelli’s interpretation; quite another to say that definiteness is relative to the state of a system, as in Everett’s: ‘According to the relational interpretation, after the first measurement the quantity q has a given value and only one for O, while in Everett’s terms the quantity q has a value for one state of O and a different value for another state of O, and the two are equally real’.33 In this quotation, Rovelli and Laudisa are clearly referring to the many-worlds version of Hugh Everett’s interpretation,34 according to which there is an ontological multiplicity of realities, one for each component of the superposed state representing the total physical system (see infra). In the many-minds version, on the contrary, there is a plurality of mental ‘viewpoints’ or ‘perspectives’ on the same indefinite world, which is in a gigantic state of superposition (Albert and Loewer, and Squires).35 Note that such a plurality of perspectives is absent from the relational point of view only in appearance, given that also in Rovelli’s interpretation definite physical quantities can actualize only relationally, that is, only once two systems are given. It then follows that for any quantum system S, there are many possible ‘perspectives’ that can be associated with it, one for each possible correlation: according to the ‘perspective’ of observer O, event q has happened, but according to another perspective associated with O’, q has not happened, and both versions must be regarded as legitimate. The analogy that Rovelli’s interpretation suggests between quantum perspectives on superposed states on the one hand, and special relativistic, 3+1 ‘perspectives’ on the four dimensional reality (Minkowski spacetime) on the other, seems to indicate that – despite the intention of the author – superposition states might be regarded as a fundamental ontological ingredient, and two-systems, quantum correlations mere perspectives on an indefinite reality. In order to probe the consequences of the many-worlds interpretation for the dispositional or categorical nature of its fundamental properties, let us write the quantum state of the universe as a superposition of various worlds:36 33 C. Rovelli and F. Laudisa, ‘Relational Quantum Mechanics’, sect. 5. 34 H. Everett, ‘Relative State Formulation of Quantum Mechanics’, Review of Modern Physics, 29 (1957): 454-462. 35 This second reading, involving ‘many perspectives’ seems to be closer to the original view proposed by Everett (1957). D. Albert and B. Loewer, ‘Interpreting the Many-Worlds Interpretation’, Synthese, 86 (1988): 87-98. E. Squires, ‘Many Views of One World’, European Journal of Physics, 8 (1987): 171-173. 36 Here I follow Lev Vaidman’s notation (2002). ‘The Many-Worlds Interpretation of Quantum Mechanics’, Stanford Encyclopedia of Philosophy (2002) http://plato.stanford.edu/

264

Dispositions and Causal Powers

|ΨUniverse > = ∑i ci | Ψ World-i >, ∑i |ci |2 = 1 Notice first of all that each world behaves like a quasi-classical world of definite properties, in which ‘fuzzy’ properties (i.e. superpositions) are absent. Consequently, there is a sense in which quantum dispositions are ‘locally’ reducible, or better, eliminable in each semi-classical world composing the superposition. However, since the whole universe is in a global state of superposition, we could as well conclude that, relative to the quantum state of the universe, which is a perfectly definite quantum state from a mathematical viewpoint, the physical properties of the universe are really indefinite. Relative to the quantum state of the universe, we cannot claim that the moon has a definite position, given that the definiteness of such a property is only a perspectival matter, depending on, and varying with, different worlds, branches, or minds.37 Consequently, in the many-worlds view there seems to be also a sense in which the properties of the quantum universe (say, having a certain density at a certain time) can be understood as being irreducibly dispositional, i.e. as being capable of having different definite ‘manifestations’ in different worlds or branches or minds, all of them being real. That such a dispositional reading of Everett-type interpretations is not so implausible can be gathered also by the notion of a centred-world put forward by Vaidman,38 and by him attributed to Saunders:39 if the world is centred on a human being, only perceived states are definite, and non-perceived ones are really superposed. Reality in itself is an entangled mess, and has the ungrounded, irreducible disposition to correlate to our brain states in such a way that we perceive the world as having definite properties. Never an interpretation of a physical theory has put more emphasis on the radical gap between the way the world is (reality) and how it illusorily appears to us.40 Consequently, the question whether in the many-worlds view quantum dispositional properties are irreducible either admits no definite answer or it receives a positive one. Within the former alternative, by focusing alternatively on the quantum state of the universe or on the single worlds composing it, we oscillate from the irreducibility to the reducibility of quantum superpositions. In another, more plausible reading, however, corresponding to the latter alternative, what is really real or mind-independent is the superposed state of the quantum universe, and the

entries/qm-manyworlds/. 37 Replace ‘worlds’ with ‘minds in a single individual’ and you get the many-minds view. 38 L. Vaidman, ‘The Many-Worlds Interpretation of Quantum Mechanics’, sect. 2. 39 S. Saunders, ‘Time, Quantum Mechanics and Decoherence’, Synthese, 102 (1995): 235-266. 40 For the problem of the empirical coherence of Everett’s type views of QM, see J. Barrett, The Quantum Mechanics of Minds and Worlds (Oxford, 1999).

Dispositions, Relational Properties and the Quantum World

265

definiteness of the classical world is somehow less fundamental if not illusory, given that the classical world in which superpositions are eliminated is a world centred on the brains of human beings. Dispositions in Spontaneous Collapse Theories (GRW) In the so-called GRW theories (from the acronym of their main inventors, Gian Carlo Ghirardi, Alberto Rimini and Tullio Weber),41 the definite, macroscopic world of our experience, threatened to be in a nebulous state by the universal validity of the principle of superposition, is obtained via a modification of the linearity of Schrödinger’s equation.42 In GRW’s original model (1985), on which we will focus, the wave function of a system is multiplied by a localization function, which physically represents a spontaneous localization in a ‘limited’ region of space of a previously non-localized quantum system. Apart from technicalities, that in this context have no importance, it is essential to note that according to GRW the fundamentally stochastic nature of the localization mechanism is not grounded in any categorical property of the quantum system: the theory at present stage is purely ‘phenomenological’, in the sense that no ‘deeper mechanism’ is provided to account for the causes of the localization. ‘Spontaneous’, as referred to the localization process, therefore simply means ‘uncaused’. In the attempt to unify the dynamics of microscopic and macroscopic systems, GRW suppose in other words that all quantum systems have an irreducibly probabilistic disposition (a propensity) to localize in a region of space whose dimension is approximately 10-5 cm, with a frequency f given by 10-16 sec-1. The probability that such a process occurs is defined as f times a second: this is tantamount to assume that a microscopic system (say, a proton) undergoes a localization process, on average, once every 1016 seconds (approximately corresponding to once every hundred million years) and this hypothesis explains why isolated quantum systems can typically remain for a very long time in non-localized or superposed state (i.e. they are spread across a large region of space). However, since a macroscopic system is constituted in average by 1023 atomic components, and since in GRW’s model the localization of a single particle drives the collapse of all the others, it follows that the components of a macroscopic apparatus that are correlated with the particle that we want to measure will undergo a localization every 10-7 seconds. In fact, the average number of particles that will collapse spontaneously in a second is given by 10-16 x 1023 = 107, which means the macroscopic apparatus remains in a state of indefinite position (i.e. in a superposition

41 G.C. Ghirardi, A. Rimini and T. Weber, ‘Unified Dynamics for Microscopic and Macroscopic Systems’, Physical Review, 34 (1986): 470-491. 42 A recent, readable survey of collapse theories and relative references is given in G.C. Ghirardi, ‘Collapse Theories’, Stanford Encyclopedia of Philosophy, (2002), http://plato. stanford.edu/entries/qm-collapse.

266

Dispositions and Causal Powers

of two position states) for no more that 10-7 seconds: as Bell43 put it, Schrödinger’s cat remains neither dead nor alive for no more than a split second. For our purposes, it is essential to stress once again that the new uncaused tendency to ‘swerve’ attributed to the atoms by GRW (in Lucretius’ De Rerum Natura we read of a ‘clinamen’ accidentally deviating the vertical fall of the Democritean atoms)44 is an irreducibly dispositional property, that becomes actual or is manifested in ways that could call for – were the new theory prove to be successful in overcoming its present difficulties with a relativistic extension and were it experimentally confirmed – the introduction of new constants of nature. Furthermore, despite the fact that the irreducibly stochastic propensity to localize is not grounded in any categorical properties of the quantum system, it is nevertheless strongly explanatory of the definiteness of the macroscopic world of our experience. We should therefore change our prejudices concerning ungrounded dispositional properties as being always explanatorily empty: the explanatory power of GRW is given by the unification that the modified equation accomplishes between the dynamical evolution of quantum systems and the classical evolution of macroscopic systems.45 QM and the Problem of the Reducibility of Dispositional to Categorical Properties Summarizing, we have seen that in all of the most important interpretations of QM i. dispositional properties – whether probabilistic or deterministic – have a crucial role; ii. the only interpretation in which we seem to have reductionism about dispositional, contextual2 properties is Bohm’s. Before trying to draw our conclusion about these two claims, it is important to recall their relevance for our purpose: if in the most-debated interpretations of QM there are no categorical properties to which one could reduce the dispositional properties, the philosophical thesis that tries to reduce the latter to the former is in serious difficulties, because it is in conflict with the most fundamental theory of matter of our day.

43 J. Bell, ‘Are There Quantum Jumps?,’ in C.W. Kilmister (ed.), Schrödinger. Centenary Celebration of a Polymath (Cambridge, 1987), p. 44. 44 This historical parallel is due to Van Fraassen. See G.C. Ghirardi, Un’occhiata alle carte di Dio (Milano, 1997), pp. 379-381. 45 Friedman and Kitcher have both proposed a theory of explanation according to which to explain means to unify. M. Friedman, ‘Explanation and Scientific Understanding’, Journal of Philosophy, 71 (1974): 5-19 and P. Kitcher, ‘Explanation, Conjunction and Unification’, Journal of Philosophy, 73 (1976): 207-212.

Dispositions, Relational Properties and the Quantum World

267

In order to justify the last statement, the three following comments to i above are appropriate. First: the fact that even the Copenhagen interpretation of QM, that is closest to instrumentalism – and, therefore, closest to a philosophical position that resists from drawing ontological lessons from physical theories – is forced to invoke a strong form of contextualism2 is a strong point in favour of the claim that quantum properties are essentially dispositional. Second: the fact that, in interpretations as ontologically and methodologically different as the Copenhagen and the hidden variables interpretation, the contextualism2 of ‘phenomena’ (i.e. their dispositionality) is regarded as fundamental seems to teach us something very deep about the holistic ontology of QM, independently of our philosophical tastes. For Bohr as well as for Bohm, such contextualism refers to the inseparability of classical apparatuses and the behaviour and properties revealed by QM-systems, in the same sense in which, in the special theory of relativity, one cannot separate space from time except by making a conventional choice of an inertial frame. Despite the big difference between Bohr and Bohm, which lies of course in the latter’s postulation of non-contextual, always defined observables (positions of particles), we should be aware as well of the similarities, which have been stressed independently also by Bell. Third, to the extent that it is not absurd to attribute to entity realists like Bohr the view that microsystems have dispositional, probabilistic tendencies to display well-defined outcomes in given experimental contexts – whose function is to somehow ‘extract’ ‘latent aspects’ from a real (i.e. mind-independent) entity – Bohr’s and Heisenberg’s positions appear to be very similar. Unfortunately, claiming that a micro-system M has a ‘real disposition’ to show a certain behaviour in a measurement context has little explanatory power, as it just amounts to saying that if we measure the ‘fuzzy entity’ M we get a definite outcome. Bohr’s obsession with the language of classical physics as the transcendental condition of possibility to talk about a quantum system may have the unfortunate consequence of preventing us from learning more about the nature of quantum dispositions. In a word, Bohr’s and Heisenberg potentialities are as ungrounded as GRW’s ‘hits’, but while the latter’s dispositions to localize play an extremely important explanatory function, the former types of dispositions don’t. So, contrary to the common cliché according to which ungrounded dispositions carry as much explanatory power as the famous virtus dormitiva in Moliere’s comedy, it is not the ungroundedness of dispositions that matters, but the richer content of the GRW theory and therefore the different context of explanation. Coming now to ii, and wanting to assess at the same time Clifton’s and Pagonis’s claim that Bohmian dispositions are unremarkable despite their contextuality2, we must avoid possible sources of misunderstandings. We have seen that in Bohm’s theory there cannot be any talk of a possessed value of a contextual2 property before measurement and that, in this sense, there cannot be any pre-existing, real property of ‘having a spin in a certain direction’ to be revealed by measurements. Having ‘no definite spin in a given direction’ seems to imply ‘having no spin at all in that

268

Dispositions and Causal Powers

direction’, in the same sense in which ‘having no definite colour’ implies ‘having no colour at all’. But then a remarkable difference between a quantum and a classical disposition would seem to emerge. On the one hand, it is meaningless to ask what is the possessed value of the spin of the particle independently of a specific measurement context. On the other hand, at least within a realist view of ‘classical’, macroscopic dispositions, it is clearly not meaningless to refer to the fragility of a glass independently of, and before of, a ‘breaking context’. Isn’t this a remarkable difference between macroscopic dispositions and quantum dispositions in Bohmian mechanics? By claiming that Bohmian dispositions are ‘nothing to write home about’, have Clifton and Pagonis46 not overlooked something? Furthermore, in any minimal realism about dispositions, with which I would agree, a glass possesses the fragility independently of the breaking context, and the disposition ‘fragility’ can be attributed to a glass even if it will never break. Let us agree that there is a real property of ‘being fragile, or being disposed to break in such and such a context’, even before and independently of its manifestation; in monistic views, such a property exists precisely because it is identical with the microstructural categorical property of the glass, plus context. However, this plausible realism (whether non-reductive or reductive) about macroscopic dispositions seems to mark a difference with Bohmian spin-directions, since there is no way for us to refer to a contextually1 possessed value of spin-direction independently of a measurement context. This criticism of Clifton and Pagonis, however, is based on a subtle mistake. What we should say to defend the perfect analogy between a Bohmian ‘spin in a given direction’ and classical dispositions like those possessed by glass is that the latter’s ‘propensity to break in a certain context’ corresponds to ‘the propensity to manifest a definite spin in a certain measurement context’, which is clearly possessed by Bohmian particles also before measurement, and independently of it. In the Bohmian case, the correct analogy with the classical disposition ‘fragility’ is not given by ‘having a definite spin’ but by ‘the disposition to have a definite spin in a given direction’, which is linked to the fact that elementary particles can be deflected by magnetic fields, analogously to tiny magnetic needles that in some circumstances do not point anywhere (when they are in a superposition of spin directions). It follows that the manifestation of the two dispositions is, respectively, the breaking event in one case and the acquisition of a definite spin in the other. Accordingly, and even from the viewpoint of a realist position about dispositions, Clifton and Pagonis are correct in holding that the contextualism of values in Bohm’s theory is not remarkably different from the dispositionality of ordinary classical properties. Coming now to the reducibility thesis in the Bohmian theory, we should ask whether the property ‘being disposed to have a certain definite value of spin’ can 46 R. Clifton and C. Pagonis, ‘Unremarkable Contextualism: Dispositions in the Bohm theory’.

Dispositions, Relational Properties and the Quantum World

269

be regarded as being identical to the categorical basis, given by the non-contextual position of the particle whose spin is to be measured, the magnetic moment of the particles, plus a specific measurement context (a certain orientation of a Stern-Gerlach apparatus). Given the deterministic nature of the theory, it is certainly appropriate to claim that the measurement context, together with the non-contextual observable (position) and the propensity to manifest a certain spin in a given direction, uniquely determines or causes the particle to having such and such a spin. This highly plausible talk would seem to make room for a dualistic theory of a Placean kind, to the extent that the latter is committed to the view that the categorical base causes (and is not identical with) the disposition, and the latter in its turn, together with the context, causes the manifestation. According to Place,47 the fragility is caused and explained by the molecular structure of the glass, and it is the fragility that causes and explains the breaking of the glass by a stone. If this causal analysis were correct, the reducibility which Clifton and Pagonis defend, which they read as determinability, would be a misnomer, for the simple reason that an effect (the possession of a spin in a definite direction) cannot be identical to its cause (the position of the particle, its magnetic properties and the orientation of the Stern Gerlach apparatus). However, Place’s account of dispositional properties seems to multiply causes without necessity. The stone is certainly a cause of the breaking, together with the molecular structure of the window glass: if the structure of the window had been different, the stone would have not broken it. But claiming that fragility, referred to the glass as macroscopic entity, is an additional cause of the breaking introduces an unnecessary link in the narrative: as we have seen in the previous chapter, the predicate ‘fragile’ has a predictive function that refers to certain contexts, but the cause of the breaking and its true explanation lies in the microscopic structure of the glass. Fragility by itself can provide only very fragile explanations. But not because it is a disposition: as we have seen above, in some contexts dispositions are explanatory. This discussion seems to apply perfectly also to quantum, Bohmian dispositions. The quantum dispositional property ‘propensity to show a certain spin’ should be regarded as identical with a complex state of affair, constituted by the possessed position and other purely contextual1 properties of the particle. Note that we cannot claim that the categorical property (position) is a sufficient cause for the manifestation of the disposition, in the same sense in which the microstructure of glass is not sufficient for the breaking event.

47 U.T. Place, ‘Dispositions as Intentional States’, pp. 19-32.

270

Dispositions and Causal Powers

Conclusions Since only Bohmian dispositions are reducible to context and non-contextual variables,48 the possibility of considering dispositions as reducible to intrinsic properties seem to depend on the prospect of success for the Bohmian interpretation of QM. The fact that the non-local contextuality of the fundamental physical observables is so widespread a feature of various interpretations of QM confirms that the quantum world cannot be conceived of as inhabited by entities whose identity is given by intrinsically possessed properties. The relational aspect of such entities is much more important to understand what they really are. One of the difficulties in the literature about dispositions is to find a clear-cut criterion to separate the dispositional from the categorical. Given this difficulty, some authors even deny there is an interesting distinction to be drawn. However, in the context of QM ‘categorical’ certainly entails ‘intrinsically possessed’ (contextual1), so that relational entails ‘dispositional’ or contextual2: it then follows that relationality is sufficient for dispositionality, at least in the context of the philosophy of QM. If relationalism about quantum properties just is another name for strong contextualism (contextualism2), then irreducible relationalism entails irreducible dispositionality. And the irreducibility of quantum properties depends on its turn on the well-known, universally agreed-upon lack of definiteness of the quantum world: dispositions in QM are irreducible simply because there are no categorical, non-dispositional properties to which they can be reduced.49

48 The proviso to add to this conclusion involves the many-worlds interpretation, in the sense explained above. 49 I thank Federico Laudisa for having read a previous version of this manuscript and the audience in Paris for valuable suggestions and criticisms. I also thank Max Kistler for his suggestions in matters of both style and content.

Chapter 13

Are Specific Heats Dispositions? Anouk Barberousse

Introduction My aim in this paper is a modest one: I will present some constraints physical theories set on the disposition debate, by means of a particular example. I hope this example will not only shed light on the discussion, but also show that many dispositions rest on generic laws linking the macroscopic scale to the microscopic one. This fact, it seems to me, makes the debate about the nature of dispositions depend on a more general one about the various levels at which the material world is analysed in physics. Specific Heats as Bona Fide Dispositions Specific heats are, at least apparently, paradigmatic dispositions. The specific heat of a sample of a given substance in a given physical state (solid, liquid or gaseous) may indeed be defined as the quantity of heat this sample must absorb in order to increase its temperature by 1 degree. This rough definition may be made more precise in different ways, for instance by giving the following example. The predicate ‘having a specific heat of 4.18 joules deg.- 1gram.-1’ may be defined by the following subjunctive conditional: If the temperature of the sample (here, a sample of water) was raised by 1 degree, provided its temperature remains within a certain range (around 14°C for instance), it would absorb an amount of heat of 4.18 joules per gram.

This seems to be a typical disposition-expressing statement, namely, a ‘statement about how a thing [here, a sample of a material substance] is disposed to respond to a stimulus’.1 Another refinement of the first, rough definition of specific heats may be given by exhibiting the quantitative relations specific heat-predicates appear in within the framework of a physical theory, namely thermodynamics:

1 D. Lewis, ‘Finkish Dispositions’, in H. Sankey (ed.), Causation and Laws of Nature (Dordrecht, 1999), p. 143.

Dispositions and Causal Powers

272

i. We can first give an operational definition of the quantity called ‘heat capacity’ or ‘specific heat’: ⎛ dQ ⎞⎟ C y ≡ ⎜⎜ ⎜⎝ dT ⎟⎟⎠ y dQ refers to an infinitesimal quantity of heat which cannot necessarily be expressed by an exact differential, y refers to a macroscopic parameter, or to a set of such parameters, which is kept constant. Cy is defined in the limit as dQ → 0 (or dT → 0 ). It is in general a function of T and y: Cy=Cy(T, y) Whereas the amount of heat dQ which needs to be added to produce a given temperature change dT of a homogeneous sample is proportional to the quantity of matter contained therein, the specific heat per mole (or per gram) only depends on the nature of the substance, not on the amount present in the sample. ii. We can further define the ‘specific heat per mole’: cy ≡

1 1 ⎛ dQ ⎞⎟ C y = ⎜⎜ ⎟ ν ν ⎜⎝ dT ⎟⎠y

where ν is the number of moles contained in the sample under study, as well as the ‘specific heat per gram’: c 'y ≡

1 1 ⎛ dQ ⎞⎟ C y = ⎜⎜ ⎟ m m ⎜⎝ dT ⎟⎠y

where m is the mass of the sample. The specific heats which are mostly used are the specific heat at constant volume cv and the specific heat at constant pressure cp. A last point is that specific heats, as most, and perhaps all, dispositions investigated in scientific contexts, have been subjected to precise measurements. For instance,

Are Specific Heats Dispositions?

273

the specific heat of water was first determined by Joule, through very sophisticated experiments he made from 1843 to 1849: at a pressure of 1 atmosphere, and at 15°C (or 288.2°K), its value is 4.18 joules deg.-1 gram.-1. The debate on the nature of dispositions has focused on two types of dispositional predicates: predicates of ‘everyday’ dispositions, like fragility, and predicates of scientific dispositions. It seems to me desirable, in order to raise the various questions which have emerged from the debate in the clearest possible way, not to step back, in the case of scientific dispositions, from the scientific context in which they are conceptually and experimentally investigated. The best strategy to adopt in order to solve the problems concerning dispositions is indeed likely to be the following: in so far as at least some disposition concepts are scientific concepts in the sense that they are used within scientific investigations, the questions they raise have to be examined in connection with these scientific uses, not just by analogy with what can be said about ‘everyday’ dispositions. With regards to specific heats, the relevant theories are thermodynamics and statistical mechanics, both classical and quantum. What is specific to thermodynamics is that within it, the quantities dealt with are macroscopic ones, and the equations usually contain measurable quantities. Specific heats are such measurable quantities. Not only does the predicate ‘having a specific heat of x joules degrees-1 moles-1’ seem genuinely dispositional, but the property it refers to also seems to be an unproblematic causal power, if anything is: specific heats were once called ‘caloric capacities’, and are still called ‘Wärmekapazität’ in German, this name being connected to at least one way of conceiving them (as a measure of a sample’s capacity to deliver heat). And indeed, like other bona fide scientific dispositions, specific heats appear in many thermodynamic laws, as in the following examples: i.

⎛∂ S ⎞ C y = T ⎜⎜ ⎟⎟⎟ ⎜⎝ ∂T ⎠ y since, according to the second law of thermodynamics, dQ =TdS.

ii. In the situation where all external parameters are kept constant, i.e. whenever the system does no macroscopic work, we have ⎛∂ E ⎞ ⎛∂ S ⎞ CV = T ⎜⎜ ⎟⎟⎟ = ⎜⎜⎜ ⎟⎟⎟ ≥ 0 ⎜⎝ ∂T ⎠ ⎜⎝ ∂T ⎟⎠V V (since dW = 0 , the first law of thermodynamics is: dQ = d E , in infinitesimal processes). iii. In ideal gases, we have dE = ν cV dT

Dispositions and Causal Powers

274

and, since pdV = ν RdT and R=Nak (Na is the Avogadro number and k Boltzmann’s constant), we have cp = cV + R The ratio γ≡

cp cV

= 1+

R cV

can be determined from the velocity of sound in the gas, or measured directly. iv. Finally, C p − CV = VT

α2 κ

where 1 ⎛ ∂V ⎞ α = ⎜⎜ ⎟⎟⎟ v ⎜⎝ ∂T ⎠ p is the volume coefficient of expansion and 1 ⎛ ∂V ⎞ κ = − ⎜⎜⎜ ⎟⎟⎟ V ⎝ ∂ p ⎟⎠T is the isothermal compressibility. Moreover, specific heats seem to be properties of ‘things’ rather than of kinds.2 It is this sample of gas that must absorb some quantity of heat in order to increase its temperature by 1 degree, in virtue of its being a material thing. The power to absorb this precise amount of heat may of course be treated as a property of a kind, when this kind is defined as the set of all material samples yielding this amount of heat in the relevant conditions. But the disposition to absorb some amount of heat is not characteristic of any kind of material objects, when such a kind is defined by other characteristics, like being a sample of H2. This remark relates to specific heats as 2 Cf. M. Kistler, ‘Some Problems for Lowe’s Four-Category Ontology’, Analysis, 64.2 (2004): 146-151.

Are Specific Heats Dispositions?

275

well as to other thermodynamic dispositions, and even perhaps to all dispositions. So, Lowe’s claim that ‘individual objects possess their various natural “powers” in virtue of belonging to substantial kinds which are subject to appropriate laws’,3 or else that ‘a thing’s properties and the laws they occur in determine the kinds it belongs to and hence which dispositional predicates applies to it [the thing]’4 is only true in so far as determinate values of specific heats are considered, but false if the generic power of yielding heat is, because this generic power does not pick out any distinct kind of material things. Let us go further in the analysis of the physical and metaphysical characteristics of samples of matter having the capacity to yield amounts of heat. The property of being a sample of some substance, in the current sense of this expression,5 involves another one, the consequences of which are seldom examined, but which is crucial anyway: the sample must be of macroscopic size. This is so even if, in so far as it is large enough,6 its specific heat per mole does not depend on its mass, i.e. on the quantity of matter it contains. The condition that the system be of macroscopic size may be defined in the following way: a macroscopic object is composed of such a number of components7 that at least some of its properties result from the interactions among them, these interactions being impossible to represent nonstatistically.8 The consequences of this definition for the analysis of thermal processes will be presented in the second part of this paper. To sum up, specific heats seem to be unproblematic dispositions: they are wellidentified properties; they may be, and have been, subjected to precise measurements; they are an important part of the knowledge one can gain about the different substances; and they appear in many scientific, namely thermodynamic, laws. The fact that specific heats are dispositions in the least controversial sense this word can 3 Cf. E.J. Lowe, ‘Dispositions and Laws’, Metaphysica, 2 (2001): 13. 4 Cf. E.J. Lowe, Kinds of Being: A Study of Individuation, Identity and the Logic of Sortal Terms (Oxford, 1989), p. 163. 5 Excluding microscopic samples, constituted of a few atoms. The discussion on thermal dispositions would be meaningless in such cases. 6 ‘Being large enough’ here is just a synonym for ‘being of macroscopic size’. 7 The determination of the number of components which would make this definition more precise is a difficult matter: for some thermodynamic properties of not too condensed gases, computer simulations suggest that a few thousands (pseudo)-molecules would suffice, cf. D. Levesque and L. Verlet, ‘Computer “Experiments” on Classical Fluids. III. TimeDependent Self-Correlation Functions’, Physical Review A, 2 (1970): 2514. The point is that too few components cannot give rise to the properties which are usually conceived as powers, cf. for instance J.L. Lebowitz, ‘Macroscopic Laws, Microscopic Dynamics, Time’s Arrow and Boltzmann’s Entropy’, Physica A, (1993): 1-27. 8 M. Kistler, ‘The Causal Efficacy of Macroscopic Dispositional Properties’, this volume, p. 104, uses the word ‘macroscopic’ in a less specific way. The difference between these two proposed definitions of the distinction between ‘being of microscopic size’ and ‘being of macroscopic size’ may derive from the difference between our purposes: this distinction is perhaps best seen as a context-relative one, depending on the specific problem one is investigating.

276

Dispositions and Causal Powers

take seems to imply that dispositions or powers are scientifically admissible after all. Why indeed should they be suspected of metaphysical impropriety? Why should the predicate ‘having a specific heat of x joules per degree’ be eliminated from the scientific vocabulary? It is precisely because they are causal powers that they can be measured, like many properties investigated in physics; and doing without them would probably imply a certain loss of knowledge.9 Now it is well known that their being a material thing’s power to yield such or such quantity of heat when its temperature is increased by 1 degree and their appearing in thermodynamic laws is not all that is scientifically relevant and interesting about specific heats. Like other macroscopic properties, they have been subjected to different, sometimes competing, explanations or analyses. It is, or, more precisely, it has once been entirely justified, in the context of the practice of physics, to ask why a sample of water, in the relevant range of temperature, yields 4.18 joules when its temperature is increased by one degree. This question is most certainly connected to the following one, even if the details of this connection are not easy to explicate: ‘What is the molecular basis of this particular value of a specific heat?’. Whether these two questions have the same answer is the subject of the following sections. The Physical Analysis of Thermal Powers Even if, as I emphasized, specific heats as such, i.e. as powers to deliver determined quantities of heat in certain circumstances, play, or have played, a non negligible role in physics, both on the experimental and on the theoretical side, they are not, today, considered as worthy of appearing among the most important quantities of physics. It is certainly not so easy to tell what are today’s most important, or fundamental, quantities of physics, be it only because the rather gerrymandered set of practices we call ‘physics’ is not unified. The notion of ‘fundamental physical quantity’ may have to be relativized to the context of a particular theory or set of theories. Now, in the general domain where specific heats are used, namely, the study of thermal phenomena, neither specific heats, nor other macroscopic quantities like pressure or temperature, are considered as fundamental. Let us go further into the description of the various theoretical contexts specific heats enter in. In thermodynamics strictly speaking, only macroscopic quantities are used, like internal energy, temperature, entropy, etc. It has once seemed possible to consider some of these macroscopic, thermodynamic quantities as the fundamental quantities of a unified physics: at the end of the nineteenth-century, the proponents of energetics, like Ostwald or Duhem, indeed thought that they could dispense with using the motion of molecules to account for all known physical phenomena. At the turn of the twentiethth-century, their opponents, who thought that fundamental quantities were mechanical, won this rather violent battle about 9 This was one main motivation for the renewal of dispositionalism; cf. R. Harré, ‘Powers’, British Journal for the Philosophy of Science, 21 (1970): 81-101.

Are Specific Heats Dispositions?

277

the foundations of physics, tolling the knell of other attempts in founding physics on macroscopic quantities. They defended a conception of physics in which the fundamental quantities were mechanical as well as microscopic. Specific heats have played a distinctive role in this history, precisely because what we, as philosophers, are willing to call their ‘basis’ was not very well captured by the kinetic theory of gases, further developed from 1901 onwards in statistical mechanics, although it was designed to do so.10 The main problem was the following: whereas the values of specific heats calculated from the principles of (classical) statistical mechanics were in agreement with the measured ones in the case of monoatomic gases, the same calculations led to more and more discrepancies between calculated and measured values as the number of atoms in the molecules at work was increasing (see Table 1). At the time, namely in the last quarter of the nineteenth-century, the origin of this problem was suspected to lie in the methods used to determine the number of degrees of freedom of polyatomic molecules, but the physicists had no way to solve it, because molecules were purely theoretical entities, whose structure was inaccessible to observation, even by indirect procedures. To put it briefly, in order to compute the values of macroscopic quantities, at least two steps are necessary. First, the values of microscopic quantities, like molecular velocities, are extrapolated by a combination of observations and theoretical principles; and second, one has to determine what the possible motions of a single molecule are: translation along three independent axes, rotation, vibration, etc. Without any detailed knowledge about the internal structure of atoms and molecules, nineteenth-century physicists were reduced to bet on these motions, and had to content themselves with conjectures deriving from analogies with observable cases, even though they were conscious that those analogies were unjustified and unsatisfactory.11 One of the consequences of these macroscopic analogies was that an important theorem of statistical mechanics, the so-called ‘equipartition theorem’, stating that at equilibrium, the total kinetic energy of a system is equally distributed among all its degrees of freedom, necessarily had wrong consequences whenever polyatomic gases were at work, because in that case, the number of degrees of freedom of their molecules was erroneously determined. It is only when quantum theory was adopted that the famous ‘specific heats problem’ was solved, because the quantization of the authorized levels of the atoms’ energies allowed to understand why the equipartition theorem failed in classical statistical mechanics. The Philosophical Analysis of Thermal Dispositions, Physically Informed After having briefly summarized the historical fate of specific heats, and sketched what problems they raised for the physical theories available 100 years ago, let us 10 Cf. J. Gibbs (1902), Elementary Principles in Statistical Mechanics (New York, 1965). 11 Cf. S. Brush, E. Garber and C.W.F. Everitt (eds), Maxwell on Heat and Statistical Mechanics. On ‘Avoiding All Personal Enquiries’ of Molecules (London, 1995).

278

Dispositions and Causal Powers

draw some conclusions from this history, in order to understand how the concept of specific heat, as referring to some macroscopic disposition, is analysed in physics, and, more generally, can be analysed in terms of some molecular basis. In other words, we now have to ask, paraphrasing Mellor,12 what in the microscopic world, by making the statement ‘this sample of water has a specific heat of 4.18 joules…’ true, makes it safe to infer ‘this sample will deliver 4.18 joules per gram’ from ‘its temperature is increased by 1 degree’. My starting point in this discussion is a quote from Mellor,13 which focuses on temperature not specific heats, but this difference has no bearing on the criticism I will propose against Mellor’s thesis. Many philosophers, misled by Kripke,14 identify temperature with its bases in gases, namely – assuming for simplicity a certainly false (because deterministic) kinetic theory of gases–the mean kinetic energy of gas molecules. But as an identity thesis, this will not do, for at least three reasons (apart from the falsity of the theory). First, the thesis implies that increasing the velocity, and hence the kinetic energy, of a single particle would automatically raise its temperature, which is absurd. Second, there is in reality no property of mean kinetic energy to identify temperature with, any more that there are in reality the 2.4 children that average families have: there are only the actual kinetic energies of individual molecules, whose mean value is what the kinetic theory relates to a gas’s temperature. Third, and most important here, not all temperatures of entities can be identified with molecular kinetic energies: in particular, radiation temperatures cannot, since radiation contains no molecules. I disagree with this alleged dismissal of the identity thesis on the first two points – the third point being inessential. True, radiation contains no molecules, but the expression ‘mean molecular energy’ is usually used (or should be used) as an abbreviation of the general definition of absolute temperature, in terms of the distribution function Z, when it is applied to material bodies. This definition also applies to radiation, where it receives another interpretation, linked to the photons’ distribution of energy. Not all temperatures can indeed be identified with molecular kinetic energies, but this does not affect the validity of a correctly formulated identity thesis along Kripke’s lines, because the identity thesis bears on the general definition of absolute temperature, not on its interpretations. However, the discussion about the first two points in Mellor’s paper only concerns temperatures of material bodies. My criticism of the first point is the following: Mellor’s claim can be understood in two ways, depending on what the referent of ‘its’ is in the phrase: ‘the velocity, and hence the kinetic energy, of a single particle would automatically raise its temperature’. If ‘its’ refers to the single particle, the mentioned increasing of temperature is non-sensical, because, among the application conditions of the 12 D.H. Mellor, ‘The Semantics and Ontology of Dispositions’, Mind, 109 (2000): 757780. 13 Ibid., p. 774. 14 S.A. Kripke, ‘Identity and Necessity’, in M.K. Munitz (ed.), Identity and Individuation (New York, 1971), pp. 135-164.

Are Specific Heats Dispositions?

279

predicate ‘temperature’, there is one asking that it has to be applied to macroscopic bodies. If on the other hand, ‘its’ refers to such a macroscopic, material body, then the claim is simply false, as seen by simply computing the effect on the mean kinetic energy of, say, 1023 oxygen molecules at standard pressure of the increasing of the velocity of one molecule by 10 m.s-1: the difference of values does not have any physical meaning – at all. In physics, orders of magnitude do matter. The second point is more tricky. There is no doubt that the question of the ‘reality’ of mean values is central to the discussion of the possible bases of the kind of dispositions we are considering here, namely temperatures, specific heats, entropies, and the like. With regards to the question of the meaning of mean values, Mellor contents himself with identifying mean kinetic energies with mean numbers of children in families, and with denying any reality to both mean values. I don’t exactly know about the consequences on a state’s economy of an increase of 0.1 in the average number of children per family, as far as concept types are considered, but I do know that statistical mechanics, and more generally statistical physics, is precisely the theory of the effects of significant increases in mean values of microscopic quantities. Statistical physics may indeed be considered as the empirical study of the measurable effects of the aggregation of large numbers of molecules in different microscopic states. If you take, say, 10 molecules, these effects will not appear; if you take 1023 of them, the aggregate will have the capacities itemized in thermodynamic laws. The ‘reality’ of mean values is thus not a simple question: denying them any ontological status implies considering that statistical physics is not an empirical theory, but, for instance, a theory of statistical inference. This option cannot be excluded a priori, but has to be argued for, and not just assumed without any justification, as in Mellor’s paper. In order to further examine the question of the relation between (thermal) dispositions and their basis, I shall now discuss two questions also raised in Mellor’s paper, namely the questions of double-counting and of overdertermination of causing.15 Mellor starts from an example of a sample of gas at constant volume, the pressure of which increases as its temperature increases, as required by Boyle’s law. His question about the causing of the increase ΔP of gas G’s pressure is the following: But what then is the real cause of ΔP? Is it ΔT, the increase in G’s temperature T, or ΔE, the increase in T’s basis in G, namely the mean kinetic energy of G’s molecules? And if we say it is both, are we not double-counting the causes of this increase in G’s pressure, and thereby overdertemine it?

To answer this question, Mellor contrasts the gas case with the ‘well-known example of a spark causing fuel to burn in oxygen’, and claims that in the latter, no overdetermination of the causing of the fire occurs, because each cause needs the other two in order to yield its effect. His first claim is:

15 D.H. Mellor, ‘The Semantics and Ontology of Dispositions’, pp. 775-776.

280

Dispositions and Causal Powers that is not so with our gas G, where neither ΔT nor ΔE seems to need the other in order to cause the increase ΔP in G’s pressure. And this poses a problem for any theory of causation which requires causes to be, in the circumstances, both sufficient and necessary for their effects. For each of these two apparent causes is, independently of the other, sufficient for the effect, neither of them can also be necessary for it; and so neither can be a sufficient and necessary cause at all, which is absurd. For if neither raising a gas’s temperature nor raising the mean kinetic energy of its molecules is what causes its pressure to increase, what is?

However, the statement that neither invoked cause is necessary for the increase in pressure seems hasty: it is true that a decreasing of the volume of the gas sample will also yield an increased pressure, but if the volume is kept fixed, an increase in temperature, or, equivalently in this case, of mean molecular kinetic energy, is a necessary condition for an increase in pressure. Moreover, the constant volume condition is indeed presupposed in the discussion, which means that, pace Mellor, an increase of T or of E is a necessary condition for an increase in P. Mellor’s argument is then as follows: Fortunately, however, that problem need not concern us, since although it could arise here, it need not. For all we need, in order to avoid overdetermination, is a counterfactual link between ΔT and it gaseous basis ΔE to ensure that, in the circumstances, we shall get ΔT if and only if we get ΔE. And such a link does not require T either to be identical to E, or to supervene on E necessarily. All it requires is a contingent law of nature to give all gases (or all gases of a certain kind or kinds) a temperature T if and only if their molecules have a mean kinetic energy E. Such a law can make T and E, and hence ΔT and ΔE, supervene contingently on each other in gases, thus enabling each of these changes to be a sufficient and necessary cause of ΔP.

This, I maintain, is the right answer to the question of which of ΔT and its basis ΔE really causes ΔP: they both do. Mellor’s conclusion is twofold: first, nothing prevents us from acknowledging the existence of real and categorical properties of both kinds, i.e. thermal and kinetic, and second, in general, all a factual property F needs, in order to have in things of any kind K a distinct and non-overdetermining factual basis B, is that among the many laws these two properties occur in is the law that all K things are F if and only if they are B.

My first point is that it is at least unusual to express the relation between temperature and mean kinetic energy of a set of molecules by a biconditional of the form ‘a gas of kind K (e.g. in the relevant range of pressure and temperature) has temperature T if and only if the mean kinetic energy of its molecules is E’. The relevant law is more precisely expressed by a sentence like: ‘in so far as a gas of kind K is correctly represented by the usual models of kinetic theory (for instance the billiardball model, or the model of molecules as centres of forces), the calculation of the mean kinetic energy of its molecules yields a result which is very close to the value

Are Specific Heats Dispositions?

281

of the temperature of a sample of such a gas as measured by the usual methods’. The biconditional formulation cannot be read as merely summing up the longer one, because it exceeds the meaning of the sentences which can be derived from the physical theories as they are actually used. It indeed overlooks the proviso restricting the validity of such claims 1) to situations in which satisfactory models are available, and 2) to the limits within which those models are valid. The quick reading, attributing, so to say, the wrong type of reality to the referents of the physical concepts, is the source of much confusion. It mixes up the rather different theoretical roles the concepts of temperature and of mean kinetic energy of molecules play in the involved theories, namely thermodynamics and the kinetic theory of gases or statistical mechanics. The thermodynamic concept of temperature appears in phenomenological relations, i.e. relations among measurable quantities, whereas the concept of molecular mean kinetic energy is not susceptible to be linked to any direct measurements, and results from various conjectures about, and idealizations of, the microscopic world. I claim that when this layer of conceptualization and modellization is ignored, namely when one directly draws ontological conclusions from scientific statements that are only meaningful within a certain theoretical framework, the meaning of these conclusions is too underdetermined for the validity of these inferences to be checked. In particular, forgetting the distinctions among the roles the different concepts play in the theories within which they are meaningful leads to ill-posed questions about the relations between macroscopic dispositions and their molecular basis. Let us now go deeper into the analysis of the relations between thermal dispositions and their molecular basis. Thermal dispositions are analysed, or explained, in the context of classical statistical mechanics (but things are not very different in quantum statistical mechanics): 1. by properties of individual molecules, like their masses and velocities, or energy levels; 2. by statistical properties of aggregates of molecules. Is the second set of properties, the statistical ones, to be identified with the first? Namely, are thermal dispositions entirely explainable by the ever-changing properties of all the individual molecules contained in a sample, like ‘having a velocity of x m.s-1 at time t’? Properties of individual molecules are only instantiated by one and the same molecule during very short time intervals, which implies that thermal dispositions are explained by dynamical, not static interactions. The example of thermal dispositions is thus useful to develop in the disposition debate, be it only because it reminds us that fragility, whose molecular basis involves mostly static interactions, is not the paradigmatic case of a disposition. Moreover, considering a dynamical basis forces us to minimize the importance of an often-used distinction, between occurrent and dispositional properties. It is sometimes said that the basis of dispositions has to be made of occurrent properties, namely lasting, unchanging ones. But the

Dispositions and Causal Powers

282

occurrent properties of individual molecules, in the case of thermal dispositions, are only ‘lasting’ in an abstract sense: examples of such properties are the molecules’ instantaneous velocities at times t1, t2, Moreover, and more importantly, mentioning the occurrent properties of molecules is by no means sufficient to establish the molecular basis of thermal dispositions. To this aim, it is equally necessary to mention what the available or possible states of molecules are.16 This can be done only if the global properties of the whole sample, like its total internal energy, or its temperature, are given. Consequently, the statistical properties of an aggregate of molecules are not identical to the set of the individual, occurrent properties of its components: as we have seen contra Mellor, if a single molecule, or even a few molecules, go much faster than the others, or much slower, their individual, occurrent properties do not contribute to the macroscopic ones. To sum up, thermal dispositions may be said to be identical to a set of properties of their molecular basis provided that statistical properties are included in this set. But they are not identical to any set of strictly occurrent properties of individual molecules: the statistical tools used to compute their values from microscopic quantities are not just formal tricks enabling us to go beyond our cognitive capacities, but do afford a sort of imitation of the relation between the microscopic and the macroscopic scales of the physical world, as shown by some computer simulations.

Table 1

The specific heats problem: For polyatomic gases, calculated and measured values of specific heats disagree

Gas

Symbol

cV (measured)

cp (measured)

cp (computed)

Helium

He

12.5

1.666

1.666

Argon

Ar

12.5

1.666

1.666

Nitrogen

N2

20.6

1.405

1.407

Oxygen

O2

21.1

1.396

1.397

Carbon dioxide

CO2

28.2

1.302

1.298

Ethane

C2H6

39.3

1.220

1.214

16 The main methodological tools of statistical physics, the various Gibbs ensembles, have been made to allow the definition and counting or measure of these possible states.

Bibliography Albert, D. and Loewer, B., ‘Interpreting the Many-Worlds Interpretation’, Synthese, 86 (1988): 87-98. Alexander, P., Ideas, Qualities and Corpuscules (Cambridge: Cambridge University Press, 1985). Alston, W.P., ‘Dispositions, Occurrences and Ontology’, in R. Tuomela (ed.), Dispositions (Dordrecht: Reidel, 1978), pp. 359-388. Ammerman, R.R., Classics of Analytic Philosophy (New York: McGraw-Hill, 1965). Anscombe, G.E.M., ‘Causality and Determination’ (1971), in E. Sosa and M. Tooley (eds), Causation (Oxford: Oxford University Press, 1993), pp. 88-104. Anscombe, G.E.M., Collected Philosophical Papers, vol. II (Oxford: Blackwell, 1981). Anstey, P.R., The Philosophy of Robert Boyle (London: Routledge, 2000). Aristotle, ‘On Coming-to-be and Passing-away’, trans. E.S. Forster, in Aristotle, On Sophistical Refutations, On Coming-to-be and Passing-away, On the Cosmos, Loeb Classical Library (Cambridge: Harvard University Press, 1955; repr. 1992), pp. 157-329. Aristotle, ‘On Interpretation’, trans. H.P. Cooke, in Aristotle, Categories, On Interpretation, Prior Analytics, Loeb Classical Library (Cambridge: Harvard University Press, 1938; repr. 1983), pp. 111-179. Aristotle, Physics, 2 vols, trans. P.H. Wicksteed and F.M. Cornford, Loeb Classical Library (Cambridge: Harvard University Press, 1929 and 1934, revised 1957, repr. 1980). Aristotle, The Metaphysics, trans. H. Tredennick, Loeb Classical Library (Cambridge: Harvard University Press, 1933, repr. 1989). Armstrong, D.M., A Materialist Theory of the Mind (London: Routledge, 1968). Armstrong, D.M. (1973), ‘Beliefs as States’, in R. Tuomela (ed.), Dispositions (Dordrecht: Reidel, 1978), pp. 411-425. Armstrong, D.M., Nominalism and Realism (Cambridge: Cambridge University Press, 1978). Armstrong, D.M., A Theory of Universals (Cambridge: Cambridge University Press, 1978). Armstrong, D.M., What is a Law of Nature? (Cambridge: Cambridge University Press, 1983). Armstrong, D.M., Martin, C.B. and Place, U.T., ed. T. Crane, Dispositions: A Debate (London: Routledge, 1996).

284

Dispositions and Causal Powers

Armstrong, D.M., ‘Place’s and Armstrong’s Views Compared and Contrasted’, in D.M Armstrong, C.B. Martin and U.T. Place, Dispositions: A Debate (London: Routledge, 1996), pp. 33-48. Armstrong, D.M., A World of States of Affairs (Cambridge: Cambridge University Press, 1997). Armstrong, D.M., ‘The Causal Theory of Properties: Properties according to Shoemaker, Ellis and others’, Philosophical Topics, 26 (1999): 25-37; also published in Metaphysica, 1 (2000): 5-20. Ashcroft, N.W. and Mermin, N.D., Solid State Physics (Philadelphia: Saunders College, 1976). Avicenna, Liber de Philosophia prima sive Scientia divina, ed. S. Van Riet, Louvain, E. Peeters - Leiden, E.J. Brill, I-IV (1977), V-X (1980). Bacon, J., Universals and Property Instances (Oxford: Blackwell, 1995). Barbour, J.B., Absolute or Relative Motion, vol. 1 (Cambridge: Cambridge University Press, 1989). Barrett, J., The Quantum Mechanics of Minds and Worlds (Oxford: University Press, 1999). Bartels, A., ‘The Idea which we call Power. Naturgesetze und Dispositionen’, Philosophia Naturalis, 37 (2000): 255-268. Bell, J., ‘Are There Quantum Jumps?,’ in C.W. Kilmister (ed.), Schrödinger. Centenary Celebration of a Polymath (Cambridge: Cambridge University Press, 1987). Bennett, J., Events and their Names (Cambridge: 1988). Berkeley, G., Treatise Concerning the Principles of Human Knowledge (Dublin, Jeremy Pepyat, 1710). Berkeley, G., ‘Of Motion (De Motu)’ (1721), in A.A. Luce and T.E. Jessop (eds), The Works of George Berkeley, Bishop of Cloyne, (London: Thomas Nelson and Sons, 1951), vol. IV, pp. 11-30 (Latin) and pp. 31-52 (English translation). Bhaskar, R., A Realist Theory of Science (Brighton: Harvester, 1978). Bigelow, J., The Reality of Numbers (Oxford: Oxford University Press, 1988). Bird, A., ‘Dispositions and Antidotes’, Philosophical Quarterly, 48 (1998): 227234. Bird, A., ‘Further Antidotes: A Reply to Gundersen’, Philosophical Quarterly, 50 (2000): 229-233. Bird, A., ‘Necessarily Salt Dissolves in Water’, Analysis, 61 (2001): 267-275. Bird, A., ‘On Whether Some Laws Are Necessary’, Analysis, 62 (2002): 257-274. Bird, A., ‘The Dispositionalist Conception of Laws’, Foundations of Science, 10 (2005): 353-370. Blackburn, S., ‘Filling in Space’, Analysis, 50 (1990): 62-65. Bloch, L., Les Origines de l’Ether et la physique de Newton (Paris: Alcan, 1908). Bohr, N., Atomic Physics and Human Knowledge (New York: Science Editions, 1958). Boscovich, R.G., Theoria (Venice, Typographia Remondiniana, 1763), published as A Theory of Natural Philosophy (Boston: MIT Press, 1966).

Bibliography

285

Boyle, R., ‘The Origins and Forms of Qualities’, (1772), in M.A. Stewart (ed.), Selected Philosophical Papers of Robert Boyle (Manchester: Manchester University Press, 1979), pp. 1-96. Braddon-Mitchell D. and Jackson F., The Philosophy of Mind and Cognition (Oxford: Blackwell, 1996). Bridgman, P.W., The Logic of Modern Science (New York: The Macmillan Company, 1927). Bruner, J.S., In Search of Mind (New York: Harper and Row, 1983). Brush, S., Garber, E. and Everitt, C.W.F. (eds), Maxwell on Heat and Statistical Mechanics. On ‘Avoiding All Personal Enquiries’ of Molecules (London: Associated University Presses, 1995). Campbell, K., Abstract Particulars (Oxford: Blackwell, 1990). Carnap, R. ‘Eigentliche und Uneigentliche Begriffe’, Symposion: Philosophische Zeitschrift für Forschung und Aussprache, Bd. 1, H. 4, Berlin, 1927. Carnap, R., Der Logische Aufbau der Welt (Berlin: Weltkreis-Verlag, 1928; Hamburg: F. Meiner, 1998) trans. R.A. George, The Logical Structure of The World and Pseudoproblems in Philosophy (Chicago: Open Court, 2003). Carnap, R., ‘Testability and Meaning’, I, Philosophy of Science, 3 (1936): 419-471. Carnap, R., ‘Testability and Meaning’, II, Philosophy of Science, 4 (1937): 1-40. Carnap, R., Logische Syntax der Sprache (Wien: Springer, 1934), trans. A. Smeaton, Logical Syntax of Language (London: Paul, Trench, Trubner & Co., 1937; repr. La Salle, Ill.: Open Court, 2002). Carnap, R., ‘The Methodological Character of Theoretical Concepts’, in H. Feigl and M. Scriven (eds), Minnesota Studies in the Philosophy of Science, vol. 1 (Minneapolis: University of Minnesota Press, 1956), pp. 38-76. Carnap, R., Philosophical Foundations of Physics (New York: Basic Books, 1966). Carnap, R., ‘Meaning Postulates’, in Meaning and Necessity (Chicago: University of Chicago Press, 1947; 2nd edition, 1956), pp. 222-229. Carnap, R. Meaning and Necessity, a Study in Semantics and Modal Logic (Chicago: Chicago University Press, 1947; 2nd edition 1956). Cartwright, N., How the Laws of Physics Lie (Oxford: Clarendon Press, 1983). Cartwright, N., Nature’s Capacities and their Measurement (Oxford: Clarendon Press, 1989). Cartwright, N., ‘Aristotelian Natures and the Modern Experimental Method’, in J. Earman (ed.), Inference, Explanation, and other Frustrations: Essays in the Philosophy of Science (Berkeley: University of California Press, 1992), pp. 4471. Cartwright, N., ‘Precis of Nature’s Capacities and Their Measurement’, Philosophy and Phenomenological Research, 55, 1 (1995): 153-156. Cartwright, N., The Dappled World: A Study of the Boundaries of Science (Cambridge: Cambridge University Press, 1999). Cartwright, N., ‘In Favor of Laws that Are Not Ceteris Paribus After All’, Erkenntnis, 57, 3 (2002): 425-439. Causey, R.L., Unity of Science (Dordrecht: Reidel, 1977).

286

Dispositions and Causal Powers

Chalmers, A., ‘True Fundamental Laws in a Dappled, Patchwork World’, Popper Centenary Conference Lecture, ms. (Adelaide: Flinders University, 2002). Chisholm, R., ‘The Contrary-to-Fact Conditional’ (1946), in H. Feigl and W. Sellars (eds), Readings in Philosophical Analysis (New York: Appleton-Century-Croft, 1949), pp. 482-497. Chisholm, R., ‘Freedom and Action’, in K. Lehrer (ed.), Freedom and Determinism (New York: Random House, 1966), pp. 11-44. Chisholm, R., Person and Object (London: Allen and Unwin, 1976). Choi, S., ‘Improving Bird’s Antidotes’, Australasian Journal of Philosophy, 81, 4 (2003): 573-580. Chomsky, N., Language and Mind (New York: Harcourt Brace, 1968). Churchland, P.S. and Sejnowski, T.J., The Computational Brain (Cambridge: MIT Press, 1992). Clark, R., ‘Concerning the Logic of Predicate Modifiers’, Noûs, 4 (1970): 311-335. Clifton, R. and Pagonis, C., ‘Unremarkable Contextualism: Dispositions in the Bohm theory, Foundations of Physics, 25, 2 (1995): 281-296. Comte, A., Discours sur l’esprit positif (Paris: Carilian-Gœury et Dalmont, 1844; Paris: Vrin, 1983). Daly, C., ‘Tropes’, in D.H. Mellor and A. Oliver (eds), Properties (Oxford: Oxford University Press, 1997). Darrigol, O., Electrodynamics from Ampère to Einstein (Oxford: Oxford University Press, 2000). David, M., ‘Kim’s Functionalism’, in J.E. Tomberlin (ed.), Philosophical Perspectives, 11, Mind, Causation, and World, (Oxford: Blackwell, 1997), pp. 133-148. Davidson, D., ‘Causal Relations’, in D. Davidson, Essays on Actions and Events (Oxford: Clarendon Press, 1980), pp. 149-162. Davidson, D., Essays on Actions and Events (Oxford: Clarendon Press, 1980). Davidson, D., ‘Actions, Reasons, and Causes’, in D. Davidson, Essays on Actions and Events, (Oxford: Clarendon Press, 1980), pp. 3-19. Descartes, R., Les Principes de la philosophie, trans. E.S. Haldane and G.R.T. Ross, The Principles of Philosophy, in The Philosophical Works of Descartes, vol. 1 (Cambridge: Cambridge University Press, 1912), pp. 201-302. Descartes, R., ‘Objections and Replies’, trans. E.S. Haldane and G.R.T. Ross, in The Philosophical Works of Descartes, vol. 2 (Cambridge: Cambridge University Press, 1912), pp. 1-344. Descartes, René, Oeuvres complètes, ed. C. Adam and J. Tannery, 11 vols, revised edition, Vrin, Paris, 1964-1974. Descartes, R., Philosophical Letters, 3 vols, trans. and ed. by A. Kenny (Oxford: Clarendon Press, 1970). Descartes, R., The Philosophical Writings of Descartes, trans. J. Cottingham, R. Stoothoff, D. Murdoch, 3 vols. (Cambridge: Cambridge University Press, 1984). Descombes, V., Les Institutions du sens (Paris: Editions de minuit, 1996).

Bibliography

287

Descombes, V., La denrée mentale (Paris: Minuit, 1995), trans. The Mind’s Provisions (Princeton: Princeton University Press, 2001). Descombes, V., ‘L’action’, in D. Kambouchner, Notions de Philosophie II (Paris: Folio, Gallimard, 1996). Dieks, D., Studies in the Foundations of Physics, PhD thesis (Utrecht University, 1981). Dijksterhuis, E.J., The Mechanization of the World (Oxford: Clarendon Press, 1961; Princeton: Princeton University Press, 1985). Dorato, M., Time and Reality. Spacetime Physics and the Objectivity of Temporal Becoming (Bologna: Clueb, 1995). Dorato, M., ‘Properties and Dispositions: Some Metaphysical Remarks on Quantum Ontolgy’ in A. Bassi, D. Dürr, T. Weber, N. Zanghì (eds), Quantum Mechanics, AIP Conference Proceedings, 884 (2006): 139-157. Dretske, F., ‘Laws of Nature’, Philosophy of Science, 44 (1977): 248-268. Drewery, A., ‘Dispositions and Ceteris Paribus Laws’, British Journal for the Philosophy of Science, 52 (2001): 723-733. Ducasse, C.J., ‘On the Nature and Observability of the Causal Relation’, (1926), in E. Sosa and M. Tooley (eds), Causation (Oxford: Oxford University Press, 1993), pp. 125-136. Dugas, R., La Mécanique au 17ème siècle (Paris: Dunod, 1954). Earman, J. ‘In Defence of Laws: Reflections on Bas van Fraassen’s Laws and Symmetries’, Philosophy and Phenomenological Research, 53 (1993): 413-419. Earman, J. and Roberts, J., ‘Ceteris Paribus, there is no Problem of Provisos’ Synthese, 118 (1999): 439-478. Earman, J., Roberts, J. and Smith, S., ‘“Ceteris Paribus” Lost’, Erkenntnis 57, 3 (2002): 281-301. Ellis, B.D., ‘Causal Powers and Laws of Nature’, in H. Sankey (ed.), Causation and Laws of Nature (Dordrecht: Kluwer, 1999), pp. 21-42. Ellis, B.D., ‘The New Essentialism and the Scientific Image of Mankind’, Epistemologia, 23 (2000): 189-210. Ellis, B.D., Scientific Essentialism (Cambridge: Cambridge University Press, 2001). Ellis, B.D., The Philosophy of Nature: A Guide to the New Essentialism (Chesham: Acumen, 2002). Ellis, B.D., ‘The New Essentialism and the Scientific Image of Mankind’, (NESIM) http://www.rescogitans.it/ita/scenari/tilgher/Ellis.htm. Ellis, B.D. and Lierse, C.E., ‘Dispositional Essentialism’, Australasian Journal of Philosophy, 72 (1994): 27-45. Everett, H., ‘Relative State Formulation of Quantum Mechanics’, Review of Modern Physics, 29 (1957): 454-462. Fales, E., ‘Essentialism and the Elementary Constituents of Matter’, Midwest Studies in Philosophy, 11 (1986): 391-402. Fales, E., Causation and Universals (London: Routledge, 1990).

288

Dispositions and Causal Powers

Faraday, M., Experimental Researches in Electricity (London: The Royal Society, 1839-55, repr. Santa Fe (N.M.): Green Lion Press, 2000). Faye, J., Niels Bohr: His Heritage and Legacy (Dordrecht: Kluwer, 1991). Fisk, M., ‘Capacities and Natures’, in R. Tuomela (ed.), Dispositions (Dordrecht: Reidel, 1978), pp. 189-210. Frege, G., ‘Über die Grundlagen der Geometrie’, Jahresbericht der Deutschen Mathematiker-Vereinigung, vol. 15 (1906): 379-380. Friedman M., ‘Explanation and Scientific Understanding’, Journal of Philosophy, 71 (1974): 5-19. Funkenstein, A., Theology and the Scientific Imagination (Princeton: Princeton University Press, 1986). Galilei, Galileo, ‘Discorsi e dimostrazioni matematiche intorno a due nuove scienze’ (1638), in Le Opere di Galileo Galilei, Edizione Nazionale (Firenze: G. Barbera, 1898), vol. 8; trans. H. Crew and A. de Salvio: Dialogues Concerning Two New Sciences (New York: McGraw-Hill, 1954). Gallistel, C.R., The Organization of Learning (Cambridge: Mass., MIT Press, 1990). Geach, P.T. and Anscombe, G.E.M., Three Philosophers: Aristotle, Aquinas, Frege (Oxford: Blackwell, 1963). Ghirardi, G.C., Rimini, A. and Weber, T., ‘Unified Dynamics for Microscopic and Macroscopic Systems’, Physical Review, 34 (1986): 470-491. Ghirardi, G.C., Un’occhiata alle carte di Dio (Milano: Il Saggiatore, 1997). Ghirardi, G.C., ‘Collapse Theories’, in Stanford Encyclopedia of Philosophy, plato. stanford.edu/entries/qm-collapse (2002). Gibbs, J.W., Elementary Principles in Statistical Mechanics, (New York, Scribner’s, 1902; and New York: Dover Publications, 1965). Gibson, J.J., The Ecological Approach to Visual Perception (New York: Howard Mifflin, 1979). Gilbert, W., De Magnete (London: Peter Short, 1600). Gilson, E., Etudes sur le rôle de la pensée médiévale dans la formation du système cartésien (Paris: Vrin, 1967). Goldman, A., A Theory of Human Action (Englewood Cliffs, N.J.: Prentice Hall, 1970). Goldstein, S., ‘Bohmian Mechanics’, in Stanford Encyclopedia of Philosophy, plato. stanford.edu/entries/qm-bohm (2001). Goodman, N. (1955), Fact, Fiction and Forecast (Indianapolis: Bobbs-Merrill, 1973). Grajewski, M.J., The Formal Distinction of Duns Scotus (Washington: The Catholic University of America Press, 1944). Guéroult, M., Leibniz. Dynamique et métaphysique (Paris: Aubier-Montaigne, 1967). Gunderson, L.B., ‘Bird on Dispositions and Antidotes’, Philosophical Quarterly, 50 (2000): 227-229. Hacker, P.M.S., ‘Events and Objects in Space and Time’, Mind, 91 (1982): 1-19.

Bibliography

289

Hampshire, S. ‘Dispositions’, in S. Hampshire, Freedom of Mind and Other Essays (Oxford: Clarendon Press, 1972), pp. 37-38. Hanson, S.O., ‘The Emperor’s New Clothes’, in G. Crocco, L. Fariñas de Cerro and A. Herszig (eds), Conditionals: from Philosophy to Computer Science (Oxford: Clarendon Press, 1995), pp. 13-31. Harper, W.L., Stalnaker, R. and Pearce, G. (eds), Ifs (Dordrecht: D. Reidel, 1981). Harré, R., ‘Powers’, British Journal for the Philosophy of Science, 21 (1970): 81101. Harré, R., The Principles of Scientific Thinking (London: MacMillan, 1970). Harré, R. and Madden, E.H., ‘Natural Powers and Powerful Natures’, Philosophy 48 (1973): 209-230. Harré R. and Madden, E.H., Causal Powers (Oxford: Blackwell, 1975). Harré, R., Varieties of Realism (Oxford: Blackwell, 1986). Harré, R., Laws of Nature (London: Duckworth, 1993). Harré, R., ‘Is There a Basic Ontology for the Physical Sciences?’, Dialectica, 51 (1997): 17-34. Hawkins, R.D. and Kandel, E.R., ‘Is There a Cell-Biological Alphabet for Elementary Forms of Learning?’, Psychological Review, 91 (1984): 375-391. Heil, J., Philosophy of Mind. A Contemporary Introduction (London: Routledge, 1998). Heisenberg, W., Physics and Philosophy: The Revolution in Modern Science (New York: Harper and Row, 1958). Hempel, C.G. and Oppenheim, P. (1948), ‘Studies in the Logic of Explanation’, in C.G. Hempel, Aspects of Scientific Explanation and other Essays in the Philosophy of Science (New York: The Free Press, 1965). Hempel, C.G., ‘Carnap and the Philosophy of Science’, in P.A Schilpp, The Philosophy of R. Carnap (La Salle, Ill.: Open Court, 1963), p. 694. Hempel, C.G., Philosophy of Natural Science (Englewood Cliffs, N.J.: Prentice Hall, 1966). Hempel, C.G., ‘Provisos: A Problem Concerning the Inferential Function of Scientific Theories’, in A. Grünbaum and W.C. Salmon (eds), The Limitations of Deductivism (Berkeley: University of California Press, 1988), pp. 19-36. Henninger, M.G., Relations (Oxford: Clarendon Press, 1989). Hintikka, J., Time and Necessity (Oxford: Clarendon Press, 1973). Holt, P.J., ‘Causality and Our Conception of Matter’, Analysis, 37 (1976): 20-29. Holton, R., ‘Dispositions All The Way Round’, Analysis, 59 (1999): 9-14. Hoover, K., Causality in Macroeconomics (Cambridge: Cambridge University Press, 2001). Hume, D., A Treatise of Human Nature (1739-40), L.A Selby-Bigge (ed.), (Oxford: Clarendon Press, 1888; new edn. by L.A. Selby-Bigge and P.H. Nidditch, 1955). Hume, D., ‘An Enquiry Concerning Human Understanding’ (1748), in Enquiries Concerning Human Understanding and Concerning the Principles of Morals, L.A. Selby-Bigge and P.H. Nidditch (eds), 3rd edition (Oxford: Clarendon Press, 1975).

290

Dispositions and Causal Powers

Hutchison, K., ‘Dormitive Virtues, Scholastic Qualities, and the New Philosophies’, History of Science, vol. 29, no. 85 (1991): 245-278. Hüttemann, A., ‘Laws and Dispositions’, Philosophy of Science 65 (1998): 121135. Ishiguro, H., Leibniz’s Philosophy of Language, 2nd edition (Cambridge: Cambridge University Press, 1990). Jackson, F., Pargetter, R. and Prior, E.W., ‘Functionalism and Type-Type Identity Theories’, Philosophical Studies, 42 (1982): 209-225. Jackson, F. and Pettit, P., ‘Functionalism and Broad Content’, Mind, 97 (1988): 381400. Jackson, F. and Pettit, P., ‘Program Explanation: A General Perspective’, Analysis, 50 (1990): 107-117. Jackson, F. and Pettit, P., ‘Causation in the Philosophy of Mind’, Philosophy and Phenomenological Research, 50, Suppl. (1990): 195-214. Jackson, F., From Metaphysics to Ethics: A Defence of Conceptual Analysis (Oxford: Clarendon Press, 1998). Jammer, M., The Philosophy of Quantum Mechanics (New York: Wiley, 1974). Jansen, L., Tun und Können. Ein systematischer Kommentar zu Aristoteles’ Theorie der Vermögen im neunten Buch der Metaphysik (Frankfurt-London-Miami-New York, Hänsel-Hohenhausen, 2002). Jansen, L. and Strobach, N. (2003), ‘The so-called materially valid inferences and the logic of concepts’, in B. Löwe, W. Malzkorn, T. Räsch (eds.), Foundations of the Formal Sciences II. Applications of Mathematical Logic in Philosophy and Linguistics, Rheinische Friedrich-Wilhelms-Universität Bonn, November 10-13, 2000, Trends in Logic, vol. 17 (Dordrecht: Kluwer, 2003), pp. 113-118. Jansen, L., ‘Dispositionen und ihre Realität’, in C. Halbig and C. Suhm (eds), Was ist wirklich? Neuere Beiträge zu Realismusdebatten in der Philosophie (Frankfurt a.M.: Ontos-Verlag, 2004). Johnston, M., ‘How to Speak of the Colors’, Philosophical Studies, 68 (1992): 221263. Joseph, G., ‘The Many Sciences and the One World’, Journal of Philosophy, vol. 77 (1980): 773-791. Kaila, E., R.S. Cohen (ed.), Reality and Experience: Four Philosophical Essays (Dordrecht: D. Reidel, 1979). Kaila, E., ‘Der logistische Neupositivismus’, (1930), trans. in E. Kaila, Reality and Experience, Vienna Circle Collection, vol. 12 (Dordrecht: D. Reidel, 1979). Kandel, E., ‘Cellular Mechanisms of Learning and Memory’, in E.R. Kandel, J.H. Schwartz and T.M. Jessell (eds), Essentials of Neural Science and Behavior (London: Prentice Hall, 1995), pp. 667-694. Kant, I., ‘Metaphysische Anfangsgründe der Naturwissenschaft’, (1786), in I. Kant, Werke in zwölf Bänden (Frankfurt: Suhrkamp, 1968), vol. IX, pp. 7-135; trans. J. Ellington, The Metaphysical Foundations of Natural Science (Indianapolis: Bobbs-Merrill, 1970). Kenny, A.J.P., Will, Freedom and Power (Oxford: Oxford University Press, 1975).

Bibliography

291

Kim, J., Supervenience and Mind (Cambridge: Cambridge University Press, 1993). Kim, J., ‘Causes and Counterfactuals’, Journal of Philosophy, 70 (1973): 570-572, reprinted in E. Sosa and M. Tooley (eds), Causation (Oxford, 1993), pp. 205207. Kim, J., ‘Does the Problem of Mental Causation Generalize?’, Proceedings of the Aristotelian Society, 97 (1997): 281-297. Kim, J., Mind in a Physical World (Cambridge: MIT Press, 1998). Kistler, M., ‘Causes as Events and Facts’, Dialectica, 53 (1999): 25-46. Kistler, M., Causalité et lois de la nature (Paris: Vrin, 1999). Kistler, M., ‘The Causal Criterion of Reality and the Necessity of Laws of Nature’, Metaphysica, 3 (2002): 57-86. Kistler, M., ‘Causation in contemporary analytical philosophy’, Quaestio 2 (2002): 635-668. Kistler, M., ‘Erklärung und Kausalität’, Philosophia naturalis, 39, Heft 1 (2002): 89-109. Kistler, M., ‘Laws of Nature, Exceptions and Tropes’, Philosophia Scientiae, 7 (2003): 189-219. Kistler, M., ‘Some Problems for Lowe’s Four-Category Ontology’, Analysis, 64.2 (2004): 146-151. Kistler, M., ‘Necessary Laws’, in J. Faye, P. Needham, U. Scheffler and M. Urchs (eds), Nature’s Principles (Dordrecht: Springer, 2005), pp. 201-227. Kitcher, P., ‘Explanation, Conjunction and Unification’, Journal of Philosophy, 73 (1976): 207-212. Kitcher, P., ‘1953 and All That, A Tale of Two Sciences’, Philosophical Review, 93 (1984): 335-373. Kitcher, P., ‘Explanatory Unification and the Causal Structure of the World’, in P. Kitcher and W. Salmon (eds), Scientific Explanation (Minneapolis: University of Minnesota, 1989), pp. 410-505. Knuutila, S., Modalities in Medieval Philosophy (London: Routledge, 1993). Kochen, S. and Specker E., ‘The Problem of Hidden Variables in Quantum Mechanics’, Journal of Mathematics and Mechanics, 18 (1967): 1015-1021. Kripke, S.A., ‘Identity and Necessity’, in M.K. Munitz (ed.), Identity and Individuation (New York: New York University Press, 1971), pp. 135-164. Kripke, S.A., ‘Naming and Necessity’, in D. Davidson and G. Harman (eds), Semantics of Natural Language (Dordrecht: Reidel, 1972), pp. 253-355, and pp. 763-769; repr. as monograph: S.A. Kripke, Naming and Necessity (Cambridge: Harvard University Press, 1980). Laudisa, F., Correlazioni pericolose (Padova: Il Poligrafo, 1998). Lebowitz, J.L., ‘Macroscopic Laws, Microscopic Dynamics, Time’s Arrow and Boltzmann’s Entropy’, Physica A, 194 (1993): 1-27. Lehrer, K. (ed.), Freedom and Determinism (New York: Random House, 1966). Leibniz, G.W., Philosophical Papers and Letters, L.E. Loemker (ed.) (Chicago: Free Press, 1956).

292

Dispositions and Causal Powers

Leibniz, G.W., Die philosophischen Schriften von Gottfried Wilhelm Leibniz, 7 vols., ed. C.J. Gerhardt (Berlin: Weidmannsche Buchhandlung, 1875-1890). Leibniz, G.W., New Essays on Human Understanding, transl. and ed. P. Remnant and J. Bennett (Cambridge: Cambridge University Press, 1981). Leibniz, G.W., Philosophical Texts, trans. R. Francks and R.S. Woolhouse, (Oxford: Oxford University Press, 1998). Levesque, D. and Verlet, L., ‘Computer “Experiments” on Classical Fluids. III. TimeDependent Self-Correlation Functions’, Physical Review A, 2 (1970): 2514. Lewis, D., ‘Psychophysical and Theoretical Identifications’, Australasian Journal of Philosophy, 50 (1972): 249-258. Lewis, D., Counterfactuals (Oxford: Blackwell, 1973). Lewis, D., ‘Causation’, Journal of Philosophy, 70 (1973): 556-567, reprinted in his Philosophical Papers, vol. 2 (Oxford, 1986), pp. 159-172. Lewis, D., On the Plurality of Worlds (Oxford: Blackwell, 1986). Lewis, D., Philosophical Papers, vol. 2 (Oxford: Oxford University Press, 1986). Lewis, D., ‘Finkish Dispositions’, Philosophical Quarterly, 47 (1997): 143-158. Lewis, D., Papers in Metaphysics and Epistemology (Cambridge: Cambridge University Press, 1999). Lewis, D., ‘Humean Supervenience Debugged’, Mind 1994, repr. in his Papers in Metaphysics and Epistemology 1999, pp. 224-247. Lewis, D., ‘Finkish Dispositions’, in H. Sankey (ed.), Causation and Laws of Nature (Dordrecht: Kluwer Academic Publishers, 1999), pp. 143-155. Lewis, D. and Langton, R., ‘Defining “Intrinsic”’, in D. Lewis, Papers in Metaphysics and Epistemology 1999, pp. 116-32. Lipton, P., ‘All Else Being Equal’, Philosophy, 74 (1999): 155-168. Locke, J., An Essay Concerning Human Understanding (1689), ed. P.H. Nidditch (Oxford: Clarendon Press, 1975). Lowe, E.J., Kinds of Being: A Study of Individuation, Identity and the Logic of Sortal Terms (Oxford: Blackwell, 1989). Lowe, E.J., ‘Dispositions and Laws’, Metaphysica, 2 (2001): 5-23. Lowe, E.J., A Survey of Metaphysics (Oxford: Oxford University Press, 2002). Lowe, E.J., ‘A Defence of the Four-Category Ontology’, in C. Moulines and K.G. Niebergall (eds), Argument und Analyse. Akten des vierten Internationalen Kongresses der Gesellschaft für Analytische Philosophie (Paderborn: Mentis, 2002). Lowe, E.., ‘Recent Advances in Metaphysics’, Facta philosophica 5 (2003): 3-24. Macdonald, C., ‘Tropes and Other Things’, in S. Laurence and C. Macdonald (eds), Contemporary Readings in the Foundations of Metaphysics (Oxford: Blackwell, 1998), pp. 329-350. Mach, E. The Analysis of Sensations, trans. T.J. McCormack (Chicago: Open Court, 1914). Machiavelli, N., The Prince, ed. Q. Skinner and R. Price (Cambridge: Cambridge University Press 1997). Mackie, J.L., ‘Dispositions, Grounds, and Causes’, Synthese, 34 (1977): 361-370.

Bibliography

293

Mackie J.L., ‘Causes and Conditions’, in E. Sosa and M. Tooley (eds), Causation (Oxford: University Press, 1993). McKitrick, J., ‘The Bare Metaphysical Possibility of Bare Dispositions’, Philosophy and Phenomenological Research, 64, no. 2 (2003): 349-369. McLaughlin, B., ‘Disposition’, in E. Sosa and J. Kim (eds), A Companion to Metaphysics (Oxford: Blackwell, 1995), pp. 121-124. McTaggart J., ‘The Unreality of Time’, Mind, 68, vol. 17 (1908): 457-474. Malebranche, N., ‘Eclaircissements sur la recherche de la vérité’ (1681), in N. Malebranche, Oeuvres, vol. I. G. Rodis-Lewis (ed.), bibliothèque de la Pléïade, (Paris: Gallimard, 1979), pp. 789-1126, trans. T.M. Lennon and P.J. Olscamp, ‘Elucidations of The Search After Truth’, in The Search After Truth, Elucidations of The Search After Truth (Columbus: Ohio State University Press, 1980), pp. 531-753. Malzkorn, W., ‘Realism, Functionalism and the Conditional Analysis of Dispositions’, The Philosophical Quarterly, 50 (2000): 452-469. Mansion, A., Introduction à la physique aristotélicienne, Louvain-la-Neuve, 1987. Margenau, H., ‘Advantages and Disadvantages of Various Interpretations of the Quantum Theory’, Physics Today, 7 (1954): 6-13. Mansion, A., Introduction à la physique aristotélicienne (Louvain-la-Neuve, 1987). Martin, C.B., ‘Dispositions and Conditionals’, Philosophical Quarterly, 44 (1994): 1-8. Martin, C.B., ‘Final Replies to Place and Armstrong’, in D.M. Armstrong, C.B. Martin and U.T. Place, ed. T. Crane, Dispositions: A Debate (London: Routledge, 1996). Martin, C.B. and Heil, J., ‘Rules and Powers’, Philosophical Perspectives, 12 (1998): 283-312. Maxwell N., ‘Quantum Propensiton Theory: A Testable Resolution of the Wave/ Particle Dilemma’, British Journal for the Philosophy of Science, 39 (1988): 1-50. Mellor, D.H., ‘In Defence of Dispositions’, Philosophical Review, 83 (1974): 157181. Mellor, D.H., ‘Counting Corners Correctly’, Analysis, 42 (1974): 96-97. Mellor, D.H., Matters of Metaphysics (Cambridge, Cambridge University Press, 1991). Mellor, D.H. and Oliver, A. (eds), Properties (Oxford: Oxford University Press, 1997). Mellor, D.H., ‘The Semantics and Ontology of Dispositions’, Mind, 109 (2000): 757-780. Menzies, P., ‘Against Causal Reductionism’, Mind, 97 (1988): 551-574. Menzies, P., ‘Capacities, Natures and Pluralism: A New Metaphysics for Science?’, Philosophical Books, 43 (2002): 261-270. Mermin N.D., ‘What is quantum mechanics trying to tell us?’, American Journal of Physics, 66 (1998): 753-767. Mill, J.S., A System of Logic, (London: 1949).

294

Dispositions and Causal Powers

Molière, J.-B. P. de, Le malade imaginaire (1673), in Molière, G. Couton (ed.), Oeuvres completes (Paris: Gallimard (La Pléiade), 1971), vol. II, pp. 1091-1178; trans. A.R. Waller, ‘The imaginary invalid’, in The Plays of Molière, 8 vols (Edinburgh: John Grant, 1926). Molnar, G., ‘Are Dispositions Reducible?’, Philosophical Quarterly, 49 (1999): 117. Molnar, G., Powers: A Study in Metaphysics, ed. S. Mumford, (Oxford: Oxford University Press, 2003). Morange, M., La part des gènes (Paris: O. Jacob, 1998); trans. M. Cobb, The Misunderstood Gene (Cambridge: Harvard University Press, 2001). Mumford, S., ‘Ellis and Lierse on Dispositional Essentialism’, Australasian Journal of Philosophy, 73 (1955): 606-612. Mumford, S., Dispositions (Oxford: Oxford University Press, 1998). Mumford, S., ‘Realism and the Conditional Analysis of Dispositions: Reply to Malzkorn’, Philosophical Quarterly, 51 (2001): 375-378. Mumford, S., Laws in Nature (London: Routledge, 2004). Mumford S., ‘L’état des lois’, in J.-M. Monnoyer (ed.), La Structure du Monde: Objets, Propriétés, États de Choses. Renouveau de la Métaphysique dans l’école Australienne de Philosophie, no. Hors Série de Recherches sur la Philosophie et le Langage (Paris: J. Vrin, 2004), pp. 235-262. Mumford, S., ‘The Ungrounded Argument’, Synthese 149 (3) (2006): 471-489. Murdoch D., Niels Bohr’s Philosophy of Physics (Cambridge: Cambridge University Press, 1987). Newton, Sir I., Opticks, or Treatise of the Reflections, Refractions, Inflections and Colours of Light, repr. from the 4th edn of 1730, ed. E.T. Whittaker (New York: Whittlesey House, McGraw-Hill, 1931). Ockham, W. of, ‘Summa logicae’, in Opera philosophica, vol. I, ed. P. Boehner, G. Gal and S. Brown, The Franciscan Institute of St. Bonaventure (New York: University St. Bonaventure, 1980). Ockham, W. of, Summa logicae, 2 vols., transl. M.J. Loux (South Bend, Ind.: St. Augustine’s Press, 1998). Ockham, W. of, ‘Quodlibeta septem’, I, in Opera theologica, vol. IX (The Franciscan Institute of St. Bonaventure University: St. Bonaventure, New York, 1980). Ockham, W. of, Quodlibetal Questions, 2 vols. Transl. Alfred J. Freddoso (New Haven: Yale University Press, 1991). Oliver, A., ‘The Metaphysics of Properties’, Mind, 105 (1996): 1-80. Ostler, G. (ed), The Little Oxford Dictionary of Current English, 6th edn, J. Swannell (Oxford: Clarendon Press, 1986). Pap, A., ‘Disposition concepts and extensional logic’, Minnesota Studies in the Philosophy of Science, vol. II (1958): 196-224. Parsons, T., Events in the Semantics of English (Cambridge: MIT Press, 1990). Peirce C.S., The Collected Papers of C.S. Peirce, 8 vols (Cambridge: Harvard University Press, 1939-1958).

Bibliography

295

Pietroski, P. and Rey, G., ‘When Other Things Aren’t Equal: Saving Ceteris Paribus Laws from Vacuity’, British Journal for the Philosophy of Science, 46 (1995): 81-110. Place, U.T., ‘Intentionality as the Mark of the Dispositional’, Dialectica, 50 (1996): 91-120. Place, U.T., ‘Intentionality and the Physical: a Reply to Mumford’, Philosophical Quarterly, 49 (1999): 225-31. Place, U.T., ‘Dispositions as Intentional States’, in D. Armstrong, C.B. Martin. and U.T. Place, ed. T. Crane, Dispositions. A Debate (London: Routledge, 1996), pp. 19-32. Plato, ‘Sophist’, in E. Hamilton and H. Cairns (eds), The Collected Dialogues of Plato (Princeton: Princeton University Press, 1961), pp. 957-1017. Popper, K., ‘The Propensity Interpretation of the Calculus of Probability, and the Quantum Theory’, in S. Körner (ed.), Observation and Interpretation (London: Butterworth, 1957), pp. 65-70. Popper, K., Logik der Forschung (Vienna: Springer, 1935; 3rd augmented edition Tübingen: J.C.B. Mohr, 1969). Popper K., Quantum Theory and the Schism in Physics (London: Hutchinson, 1982). Popper, K., A World of Propensities (Bristol: Thoemmes, 1990). Price, H.H., Thinking and Experience (London: Hutchinson, 1953). Prior, E., Pargetter, R. and Jackson, F., ‘Three Theses about Dispositions’, American Philosophical Quarterly, 19 (1982): 251-257. Prior, E., ‘The Dispositional/Categorical Distinction’, Analysis, 42 (1982): 93-96. Prior, E., Dispositions (Aberdeen: Aberdeen University Press, 1985). Psillos, S., ‘Salt Dissolves in Water, but not Necessarily’, Analysis, 62 (2002): 255257. Putnam, H., ‘The Nature of Mental States’, (1967), reprinted in H. Putnam, Mind, Language, and Reality: Philosophical Papers, vol. 2 (Cambridge: Cambridge University Press, 1975). Quine, W.V., The Roots of Reference (La Salle, IL.: Open Court, 1971). Quine, W.V., The Ways of Paradox and Other Essays revd. edn, (Cambridge, Mass., Harvard University Press, 1966). Redhead, M., Incompleteness, Non-locality and Realism (Oxford: Clarendon Press, 1987). Reid, T., Essays on the Active Powers of the Human Mind (1788) (Cambridge: MIT Press, 1969). Rescorla, R.A., ‘Pavlovian Conditioning. It’s Not What You Think It Is’, American Psychologist, 43 (1988): 151-160. Robinson, H., Matter and Sense (Cambridge: Cambridge University Press, 1982). Rodriguez-Pereyra, G., Resemblance Nominalism – A Solution to the Problem of Universals (Oxford: Oxford University Press, 2002). Rosenberg, A., The Structure of Biological Science (Cambridge: Cambridge University Press, 1985).

296

Dispositions and Causal Powers

Rosenthal, J., Wahrscheinlichkeiten als Tendenzen. Eine Untersuchung objektiver Wahrscheinlichkeitsbegriffe (Paderborn: Mentis, 2004). Rovelli, C., ‘Relational quantum mechanics’, International Journal of Theoretical Physics, 35 (1996): 1637-1678. Rovelli, C. and Laudisa, F., ‘Relational Quantum Mechanics’, in Stanford Encyclopedia of Philosophy, http://plato.stanford.edu/entries/qm-relational/ (2002). Rozemond, M., Descartes’s Dualism (Cambridge: Harvard University Press, 1998). Russell, B., ‘Les paradoxes de la logique’, Revue de Métaphysique et de Morale, 14 (1906): 294-317. Russell, B., ‘On the Relation of Universals and Particulars’, Proceedings of the Aristotelian Society, 12 (1911-1912): 1-24. Russell, B., The Problems of Philosophy (New York: H. Holt and Company, 1912). Russell, B., ‘On the Notion of Cause’, Proceedings of the Aristotelian Society, 13 (1912): 1-26; repr. in B. Russell, Mysticism and Logic (1917) (London: Routledge, 1986), pp. 171-196, and in The Collected Papers of Bertrand Russell, vol. 6, Logical and Philosophical Papers 1909-13, John G. Slater (ed.), (London and New York: Routledge, 1992), pp. 193-210. Ryle, G., The Concept of Mind (London: Hutchinson, 1949). Saunders, S., ‘Time, Quantum Mechanics and Decoherence’, Synthese, 102 (1995): 235-266. Schaffer, J., ‘The Metaphysics of Causation’, Stanford Encyclopedia of Philosophy (http://plato.stanford.edu/entries/causation-metaphysics/) (2003). Schilpp, P.A., The Philosophy of R. Carnap (La Salle: Open Court, 1963). Searle, J., The Construction of Social Reality (New York: Free Press, 1995). Shepard, R., ‘The Analysis of Proximities: Multidimensional Scaling with an Unknown Distance Function’, Psychometrika, 27 (1962): 125-140 and 219246. Shepard, R., ‘Approximation to Uniform Gradients of Generalization by Monotone Transformations of Scale’, in D.I. Mostofsky (ed.), Stimulus Generalization (Stanford: Stanford University Press, 1965), pp. 95-110. Shepard, R., ‘Representation of Structure in Similarity Data: Problems and Prospects’, Psychometrika, 39 (1974): 373-421. Shoemaker, S., ‘Causality and Properties’ (1980), repr. in S. Shoemaker, Identity, Cause and Mind (Cambridge: Cambridge University Press, 1984), pp. 206-233. Shoemaker, S., ‘Causality and Properties’, in P. Van Inwagen (ed.), Time and Cause (Dordrecht: Reidel, 1980), pp. 109-135, repr. in D.H. Mellor and A. Oliver (eds), Properties (Oxford: Oxford University Press, 1997), pp. 228-254. Sosa, E. and Tooley, M. (eds), Causation (Oxford: Oxford University Press, 1993). Spector, M., ‘Leibniz vs. the Cartesians on Motion and Force’, Studia Leibniziana, 7, no. 1 (1975): 135-144. Spohn, W., ‘Begründungen a priori-oder: ein frischer Blick auf Dispositionsprädikate’, in W. Lenzen (ed.), Das weite Spektrum der analytischen Philosophie. Festschrift für Franz von Kutschera (Berlin: de Gruyter, 1997), pp. 323-345.

Bibliography

297

Squires, R., ‘Are Dispositions Causes?’, Analysis, 29 (1968): 45-47. Squires E., ‘Many Views of One World’, European Journal of Physics, 8 (1987): 171-173. Stalnaker, R., ‘A Theory of Counterfactuals’, in N. Rescher (ed.), Studies in Logical Theory (Oxford: Blackwell, 1968). Steward, H., The Ontology of Mind (Oxford: University Press, 1997). Strawson, P.F., ‘Causality and Explanation’, in B. Vermazen and M. Hintikka (eds), Essays on Davidson: Actions and Events (Oxford: Clarendon Press, 1985), pp. 115-136. Suarez, F., ‘Disputationes Metaphysicae’ (1597), in Opera omnia, Editio nova, C. Berton, vol. XXV and XXVI, repr. of the Paris 1866 edition, (Hildesheim, Zürich, New York: Georg OlmsVerlag, 1998). Partial trans. Alfred J. Freddoso, On Efficient Causality, Metaphysical Disputations 17, 18, and 19 (New Haven: Yale University Press, 1994). Suarez, M., ‘Propensities in Quantum Mechanics’ in History and Philosophy of Modern Physics (forthcoming). Swinburne, R., ‘Reply to Shoemaker’, in L.J. Cohen and M. Hesse (eds), Aspects of Inductive Logic (Oxford, Oxford University Press, 1983), pp. 313-320. Taylor, R., Action and Purpose (Englewood Cliffs, NJ: Prentice Hall, 1966). Thompson, I.J., ‘Real Dispositions in the Physical World’, British Journal for the Philosophy of Science, 39 (1988): 67-79. Tiercelin, C., ‘L’influence scotiste dans le projet peircien d’une métaphysique scientifique’, Revue des Sciences Philosophiques et Théologiques, numéro spécial: Jean Duns Scot et la métaphysique classique, no.1, tome 83 (1999): 117134. Tiercelin, C., ‘C.S. Peirce et le projet d’une métaphysique scientifique évolutionnaire’, Publications de la Sorbonne, Philosophie, 5 (2000): 453-463. Tiercelin, C., ‘La métaphysique et l’analyse conceptuelle’, Revue de Métaphysique et de Morale, numéro spécial: Métaphysique et ontologie: perspectives contemporaines, no. 4 (2002): 559-584. Tiercelin, C., ‘Sur la réalité des propriétés dispositionnelles’, in Cahiers de l’Université de Caen, nos. 38-39 (2002): 127-157. Tiercelin, C., ‘Le problème des universaux: aperçus historiques et perspectives contemporaines’, in J.-M. Monnoyer (ed.), La Structure du monde: objets, propriétés, états de choses (Paris: Vrin, 2004), pp. 329-353. Tooley, M., Causation: A Realist Approach (Oxford: Clarendon Press, 1987). Tooley, M., ‘The Nature of Laws’, Canadian Journal of Philosophy, 7 (1977): 667698. Tuomela, R. (ed.), Dispositions (Dordrecht: Reidel, 1978). Vaidman L., ‘The Many-Worlds Interpretation of Quantum Mechanics’, in Stanford Encyclopedia of Philosophy, http://plato.stanford.edu/entries/qm-manyworlds (2002). Van Fraassen, B., The Scientific Image (Oxford: Clarendon Press, 1980).

298

Dispositions and Causal Powers

Vuillemin, J., Physique et métaphysique kantiennes (Paris: Presses Universitaires de France, 1955). Vuillemin, J., Nécessité ou contingence (Paris: Editions de minuit, 1984). Watkins, E., Kant and the Metaphysics of Causality (Cambridge: Cambridge University Press, 2005). Whitaker, A., Einstein, Bohr and the Quantum Dilemma (Cambridge: Cambridge University Press, 1996). Wiggins, D., Sameness and Substance (Oxford: Blackwell, 1980). Wittgenstein, L., Philosophische Untersuchungen, Philosophical Investigations, bilingual edition in German and English; English transl. G.E.M. Anscombe (Oxford: Blackwell, 1953). Yakira, E., La causalité de Galilée à Kant (Paris: Presses Universitaires de France, 1994).

Index

accident, real, 2, 4, 7-14 act of measurement, see measurement, act of action, 5-6, 11, 14, 19, 30, 96, 111-2, 115n, 128, 133, 137, 140-50, 154n, 157, 161-3, 176 actualism, 91, 92n affordance, 90, 191 aggregate, 193, 279, 281-2 analytic, analyticity, 25n, 31-2, 46, 48-52, 55, 59, 65-6, 74, 109, 114, 119-20, 239, 245, 255n Anscombe, 136n, 139n, 152, 154-5, 157 anti-realism, 30 anthropomorphism, 11, 13, 133, 137, 150 antidote, 37-8, 94, 110, 113, 156n Aristotle, 3-7, 9-10, 17, 116n, 141n, 164, 167, 168n, 173n, 196 Armstrong, D.M., 27-8, 30, 67, 76-7, 83n, 88n, 90-1, 92n, 94, 103-5, 108-9, 112, 116, 120-2, 124n, 125, 134n, 146n, 250-1 Attribution of disposition, see disposition, attribution of Avicenna, 95 Bacon, Francis, 10 basis categorical, 28n, 30, 67, 70, 83, 88, 97, 104n, 110, 112, 124, 145, 269 causal, 67, 110-1, 124n, 125, 153 microscopic, 32, 104-5, 111-2, 121, 1236, 129-30, 132, 278-9 supervenience, 72-7, 79 Berkeley, 20, 192 Bhaskar, Roy, 183 Bird, Alexander, 91n, 99n, 100, 110, 115n, 156n, 240-2 Blackburn, Simon, 22, 67, 70, 82 Boscovich, Roger Joseph, 18, 37, 182, 189, 194 Bohm, David, 249, 250n, 251, 256, 259, 260n, 261, 266-70

Bohr, Niels, 190-4, 251, 256-61, 267 Boyle, 10-7, 19, 21n, 81, 189, 253 bridge principle, 203, 204 Bridgman, P.W., 45 Campbell, Keith, 76 capability, 186-7 capacity, 2-3, 5, 10, 12n, 14, 16-8, 21n, 96, 123, 140, 146-7, 149-50, 154, 156-8, 168, 187, 195-203, 205, 209, 22636, 238-9, 242-7, 272-3, 275 Carnap, Rudolf, 22n, 23-5, 43-57, 59-60, 65-6 Cartwright, Nancy, 24n, 81, 91, 165n, 166-7, 186, 204n, 208-9, 212n, 223, 225-6, 229-39, 242-3, 244n, 245, 246n, 262n categorical predicate/property, 23-4, 25n, 26-32, 37-40, 67, 70, 77, 83, 86-9, 94-6, 97n, 103-5, 106n, 111-2, 114, 116-25, 129-32, 145, 146n, 168n, 187, 195-9, 201, 209, 236, 246n, 250-6, 258-63, 265-6, 268-70, 280 categoricalism, 27-8 categoricity, 182, 184, 255 causality, causation, 1,3-4, 19, 20n, 21-2, 33, 71, 91n, 92n, 109, 138, 140, 143, 145, 185, 196, 207, 213, 216-9, 280 causal criterion of reality, 28, 31, 86 causal role, 32, 75, 84, 88, 95, 106, 108, 255n causation, substance, 196 causation, event, 140, 143, 196 ceteris paribus, 74, 91, 110, 113, 114n, 1156, 144, 183, 185, 194, 201, 208-9, 221-6, 229, 231, 235-9, 241-7 change (aristotelian), 2-6, 13, 14 Chisholm, Roderick, 47, 51, 57, 59, 65 Clifton, Robert, 249, 255, 261, 267-9 colour, 12, 15, 25n, 32, 104n, 126-129 Comte, Auguste, 20, 22 conditional analysis of dispositions, 34, 87n,

300

Dispositions and Causal Powers

153, 154, 156-159, 237, 240 conditionality, 182, 184, 196, 200-2 conditions/circumstances ideal, 35-39, 74, 163n, 208-213, 216, 218, 237, 241, normal, 109, 162, 240-1 confirmable, 45, 52-3, 54n, 56, 66 connection, necessary, see necessary connection contextualism, 255, 259-60, 267-8, 270 contingent, contingency, 56, 70, 72, 85, 87n, 93, 96, 98-100, 101n, 109-11, 11920, 121n, 122n, 280 counterfactual conditional, 25n, 26-8, 31-2, 38-9, 47, 56-63, 65-6, 70, 78-9, 93n, 103, 110n, 115, 119-20, 134n, 145, 148, 152-3, 184, 207-8, 210, 213-9, 223, 236, 240, 252, 280 counterfactualism, 60 Davidson, Donald, 23n, 30, 107, 112n, 138n, 143, 154-5 definition (of dispositional predicates), 45, 47-9, 51-7, 58n, 60, 62, 65-6, 84, 89n, 240, 271-2, 275, 278 definition in use, 43, 47, 49 dependency, ontological, 11, 83, 168 Descartes, 2, 9, 10-4, 17-20, 84, 141n, 253 description, categorical, 201 disposition/power active/passive, 3, 14, 16-7, 21, 33, 36, 77-8, 84, 133, 135, 144n, 145-7, 149-50, 174, 186-7, 233 attribution of, 7, 27-9, 31, 108-9, 110n, 115-6, 117n, 119-20, 121n, 134, 154n, 158n, 169n, 182, 184-5, 188, 194, 231n, 232 Bohmian, 267-70 determinable, 200 finkish, see fink irreducible, 68, 71, 250, 256-7, 259-61, 264, 270 observable, 190 probabilistic 34, 172-3, 175-7, 259, 265 scientific, 238, 244, 273 surefire, 172-6 ultimate/fundamental, 75, 89 ungrounded 67-8, 70-1, 73, 75, 77-80, 87n, 89, 100, 101n, 126n, 264,

266-7 dispositional aspect, 257 predicate/property, 15, 23-32, 68, 70, 92, 94, 104n, 105-107, 115, 117122, 126, 129, 237, 239, 252, 253, 266, 269, 275 dispositional-categorical distinction, 25n, 87, 94, 106n, 117-9, 131, 145, 168n dispositionalism, 28, 82, 89, 224, 226, 229, 231-2, 236, 238-9, 241-2, 246-7, 276n dispositionality (of a predicate/property), 23, 25, 93, 117n, 118, 120, 254, 259, 262, 267-8, 270 Dorato, Mauro, 192 Dretske, Fred, 90, 134n Duhem, Pierre, 276 Duns Scotus, 95, 96n dynamical conception of matter, 17, 19, 84, 192, 192, 194, 281 Earman, John, 201, 224n, 232, 234, 243 Efficacious, causal efficacy, 27-8, 30, 74, 86, 103-7, 110-2, 114, 116-7, 120-1, 129-32, 139, 185, 192 Einstein, 45, 182, 257 eliminable, ineliminable, 24n, 46, 50-1, 545, 66, 264 ellipitical (disposition ascription), 162-3, 165, 176 Ellis, Brian, 28, 67, 78, 82, 83n, 84n, 85-8, 90, 93-4, 98-9, 198 empiricism, 45-6, 54, 66, 137, 139n empiricism, logical, 22-3, 239-40 enhance, see weaken/enhance epiphenomenal, 104, 130, 131, 132, 145 epiphenomenalist, 110, 112, 132 epistemic, 24, 97, 116, 170, 204, 216, 219, 223, 254-5 essentialism (scientific, dispositional), 83, 84n, 85-6, 88, 93, 98-100 event (in causation), 1, 4, 20-3, 25, 35-6, 44-5, 70-1, 74, 78, 82-3, 86, 87n, 90-4, 96, 99, 107-8, 130, 138-41, 143, 146, 151, 158, 185, 192-3, 1967, 213-6, 218, 237, 241-2, 261-3, 268-9

Index Everett, Hugh, 262-4 exceptions (to laws of nature), 91, 110, 112, 114n, 135, 148-9, 171, 183, 221-2, 243n explanation (scientific, causal), 1-4, 11, 20, 22-4, 69, 89-90, 100, 106-8, 112-4, 123, 126, 128-9, 131, 133-8, 141, 143-4, 146-7, 150-1, 163, 176, 183, 188, 193-4, 209, 211-3, 238, 266n, 267, 269, 276 reductive, 69, 129, 131 factual consequences (of reduction sentences), 50-1, 55-6, 59, 61-2 falsification, 173, 222 family resemblance, 140, 184, 186 Faraday, 18, 37, 189 fink, finkish disposition, 94, 110, 153, 156n, 240 Frege, Gottlob, 43-6, 49, 51, 54, 55n, 56, 59, 65, 108n foundation, 95, 189, 251, 277 foundationalism, 87n form, substantial, 2, 4, 6-7, 9-10, 12-14, 17, 19, 96 function of disposition ascriptions, 161-2, 166-7, 252-3, 267, 269 functional role, 32, 77, 88-9, 121, 124n, 255n functional dependency, 36, 226 functionalist view of dispositions, 86, 87n, 93, 95, 105, 111-2 Galileo, 20, 22, 204, 207, 253 Ghirardi, Rimini, Weber (GRW), 256, 265-7 Gilbert, William, 17, 37, 182, 189 Goldman, Alvin, 152-4 Goodman, Nelson, 24-7, 31, 47, 51, 54, 57, 58n, 60, 65, 81, 118n, 145n grammar, 163, 190-1 Hampshire, Stuart, 196-201 Harré, Rom, 19n, 24n, 69, 81, 86n, 126n, 139n, 143, 185 heat, specific, 210, 212, 213, 219, 271-76 Heisenberg, 249, 258, 261, 267 Hempel, 22, 50n, 134n, 137, 142 hidden variables, 39, 256, 258-260, 267, Hume, 1, 3-4, 21-2, 31, 35-6, 73-4, 78, 91,

301

96, 135, 139, 188, 192, 241 Humean/non-Humean, 31-2, 36, 67-75, 77-80, 85, 87-8, 90-1, 96, 98, 101n, 134n, 136n, 151, 193, 223-4, 226, 236, 241-2 Hüttemann, Andreas, 223 idealization, 34-5, 91, 281 indeterminate, 31, 48n, 49, 54-5, 59-60, 66, 89, 94, 100n, 130, 196, 200 instantiation, 72, 75-6, 86, 87n, 92n, 97, 119n instrumentalism, 267 intentional, 71, 79 interference, 38, 114, 191, 199, 201, 203-5, 222, 224, 257-8 interpretation of quantum mechanics, 115n, 249-251, 256-7, 259-64, 266-7, 270 Jackson, Frank, 30, 32, 67, 100, 104-5, 119, 122n, 131 Jansen, Ludger, 191 Kant, 18, 37, 181-2, 189-91, 193-4, 196, 258-9 Kenny, A.J.P., 147n, 149n Kim, Jaegwon, 111-2, 125, 130-1 kind natural, 82, 83n, 85-6, 100, 181, 183-4, 187, 241-2 of substance, 83, 157, 169 Kistler, Max, 138n, 223 Kitcher, Philip, 216, 219, 266n Kripke, Saul, 100, 278 Laudisa, Federico, 261, 263 latency, 196 law of nature, 1, 11, 19-20, 22-4, 25n, 29, 34-6, 44-6, 49, 57, 59-60, 72-3, 81n, 82-5, 87n, 90-3, 96-100, 101n, 1134, 115n, 116-21, 122n, 123, 130, 133-8, 142-6, 151, 166, 170, 183, 195, 197, 201, 204, 207-13, 215-9, 221-36, 239, 241-7, 252, 271, 273, 275-6, 279-80 law ceteris paribus, 114, 201, 208-9, 221-7, 229, 231, 239, 241, 246 necessary, 100

302

Dispositions and Causal Powers

contingent, 280 Lehrer, Keith, 152-5 Leibniz, 17-9, 36n, 84, 184, 194 Lewis, David, 47, 58, 60, 65-6, 71-3, 76, 79, 90-1, 124n, 134, 153, 156n, 213-6, 218-9, 240 Lipton, Peter, 223, 231n, 237, 241-6 Locke, 10, 14-7, 19, 76, 84, 104n, 135, 253 Lowe, E.J., 76, 117n, 168n, 275 macroscopic, 19, 36, 104-5, 107, 109-112, 122-3, 125-6, 129-30, 132, 253, 255, 258n, 265-6, 268-9, 271-3, 275-9, 281-2 macrostructure, 188 Mach, Ernst, 43, 182-3, 188 McTaggart, John, 250 Malebranche, 3, 10n, 19 malleability, 197, 200-3 manifestation, 17, 70-1, 73-6, 78-9, 85, 88, 90n, 94, 103-4, 106, 108-16, 117n, 119-21, 123, 125-6, 128-30, 132, 146n, 149, 151, 163-5, 168n, 172-6, 191, 193, 196-7, 199-201, 203, 20712, 225-8, 230, 233-9, 244, 251-4, 257, 264, 268-9 manifest (qualities, properties), 11, 26, 38, 70, 72, 78, 113-4, 191, 197, 208-13, 225, 254 Martin, C.B., 67, 71, 106n, 110, 119n, 153, 240-1 Maxwell, James Clerk, 18 Maxwell, Nicholas, 250 measurable, 11, 36, 39, 114, 227, 235, 260, 273, 279, 281 measurement, act of, 24n, 25, 39, 249-50, 255-63, 265, 267-9 mechanism, 27, 33, 78, 83, 85, 87n, 110, 115n, 121, 124n, 129, 131, 137-8, 142-3, 145-6, 181, 188, 265 Meinongian, 71, 83n Mellor, D.H., 25n, 28, 30, 40, 67, 81n, 82, 87, 94-5, 103n, 117n, 118n, 278-80, 282 Menzies, Peter, 124n, 198, 200 metaphysics, 12, 17-20, 28-9, 31, 46, 51-2, 67-8, 71, 79-80, 85, 87, 91, 96-7, 100, 145, 168-72, 176, 181, 190, 226, 234-5, 250-1, 256, 275-6

microreduction, 122-3, 125-6, 129, 131 micro-based properties, 129 microscopic, 19, 28, 30, 32, 40, 104-5, 107, 111-2, 121-5, 129-32, 253, 255, 265, 269, 271, 275n, 277-9, 281-2 microstructure, 27, 121, 188, 269 Mill, John Stuart, 57, 106, 134n, 197 Millikan, Robert, 144 modality, 71, 78, 147-8, 187 Molière, 1-2, 109, 133, 145, 212, 267 Molnar, George, 15n, 28, 67, 71, 89n, 125-6, 145n, 240 monism, 87-8, 96, 252 Monnoyer, Jean-Maurice, 190 Mumford, Stephen, 25n, 30, 82, 86-101, 106n, 117n, 118n, 119, 126, 146n, 158n, 163n, 192, 194, 237, 240-2, 251-2, 255n natural kind, see kind, natural necessary connection, 71, 73-4, 90, 223 Newton, 3, 18, 113, 146, 189, 194, 204, 208, 218, 224, 234 non-locality, 250, 251, 256n, 260 normal conditions, see conditions, normal Ockham, 8 occurrent, 40, 71, 72, 77, 78, 87n, 106, 117n, 120, 123, 186-194, 209, 235, 251, 252, 281, 282 observable, 250, 255-258, 260, 267, 269, 270 operational, 254, 272 Oppenheim, Paul, 22 Ostwald, Wilhelm, 276 overdetermination, 130, 132, 279, 280 Pagonis, Constantine, 249, 255, 261, 267, 268, 269 Pap, Arthur, 47, 51, 56 Peirce, Charles, S., 92n, 94, 96, 97, 99, 100n Place, Ullin T., 71, 251, 269 Plato, 31, 76, 78, Popper, Karl, 23, 25n, 82, 100,114, 118, 175n, 250, Price, H.H., 122 Putnam, Hilary, 100, 241 possibility conditional, 151,

Index circumstantial, 34, 154, 155, 156n possible world, 26, 31, 47, 57n, 58, 65, 72, 78, 85, 93, 99, 170n, 184, 215-6, 218 potentiae, 8, 249, 258 power, see disposition prediction, 44, 45, 113, 114, 114n, 136, 137, 144,165, 183, 211, 223, 239, 253 propensity, 90n, 92, 100n, 175, 250, 265, 266, 268, 269 property categorical, see categorical predicate/ property dispositional, see dispositional predicate/property occurrent, see occurrent statistical, 40, 281, 282 quality, 3n, 5-7, 10-15, 97, 119, 145, quantity, 8, 11, 254, 263, 271-276 quantum holism, 251, 256, 258 quantum mechanics, 19n, 115n, 182, 238n, 249, 251, 256, 261 Quine, Willard van Orman, 25, 27, 28, 29, 30, 66, 81, 116, 121, 145 Ramsey, Frank P., 46, 57n, 58, 91n, 134n realism, 36, 76, 78n, 79, 82, 93, 96, 97, 100, 101, 105, 144n, 250n, 251, 257, 258, 268 scientific, 181, 187, 188, 194 Redhead, Michael, 250, 256 reduction sentence, 47-51, 55-6, 59, 62, 66, 94, 117n reductionism, 266 regularity, 32-3, 35-6, 46, 91-3, 135-6, 1423, 183, 188, 223-6, 231n relational, 16, 21n, 22, 96, 97, 116n, 146n, 187, 249-263, 270, Roberts, John, 201, 224n, 232, 235n, 243 Rovelli, Carlo, 39, 261-263, Russell, Bertrand, 35, 36, 45n, 76, 97, 133n, 187 Russellian fact, 108n, 109, Ryle, Gilbert, 81, 94n, 122, 127, 184, 187, 200, 201n Schrenk, Markus, 32, 36, 38, 74n, 113n,

303

169, 183 Schrödinger’s cat, 181, 266, Schrödinger’s equation, 208, 217, 259n, 265, Shepard, Roger, 126, 127, Shoemaker, Sydney, 15n, 25n, 27, 28, 30, 33n, 76, 77, 86n, 117n, 118n, 120n, 145n, 146n spontaneous collapse theories, 39, 256, 265 localization, 265 manifestation 115n Squires, Roger, 108, 109 statistical mechanics, 273, 277, 279, 281 Steward, Helen, 107, Suarez, 7-9, 13 substantial form, see form, substantial supervenience, 69, 71, 73-75, 79, 90 tendency, 9, 17, 18, 113, 114, 115, 146, 182, 186, 197, 198, 199, 203, 232, 260, 266 Tooley, Michael, 90, 134n, 136n, 138n, 140n, 216n trigger, 83n, 87n, 115n, 149, 196-7, 201-3, 227-30, 233, 237-9, 244 trope, 34, 76, 97, 134, 167, 169, 170-172 truth function, 26, 51-2, 94, 187 truthmaker, 31, 34, 38, 70, 78, 79, 85, 94, 97, 98 two-sidedness, 197, 200-1, 205 Ultra-grounding, 69, 126n “ungrounded argument”, 67-9, 71, 73, 79-80 universals, 76, 91, 92n, 95, 97n, 134n, 145n, 168n, 169, 172-176, 223 unobservable, 37, 105-6, 182, 187-90, 192, 194 van Fraassen, Bas, 182, 266n verificationism, 23, 29, 122, 123 Vienna Circle, 22, 45 weaken/enhance, 197, 201-2 Wittgenstein, 51, 100, 111n, 112n, 139n, 181, 188, 194