The Ways Things Are: Studies in Ontology 9783110325782, 9783110325102

Citation preview

Christian Kanzian / Winfried Löffler / Josef Quitterer (Eds.) The Ways Things Are Studies in Ontology

Philosophische Analyse Philosophical Analysis Herausgegeben von / Edited by Herbert Hochberg • Rafael Hüntelmann • Christian Kanzian Richard Schantz • Erwin Tegtmeier Band 44 / Volume 44

Christian Kanzian, Winfried Löffler, Josef Quitterer (Eds.)

The Ways Things Are Studies in Ontology

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.

North and South America by Transaction Books Rutgers University Piscataway, NJ 08854-8042 [email protected] United Kingdom, Ireland, Iceland, Turkey, Malta, Portugal by Gazelle Books Services Limited White Cross Mills Hightown LANCASTER, LA1 4XS [email protected]

Livraison pour la France et la Belgique: Librairie Philosophique J.Vrin 6, place de la Sorbonne; F-75005 PARIS Tel. +33 (0)1 43 54 03 47; Fax +33 (0)1 43 54 48 18 www.vrin.fr

2011 ontos verlag P.O. Box 15 41, D-63133 Heusenstamm www.ontosverlag.com ISBN 978-3-86838-129-0 2011 No part of this book may be reproduced, stored in retrieval systems or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use of the purchaser of the work Printed on acid-free paper FSC-certified (Forest Stewardship Council) This hardcover binding meets the International Library standard Printed in Germany by CPI buch bücher.de

Contents

Preface

7 

1. Before Ontology Metaphysics and Higher-Order Modal Logic (TIMOTHY WILLIAMSON)

17 

On Metaphysical Possibility – in Classical and Analytic Philosophy (OTTO MUCK)

37 

Beyond the Ultimate (LUDGER JANSEN)

57 

A Plea for a Naturalistic Ontology (JOHANNES L. BRANDL)

73 

Ockham on the Nature of Concepts (GERHARD LEIBOLD)

93 

What is Aristotelian Metaphysics? (FRIEDO RICKEN)

105 

2. Ontology at Work Experience of Change and Change of Experience (E.J. LOWE)

121 

On the Unifying Role of Laws of Nature (PAUL WEINGARTNER)

131 

Finitist Objects (SERGIO GALVAN)

149 

6

Ontology and Ordinary Objects (LYNNE RUDDER BAKER)

167 

Can One Separate Me From My Life? (CHRISTIAN TAPP)

181 

Dimensions of the Mental (FRANZ VON KUTSCHERA)

191 

3. Ontology in Application Philosophy and Neuroscience (ANTONELLA CORRADINI)

203 

How Molinists Can Have Their Cake and Eat It Too (GODEHARD BRÜNTRUP and RUBEN SCHNEIDER)

221 

Scientia Media and Freedom to Do Otherwise (CHRISTOPH JÄGER)

241 

Dualism and Resurrection (UWE MEIXNER)

263 

Living like God in Switzerland? (NIKO STROBACH)

277

The Authors

299

Preface

This volume on ontology is a means of expressing our congratulations to Edmund Runggaldier on the occasion of his 65th birthday. It is a symbol of gratitude for everything he has done and is still doing for the international scientific community as well as for philosophy in Austria, in Innsbruck, and especially for the members of his home institution, the Department of Christian Philosophy at the University of Innsbruck. The title of the volume represents two chief interests of Edmund’s philosophical work: the ontology of things and substances and their properties and characteristics. The papers were authored by distinguished philosophers whom the editors wish to thank for their kind and smooth cooperation. The contributors are colleagues and friends of Edmund’s with whom he is in frequent contact. If they are not always in absolute agreement, the resultant dialogues are always both lively and respectful. Thus, in this book, we do not intend to provide a series of uncontroversial theses but, rather, what this book offers is a collection of stimulating and incisive contributions to contemporary ontology. Edmund Runggaldier’s academic personage, his contributions to philosophy, and his leading role in Austrian and European philosophical societies do not need a lengthy introduction. Nevertheless, it may be of interest for some readers that Edmund was born in 1946 in St. Ulrich in Gröden, an Alpine valley in South Tyrol, Italy. He is a native speaker, not only of German and Italian, but also of the “minority language” Ladin. He attended high school in Milan at the Istituto Leone XIII. In 1966, he joined the Jesuit order and pursued his first studies in philosophy at Pullach near Munich, where he obtained a B.Phil. After he had spent three years in Innsbruck for theological studies, he was accepted as a D.Phil. student at the University of Oxford, under the tutelage of Alfred J. Ayer. His D.Phil. thesis Carnap’s Early Conventionalism (Amsterdam: Rodopi 1984) was one of the first monographs on Carnap, and it was a pioneering work on the early Carnap. In 1977, Edmund received his D.Phil. at Oxford, and, in the same year, he started his academic career at the Theological Faculty of Innsbruck, where he became an Assistant Professor in close cooperation with the chair of

8 Otto Muck (one of the contributors to this volume). Edmund’s habilitation thesis Zeichen und Bezeichnetes (Berlin: De Gruyter 1985) gained wide attention as a major work in philosophical semantics and one of the first in German. Edmund was appointed Associate Professor in 1985, and, a few years later in 1990, he was declared Full Professor at the Department of Christian Philosophy at the University of Innsbruck. From 1994–96, he served as head of the department, and, from 1993–95, as Dean of the Faculty of Catholic Theology. Edmund Runggaldier has presided over the Austrian Society for Philosophy (1996–98) and the Austrian Ludwig Wittgenstein Society (2000–06). From 2004–08, he was the Chairman of the Society of German-speaking philosophers at the Catholic Theology Faculties. Under his guidance, all three societies enjoyed periods of significant flourishing. He spent a semester as Visiting Scholar at the University of Notre Dame and at the Loyola University Chicago (1996–97). Visiting Professorships brought him to Fujen University in Taipei, Taiwan (2002) and “per chiara fama” to the Catholic University at Milan, Italy (2003). From 2007– 09, Edmund Runggaldier held the distinguished Guardini Chair at the Faculty of Protestant Theology at Humboldt University, Berlin. He is still a member of numerous boards and executive committees and is engaged in leading positions in international scientific communities. The list of his publications is more than 100 items long, most of his work being dedicated to topics in metaphysics, philosophy of religion, philosophy of mind, and the philosophy of language. In addition to his previously mentioned two monographs, his Analytische Sprachphilosophie (Stuttgart: Kohlhammer 1990), Was sind Handlungen? (Stuttgart: Kohlhammer 1996), and Grundprobleme der analytischen Ontologie (with C. Kanzian; Paderborn: Schöningh 1998) deserve special mention. His most recent monograph concerns Albert the Great (Die menschliche Seele bei Albertus Magnus (Münster 2010)) and offers a non-reductionistic contribution to the mind-body-problem. Among his numerous co-authored or coedited volumes, we mention only Unser Selbst. Identität im Wandel der neuronalen Prozesse (with G. Rager & J. Quitterer, Paderborn: Schoeningh 2002) and Vielfalt und Konvergenz der Philosophie (ed. with W. Löffler, Vienna: ÖBV-HPT 1999). The contributions to this volume have been grouped according to their “topoi” within the greater spectrum of ontological theorizing. The papers in section 1, “Before Ontology”, are dedicated to methodological and meta-

9

ontological topics. Section 2, “Ontology at Work” collects papers on categorial ontology. Section 3, “Ontology in Application”, addresses various topics where ontological issues come into play in practical philosophy, philosophical anthropology, and the philosophy of religion.

Section 1: Before Ontology In his essay “Metaphysics and Higher-Order Modal Logic”, Timothy Williamson tries to adjudicate between two opposing fundamental intuitions in the metaphysics of modalities, neither of which is prima facie completely irrational: “Necessitism” claims that necessarily everything is necessarily something – even if only a merely possibly concrete object –, whereas “contingentism” defends the intuition that objects could also have been nothing at all if the world had been otherwise. Williamson’s criterion is whether these intuitions can be supplied with a powerful logical theory: In a detailed argumentation, Williamson shows that necessitism in sum fares better, if a higher order modal logic (which is necessary to express many interesting modal claims) is the desired theoretical framework. Otto Muck’s contribution “On Metaphysical Possibility – in Classical and Analytic Philosophy” recalls the widely underrated contributions of mid-20th century philosophers like J. Maréchal, Bernard Lonergan, and others to the problem of how to determine essences and metaphysical possibilities. Muck contrasts them with essentialist thought in current analytic metaphysics. Essences and possibilities sometimes are depicted as existing in a separate realm beyond everyday life and scientific practice, this realm only being accessible via a sort of (contestable) intuition. Against this, Maréchal, Lonergan et al. attempt to show how our modal and essentialist discourses ultimately rest on our dealing with the actual “life-world” (Lebenswelt). Recurrent structures of problem-solving in the life-world (and similarly, in the sciences) also provide plausible accounts of the concepts of “reality” and “truth”. The title of Ludger Jansen’s contribution is “Beyond the Ultimate”. The author deals with a methodological question. What are the ontological restrictions in explanations? There are two principle alternatives for answering this question: holism and atomism. The first alternative holds that only macro-structural wholes and their properties can be accepted; the latter

10 claims that only atomistic parts may be involved in valid explanations. Ludger Jansen argues against both kinds of restrictions. In his article “A Plea for a Naturalistic Ontology”, Johannes Brandl does not intend to invent a “new type of naturalism”, but rather pleads for a nuanced account of naturalism and naturalistic programs. In contemporary debates, one finds a remarkable manifold of naturalistic positions, not only in epistemology and metaphysics, but also in ethics and theology. These positions are worth considering critically but not polemically. The ontological status of concepts is one of the most complex (yet still widely underrated) problems of philosophy. It connects to issues in, at least, epistemology, the philosophies of language, action, and mind as well as its well known connection to the problem of universals. In his paper, “Ockham on the Nature of Concepts”, Gerhard Leibold shows how some contemporary positions are prefigured in the works of one of the most eminent medieval philosophers. It turns out that Ockham’s thought has many more facets than usual allusions to Ockham would have one believe. In his article, “What is Aristotelian Metaphysics?”, Friedo Ricken reconstructs the core of Aristotle’s metaphysics in a coherent and comparatively short survey, which is based on a close philological reading of the original Greek texts. Ricken traces the ways in which Aristotle’s hinge concept, the concept of ousía, connects the various meanings of ‘being’ and establishes the relation between the domains of objects of the various sciences. In modern terms, one could say that Aristotle’s ontology takes concrete, mid-size individuals (with organisms as chief examples) as ontologically prior. Ricken finally sketches how Aristotle’s (ultimately pragmatic) ontological approach could plausibly integrate questions concerning God and the soul. Section 2: Ontology at Work Jonathan Lowe’s paper, “Experience of Change and Change of Experience,” leads to the centre of ontological research on change, time, and experience of change. Lowe holds a causal theory of perception according to which sense quales, “mental modes”, which are perceived, are caused by qualities of the perceived objects. Every experience of change is a change of such a sense experience. The passage of time depends on change, the coming into and passing out of existence of various entities such as qualities, relations, and substances.

11

One of the most discussed topics in categorical ontology recently is “laws of nature”. One’s view on laws of nature not only has implications for one’s theory of properties and relations but also for one’s views on causality and causation. Paul Weingartner’s contribution “On the Unifying Role of Laws of Nature” pertains to this discussion. Weingartner’s main point is that laws of nature unify the multitude and plurality of properties that inhere in things. To point out this theoretical function of laws of nature, Weingartner gives a differentiated view on the meaning of “law” and “laws on nature”, but also of the ways the unification at issue takes place. Sergio Galvan’s paper, “Finitist Objects” provides an example of the close connection between investigations in ontology (here: the ontology of mathematical objects) and methodological or meta-ontological considerations. Galvan shows that the philosophy of finitism in mathematics is laden with various problems (despite its being proposed as a way out of ontological and proof-theoretical problems with transfinite domains of objects). “Intuition” is supposed to grasp finitist objects. This “intuition”, however, admits of different explications, and, thus, the nature of the finitist objects may be described differently. Galvan argues for the quasi-abstract character of these objects and in favour of William Tait’s thesis that PRA is the right formal system for handling them. In “Ontology and Ordinary Objects”, Lynne Rudder Baker defends the reality of ordinary objects. According to her view, the objects of our everyday life cannot be ontologically reduced to the physical particles they consist of because they belong to a different natural kind essentially. As such, they are constituted by their ultimate material constituents but they are not identical with these constituents. After rejecting several objections against her view, Rudder Baker presents her constitution view as a coherent and economical way to defend a non-reductive ontology of ordinary objects “that we encounter and interact with in everyday life.” There is an ongoing discussion in ontology as to whether real entities have to be conceived in a three- or four-dimensional ontology. In his contribution, “Can one separate me from my life,” Christian Tapp shares Runggaldier’s thesis on the compatibility of four and three-dimensionalism and presents an integrative perspective according to which the two ontological theories are interrelated with each other. Tapp refers to the interrelationship between me and my life to demonstrate the mutual dependence of three and four-dimensionalism.

12 Can the mental be reduced to the physical or does it constitute a reality of its own? In his contribution “Dimensions of the Mental”, Franz von Kutschera presents the infinity and openness of the reflecting and calculating mind as an irreducible aspect of the mental. Moreover, the hierarchies of abstract objects as they are developed in set theory give rise to a host of contradictions and paradoxes, unless they are understood as concepts or mental constructs. This conceptualistic understanding of abstract objects does not only avoid the contradictions of the Platonic conception (abstract objects are independent of the human mind), it comprises an infinite hierarchy of entities which exceeds the physical world and therefore is a “decisive obstacle to materialism.” Section 3: Ontology in Application In her paper “Philosophy and Neuroscience”, Antonella Corradini discusses different strategies for maintaining the autonomy of philosophical and metaphysical reasoning in the era of neurophilosophy, where philosophical discourse constitutes only an appendix to neuroscientific research. After a refutation of Bennett’s and Hacker’s extreme defence of philosophy as an apriori and purely conceptual enterprise, Corradini presents her own view of a dualistic kind of emergentism as a metaphysically sound way to deal philosophically with the recent discoveries of neuroscience. Questions concerning the relationship between deterministic causation and human agency, time and modality, semantics and the ontology of counterfactuals are central issues of applied ontology. Godehard Brüntrup, Ruben Schneider, and Christoph Jäger discuss these questions on the basis of Molina’s Concept of middle knowledge: Can theological compatibilism, the idea that God’s knowledge of our future decisions is compatible with our freedom, be formulated in such a way that it does not lead to nomological compatibilism, which excludes a libertarian concept of free will? For Brüntrup and Schneider (“How Molinists Can Have Their Cake and Eat It Too”), the only solution for this problem is to assume a subject’s “counterfactual power” over the subject-related knowledge of God. In his contribution “Scientia Media and Freedom to Do Otherwise”, Jäger proposes a solution within the metaphysical framework of Molina. According to Jäger, the fact that we do not have causal influence on God’s knowledge of our future decisions does not imply that we do not causally influence

13

our decisions. Therefore, we need no counterfactual power on God’s knowledge in order to preserve a libertarian concept of will. The title of Uwe Meixner’s paper “Dualism and Resurrection” reveals clearly its content. At first, Meixner tries to define “dualism” in a weak, non-standard way. For Meixner, dualistic theories imply logically the thesis that some mental entities are at least partly nonphysical. The “I” or “the soul”, that is the “human person” is such a mental entity. Identity between this life and the glorified life is preserved for the human person but not for the entire human being. The body we will own in our glorified life is essentially different from our present organic body. Similar topics to those raised in Meixner’s article are focused on in Nico Strobach’s, “Living like God in Switzerland?”. He discusses the soul, the eternal present, and beatitude in Jean-Jacques Rousseau‘s fifth promenade. The promenade is better known among scholars of French literature than among philosophers and Rousseau might seem not to belong to Edmund Runggaldier’s main-interests. The topics Strobach discusses, however, are philosophical without doubt and are crucial to Runggaldier’s work, such as traditional ideas on divine life and its occupation, eternity, consciousness of time, the soul, and supreme happiness. Strobach provides a systematic analysis of Rousseau’s text, of the tensions it contains, and indicates what we might learn from them. We finally wish to express our gratitude to the many people who supported the publication of this volume in various ways. Thanks to the authorities of the University of Innsbruck and of the Faculty of Catholic Theology, especially to our rector Professor Tilmann Märk, and to the Dean of our Faculty, Professor Józef Niewiadomski. Special thanks go to Ontos Publishers and their director Dr. Rafael Hüntelmann for accepting this volume in the series Philosophical Analysis. Thanks go to Terrence Ehrman CSC, David Galvin, Adam Green, and Christoph Jäger for linguistic advice. This volume would have been impossible without Monika Datterl’s thorough and patient work on the lay-out, and Ksenia Scharr’s support in technical and organisational matters. Innsbruck, October 2011

C.K., W.L., J.Q.

1. Before Ontology

Metaphysics and Higher-Order Modal Logic TIMOTHY WILLIAMSON, Oxford*

One of the most fundamental issues in modal metaphysics can be summed up in a single formula of first-order modal logic: NNE □  x □  y x=y This says that necessarily everything is necessarily something. On the interpretation of interest here, the necessity is metaphysical, what would have been so whatever had been so, and the quantifiers unrestricted, ranging over absolutely everything there is. Contingentists deny NNE. Typically, they hold that each of us is a counterexample: if your parents had never met, you would have been nothing at all. Necessitists assert NNE. Typically, they hold that if your parents had never met, you would still have been something, a merely possible person, something that is not a person but could have been one. Both views save the ordinary phenomena. Deciding between them requires developing each view as a systematic theory, to facilitate the application of scientific criteria of theory choice. This paper pursues that methodology in the wider setting of higher-modal logic. In making the comparison, we can uncritically assume no form of possible worlds semantics for higher-order modal logic, for its use of an outer domain that includes the domains of all worlds is radically misleading from a contingentist perspective. Rather, we should assess principles of higher-order modal logic in the object-language for theoretical virtues such as strength, simplicity and compatibility with what we already know. 1. Comprehension and higher-order necessitism Higher-order quantification theory requires comprehension principles. In crudely nominalizing terms, they tell us when an open formula defines a member of the relevant higher-order domain. We may conveniently focus on the special case where the higher-order variable is one-place and second-order:

18

Timothy Williamson

CompM  X □  x (Xx ↔ A) Here x is an individual variable that may occur free in the formula A and X is a monadic first-level predicate variable that does not occur free in A. Any result of prefixing an instance of CompM with necessity operators and universal quantifiers of any type (binding parameters in A) in any order also counts as an instance of CompM. Roughly, CompM says that there is a property having which is strictly equivalent to being such-and-such, where A says of the value of x that it is such-and-such (the property of being such-and-such). In some logics we can derive CompM by  -generalization from the elementary λ-conversion principle □  x (λx(A)x ↔ A), and regard the property as the referent of the predicate λx(A). In any case, CompM is an attractive comprehension principle as it stands. Standard possible worlds model theory for higher-order modal logic validates all instances of CompM. Although that will not convince contingentists or even all necessitists that CompM is informally valid, it does at least show the schema to be part of one consistent and natural picture of modal reality. Independently of the formal semantics, something like CompM is the rationale for pretheoretically compelling arguments by second-order generalization. The latter consideration carries as much weight for contingentists as for necessitists. CompM requires a plenitudinous theory of properties, on which open sentences like ‘x is not red’ and ‘x is red or green’ define properties, for we can put ¬Rx or Rx  Gx for A. Thus sharing a property does not imply similarity. Whatever the virtues of a sparse theory of natural properties for other purposes, it is unsuitable for a good theory of higher-order logic, modal or non-modal. Without a reasonably strong comprehension principle, higher-order logic is impotent. We therefore adopt CompM as a working hypothesis. CompM implies a second-order analogue of the necessitist principle NNE for properties: necessarily every property is necessarily something. To see this, we must first formalize the second-order principle. For a firstorder variable x, the open sentence ‘x is something’ is formalized as  y x=y. Thus to formalize ‘X is something’ for a second-order variable X we need the analogue for predicates of the identity predicate for individual terms. The natural candidate is necessary coextensiveness. Hence the analogue of x=y is □  x (Xx ↔ Yx), so the analogue of  y x=y is  Y □  x (Xx ↔ Yx), and the analogue of NNE is this:

Metaphysics and Higher-Order Modal Logic

19

NNEM □  X □  Y □  x (Xx ↔ Yx) But if we instantiate CompM with Yx for A and the prefix □  Y □, the result is □  Y □  X □  x (Xx ↔ Yx), which is trivially equivalent to NNEM. In effect, CompM entails that properties have non-contingent being. A similar point holds for higher types: the corresponding general comprehension principle Comp entails the corresponding analogue of NNE for all types. It does not entail NNE itself. Thus the natural comprehension principle for higher-order modal logic entails the analogues of necessitism at all orders beyond the first. Contingentists have two strategies in response. They may accept that comprehension principle and argue that the difference between first-order and higher-order quantification is deep enough to justify postulating the asymmetry in logic. Alternatively, they may reject that comprehension principle and seek another that does not entail the higher-order analogues of necessitism. The next section explores the former strategy; sections 3 and 4 explore the latter. For simplicity, we will assume S5 as the background propositional modal logic for metaphysical modality. 2. Individuals and their haecceities Suppose that the contingentist accepts Comp, and in particular CompM. We can derive some connections between the first-order and second-order quantifiers. For putting x=y for A and adding the prefix □  y □ gives this instance of CompM: (1)

□  y □  X □  x (Xx ↔ x=y)

Necessarily for each thing necessarily there is the property of being that thing: having the property is necessary and sufficient for being identical with the thing. Informally write ‘X is a haecceity of y’ for □  x (Xx ↔ x=y) and ‘X is a haecceity’ for ◊  y □  x (Xx ↔ x=y). Then (1) says that necessarily everything necessarily has a haecceity. The haecceity is unique, in the strong sense that necessarily for every haecceity of y necessarily every haecceity of y is necessarily coextensive with it. For in S5 nothing of the form □A is contingent, so it is not contingent whether something is a haecceity of y; but if both X and Y are haecceities of y then they are necessarily coextensive with each other. Since necessary coextensiveness is the analogue of

20

Timothy Williamson

identity in the present setting, that suffices for uniqueness. Furthermore, a haecceity determines a unique individual, in the strong sense that necessarily for everything of which X is a haecceity necessarily everything of which X is a haecceity is identical with it. Clearly, in some sense there is a one-correspondence between individuals and their haecceities. It is tempting to go further, and say that there is a one-one correspondence between possible individuals and their haecceities. However, such talk is radically misleading. In the relevant unrestricted sense of ‘individual’, contingentists and necessitists should agree that there are no merely possible individuals; everything is actually something. For necessitists, there is no contingency in what individuals or haecceities there are, so the actual one-correspondence between individuals and their haecceities covers all possibilities. For most contingentists, there could have been individuals other than all the actual ones. In this section, we have supposed that the contingentist accepts CompM, and therefore NNEM. On this view, there actually are properties that could have been the haecceities of individuals other than all the actual ones. Thus not every haecceity is the haecceity of something; some haecceities merely could have been the haecceity of something. Haecceities as characterized above resemble the individual essences on which Alvin Plantinga bases his contingentist account of quantified modal logic. For him, a property X is an essence if and only if possibly something is not possibly something without having X, and necessarily nothing else has X. A second-order version of Plantinga’s theory would entail (1). Moreover, he holds, essences have necessary being: my essence would be uninstantiated but still something if I were nothing. That is why he uses essences as surrogates for ‘possible individuals’ to make possible worlds semantics metaphysically respectable: there are no merely possible individuals, but there are merely possibly instantiated essences. Thus (1) and NNEM generate a theory of haecceities very similar to Plantinga’s theory of individual essences. However, this also means that, given CompM, contingentism is subject to objections analogous to those made to Plantinga’s contingentist theory. The initial challenge is this. My haecceity necessarily tracks me in a modal sense, because it is necessary that it is a haecceity of me and of nothing else. Even if I were nothing, there would still be a property tracking me and only me. But how can it single me out in my absence? For example, if it tracked me by somehow being built out of me and the identity

Metaphysics and Higher-Order Modal Logic

21

relation, how could it still track me when there is no such thing for it to be built out of? A parallel challenge can be raised in terms of anti-haecceities, defined like haecceities except with ¬x=y in place of x=y and established by a corresponding application of CompM. My anti-haecceity is the property of not being me. It too necessarily tracks me in a modal sense, because it is necessary that it is an anti-haecceity of me and of nothing else. How can my anti-haecceity single me out in my absence? The underlying problem is more general. By NNEM, properties have necessary being. Nevertheless, as haecceities and anti-haecceities illustrate, by CompM they must also be capable of tracking a particular individual even when there is no such individual. Does any plausible metaphysics of properties combine these features? The contingentist may envisage properties as purely qualitative in some sense. But then the nature of individuals must permit them to be tracked, singled out, by purely qualitative properties. For example, suppose that exactly the same qualitative possibilities are open to Tweedledum and Tweedledee. If contingently there were no Tweedledum, no Tweedledee, and no other individuals specifically related to them, there would still be their haecceities, one tracking only Tweedledum, the other tracking only Tweedledee. But in those counterfactual circumstances how could a purely qualitative property single out one of them and exclude the other? A purely qualitative description would fit both or neither. Thus the purely qualitative conception of properties may well require a highly contentious form of the identity of indiscernibles for individuals, on which qualitative identity entails numerical identity. That is a far less plausible claim than the trivial form of the identity of indiscernibles that does not assume properties to be purely qualitative. We have no serious evidence against the metaphysical possibility of a symmetrical universe in which every individual can be reflected (rotated, translated) onto its double. The problem for the purely qualitative conception of properties is not just that it yields an implausible form of the identity of indiscernibles. The latter in turn yields a conception of individuals themselves as purely qualitative in nature. But they must not be purely qualitative in the way properties are supposed to be purely qualitative, for the latter way was supposed to explain why properties are necessarily something, and the contingentist denies that individuals are necessarily something. Thus ways of being purely qualitative seem to multiply.

22

Timothy Williamson

The contingentist might conceive an individual as a bundle of purely qualitative properties, held together by a primitive purely qualitative higher-order compresence relation and individuated by the properties in the bundle, where both the properties and the relation have necessary being but are only contingently instantiated: if the compresence relation were not instantiated, there would be no individual (compresence may be multigrade, in holding between different numbers of properties). The theory becomes still more elaborate once fitted out with an account of the persistence of individuals across times and possibilities, since an individual typically has many of its purely qualitative properties, such as shape, temporarily and contingently. Alternatively, if the theory denies identity through change and through contingency, not only is that yet another implausible consequence, which requires still more theoretical complexity to save the appearances, it also fits badly with the underlying motivation for contingentism, by treating a vast range of apparent contingency as an illusion. Thus the purely qualitative conception drags the contingentist into proliferating complications of metaphysical theory with no independent plausibility. They are not what most philosophers who signed on to contingentism bargained for. The notion of the purely qualitative is not altogether clear. Which properties count as purely qualitative? We may even count relational properties as purely qualitative when definable without reference to specific individuals: for example, being inside a hollow sphere, as opposed to being inside this hollow sphere. Thus purely qualitative properties may be extrinsic and relational. Such properties still seem to satisfy NNEM. They are also preserved under any symmetries of the universe. The picture is that purely qualitative properties correspond to predicates of the form λv(A), where A contains no individual constants and no free individual variables except v, but may contain quantifiers of any order. Although CompM also has instances such as (1) in which A contains constants or variables as parameters, the corresponding ‘impure’ predicate λv(A) is supposed to pick out a property that could in principle also be picked out by a ‘pure’ predicate λv(B), where B satisfies the restrictions. Unfortunately, the syntactic criterion is of limited help, since the reference of a syntactically atomic predicate such as ‘Hellenic’ may nevertheless be fixed in relation to particular people or places. The notion of a predicate that does not semantically ‘involve’ any particular individual is itself far from clear. Here the coarsegrained individuation of properties by necessary coextensiveness exacer-

Metaphysics and Higher-Order Modal Logic

23

bates the problem, because it undermines the idea that properties have constituent structure analogous to that of predicates, for necessarily coextensive predicates may have utterly dissimilar constituent structures. Thus a property’s involving an individual cannot be explained as the individual’s being a constituent of the property. Given the unclarity in the notion of the purely qualitative, the argument that the purely qualitative conception of properties yields an implausible metaphysics of individuals can only be provisional. But the onus is on a proponent of that conception to supply the requisite clarification, preferably in a way that avoids the bad metaphysical consequences. As yet there is no sign of that happening, and the argument stands. By appealing to the purely qualitative conception, contingentists talked their way into worse trouble. They do better to adopt a more minimalist approach, denying that they must provide any metaphysically deeper account of how properties and relations are individuated than one in terms of their modal application conditions, on which necessary coextensiveness is analogous to identity. The minimalist endorses CompM, and with a similar attitude its generalization Comp to higher types, while insisting that we need no more explanation of how my haecceity singles me out in my absence than that it is the property that, necessarily, applies to something if and only if it is me. One danger for minimalist contingentism is that the minimalism leaves the contingentism looking suspiciously ad hoc, its combination with Comp a messy hybrid. As seen in section 1, Comp implies necessitism at every order except the first. The onus is on the metaphysician who postulates such logical differences between orders to justify the asymmetry in treatment. Minimalism deprives the contingentist of the resources needed to provide such a justification. The problem can also be put in more metaphysical terms. According to the minimalist contingentist, if your parents had never met, there would not have been you, a merely possible person, but there would still have been your haecceity. The latter claim depends on a minimalist standard of being for properties, on which they can be even in circumstances in which they leave no trace. The former claim, that there would not have been you, depends on a non-minimalist standard of being for individuals, on which they can be only in circumstances in which they leave some trace. Of course, this talk of leaving a trace is metaphorical, but it is the minimalist contingentist who needs to cash out the metaphor. The default preference is for a

24

Timothy Williamson

uniform metaphysics, on which being is contingent at all orders or none. Given the minimalist comprehension principle CompM or Comp, secondorder being is not contingent. Therefore the default preference favours a metaphysics on which being is not contingent at any order including the first, in other words, necessitism. The minimalist contingentist may reply that there is a compelling metaphysical difference between the first-order quantifiers and their higherorder analogues, because the former range over (although perhaps not exclusively over) sparse flesh and blood ‘thick’ particulars, whose being requires causal activity and spatiotemporal location, while the latter range only over plenitudinous bloodless ‘thin’ universals, whose being requires only a well-defined application condition. But even if we grant that metaphysical difference for the sake of argument, it does not justify the asymmetry in the treatment of the quantifiers, because it does not really explain the non-contingency of higher-order being. Take my haecceity for example: of course there is actually a well-defined application condition, but to draw from that premise the conclusion that there would have been that well-defined application condition even if there had not been me is to assume something like NNEM, one of the very principles to be explained. The condition is defined by the assignment of me as value to the parameter y in the predicate λx(x=y). Without me, there is nothing to assign, so no such assignment. Indeed, ‘well-defined application condition’ is barely more than a notational variant of ‘property’, in the relevant plenitudinous sense. To put it the other way round: conceiving the second-order quantifiers as ranging over legitimate semantic values for predicates automatically validates NNEM only if conceiving the first-order quantifiers as ranging over legitimate semantic values for names automatically validates NNE. Could the minimalist contingentist claim instead that higher-order quantifiers, unlike first-order ones, are somehow not really in the scope of modal operators prefixed to them? That is to postulate a grammatical difference between first-order and higher-order quantifiers that in no way follows from any metaphysical difference in what they range over. One could just as easily postulate a reading of the first-order quantifiers on which they are not really in the scope of the modal operators prefixed to them. Such a reading would trivially validate NNE. Presumably that would not be a decisive argument for necessitism, because it would not decide in favour of NNE on a reading of the first-order quantifiers that is not ‘scoped out’. Similarly, why should the scoped out reading of NNEM settle the issue in

Metaphysics and Higher-Order Modal Logic

25

favour of NNEM on a reading of the second-order quantifiers that is not scoped out? There is such a reading, for the metaphysical differences imply nothing about the purely semantic question of scope. In any case, the grammar of a standard formal language for higher-order modal logic is clear: higher-order quantifiers do occur in the scope of modal operators. The safest line for the minimalist contingentist may be even more minimalist: to lay down Comp as a comprehension principle attractive simply for its general theoretical virtues, and derive the non-contingency of higher-order being as a corollary, but uphold the contingency of first-order being as common sense. Such an ultra-minimalist contingentist abjures any attempt to explain or justify the asymmetry on deeper metaphysical grounds, as likely just to lead to trouble. The combination of Comp with (first-order) contingentism can be shown to be internally consistent. Nevertheless, ultra-minimalist contingentism is unsatisfying. It has internal tensions about which it stays silent, rather than attempting to resolve them. On one hand, the theoretical virtues such as simplicity and strength that recommend Comp and higher-order necessitism also recommend firstorder necessitism. On the other hand, the allegedly common sense rejection of the latter threatens the former too, since common sense has difficulty with the property of being a given individual when there is no individual for it to be the property of being. Thus the motivations for the two components of ultra-minimalist contingentism appear to conflict with each other. If the appearance is deceptive, we want to know why. It is hard to rest content with the unexplained asymmetries in ultra-minimalist contingentism. We might have to in the absence of a tolerable alternative, but necessitism is such an alternative. Full-blooded contingentists may also object to all forms of minimalist contingentism that they are not contingentist enough, because they deny the contingency of being at higher orders. Since Comp requires that denial, full-blooded contingentists may therefore reject Comp, and seek an alternative comprehension principle that lacks its higher-order necessitist consequences. Since they are independently committed to not taking possible worlds semantics at face value, the validity of Comp in the most natural forms of such semantics will not impress them. The next two sections examine different sorts of alternative comprehension principle for contingentist higher-order modal logic.

26

Timothy Williamson

3. Non-modal comprehension One natural way of weakening Comp is to omit the modal operator between the two quantifiers. In the case of CompM, the result is this: CompM−  X  x (Xx ↔ A) In all other respects CompM− is like CompM. To see the point of the weakening, consider the instance of CompM− with the same prefix and the same substitution for A as (1), the haecceityyielding instance of CompM: (1−)

□  y □  X  x (Xx ↔ x=y)

This does not imply any modal tracking of individuals: when I am the value of the variable y, it requires only that in each possible circumstance, if I am something then there is a property that I alone have (but in other possible circumstances other individuals may have it, and I may lack it even though I am something), and if I am nothing then there is a property that nothing has (but in other possible circumstances anything may have it). Thus (1−) lacks the consequences of (1) that made trouble for contingentism. The constraint on the extension of X is modally local rather than global. As the visible absence of modal operators makes clear, CompM− is in some fundamental way a non-modal principle. Of course, we can necessitate its instances, as in (1−), and we can also substitute formulas with modal operators for A, but in those respects CompM− is no different from a principle of non-modal propositional logic such as the law of non-contradiction ¬(A & ¬A). CompM− is provably consistent with ¬NNE & ¬NNEM, so that it allows the first-order contingentist to be a second-order contingentist too. A similar argument shows that Comp−, the generalization of CompM− to all higher types, is consistent with the family of negations of the analogues of NNE for all types. Thus CompM− permits a thoroughgoing contingentism over all types. We can use the same sort of model to show that CompM− is also consistent with the negation of (1), so that it does not require haecceities. We can also show that not even the combination of CompM− with the analogues of NNE for all types requires haecceities.

Metaphysics and Higher-Order Modal Logic

27

Such results show that CompM− is weak enough for the contingentist’s purposes. However, it may be too weak. For example, it does not underwrite pretheoretically valid arguments by second-order generalization. To be more precise, consider (2), where T is an atomic two-place predicate constant: (2)

(Taa & ◊¬Taa) →  X (Xa & ◊¬Xa)

A typical instance of (2) can be roughly paraphrased as a banal truth such as ‘If Alice contingently talks to herself, there is something that Alice contingently does’. One can derive (2) from CompM and the being constraint that it is impossible to stand in the relation expressed by T without being something (to tie up loose ends about non-being) in any reasonable secondorder modal logic. But (2) does not follow from CompM− and the being constraint, even with the help of NNE and its higher-order analogues. The rationale for (2) comes from a conception of the values of the second-order variables as properties for which coextensiveness does not imply identity, although necessary coextensiveness does: a non-extensional but intensional view. By contrast, (2) is implausible on an extensional view, for example Boolos’ plural interpretation of second-order logic (with suitable provision for the vacuous case). No individuals are such that necessarily they are all and only those who scratch themselves, or all and only those who scratch themselves at least once but less than a million times. CompM− does not even entail the principle that any two properties have a conjunction: (3)

Y Z X

□  x (Xx ↔ (Yx & Zx))

More generally, let % express one of the sixteen binary truth-functions. Then one can show that CompM− entails (%) only in the trivial case when % expresses either the left projection, A % B being always equivalent to A, or the right projection, A % B being always equivalent to B: (%)

Y Z X

□  x (Xx ↔ (Yx % Zx))

Thus CompM− does not yield the most elementary non-trivial closure principles for which one would hope in a systematic theory of properties on the plenitudinous conception. For example, it does not require properties to be closed under negation or disjunction, in the sense of (4). Without such Boolean combinations of properties, second-order logic is crippled.

28

Timothy Williamson

We can see the weakness of CompM− from another perspective by viewing plenitudinous properties as conditions. One may inquire into the necessary or sufficient conditions for life to evolve on a planet, or for someone to have the right to vote under a given constitution. Let A be ‘life evolves on planet x’, ‘x has the right to vote under this constitution’, or whatever. There is no guarantee that our language in its current form has the expressive resources to articulate all such conditions, for instance for life. In the simplest case, asking about necessary conditions for A is asking about λX(□  x (A → Xx)); asking about sufficient conditions is asking about λX(□  x (Xx → A)). A (partial) answer to the question may be, for example, ‘A necessary condition for having the right to vote under this constitution is being at least 18 years old’ (the conditions for having the right to vote under a given constitution need not cut nature at its joints). Let B be ‘x is at least 18 years old’. The form of the answer is □  x (A → B). But that answers the original question about necessary conditions, λX(□  x (A → Xx)), only if (4) holds, to connect the question to the answer: (4)

X

□  x (Xx ↔ B)

But (4) is an instance of CompM, not of CompM−. A similar point applies even to inquiry into conditions that are not strictly necessary or sufficient, but only normally or probably or other things being equal so, for those still involve a modal correlation of some sort between the relevant A and B. In general, (4) is still needed to get the same modal correlation between A and Xx. Thus CompM− is too weak adequately to underpin the role of secondorder logic in serving theoretical inquiry. 4. Non-singular comprehension If contingentists are to formulate a comprehension principle weaker than CompM but strong enough for the purposes of theoretical inquiry, they need an alternative approach. A natural starting-point is the idea that first-order contingency induces corresponding higher-order contingency: what properties there are depends on what individuals there are, because some properties ontologically depend on individuals in the way my haecceity ontologically depends on me. Without the individual, there is no property. A comprehension principle suited to such a view will make the being of the property at issue depend on the being of the things denoted by the pa-

Metaphysics and Higher-Order Modal Logic

29

rameters and non-logical constants in the defining formula. The idea warrants no blanket ban on all instances of the comprehension schema with parameters or non-logical constants. Indeed, such instances are essential for second-order logic to serve some of its central logical and mathematical roles. For example, to capture the full intent of the principle of mathematical induction one must use a second-order formulation: MI

X

((X0 &  n (Xn → Xsn)) →  n Xn)

In mathematics, we need MI to generate consequences like (5): (5)

(Ra0 &  n (Ran → Rasn)) →  n Ran

But (5) follows from MI by universal instantiation only if (6) holds: (6)

X n

(Xn ↔ Ran)

If our comprehension principle forbids the formula A to contain parameters or non-logical constants (such as a), we cannot expect it to yield (6) (an instance of CompM−), let alone (7) (an instance of CompM): (7)

X

□  n (Xn ↔ Ran)

The proposal is instead to condition the relevant instance of CompM on the being of anything specified in the formula A. For example, in the haecceity principle (1) A is just x=y, in which the only parameter or non-logical constant is the variable y (x is not a parameter because it is bound by the universal quantifier in CompM itself). Thus the relevant being condition is just  x x=y, and the conditioned version of (1) is (8): (8)

□  y □(  x x=y →  X □  x (Xx ↔ x=y))

Unlike (1), (8) does not imply that there is my haecceity even if there is not me. Thus it avoids the feature that made (1) problematic in a contingentist setting. The same goes for the corresponding principle for anti-haecceities: since the defining formula ¬x=y contains the same non-logical constants and parameters as x=y, the qualifying condition  x x=y is the same: (9)

□  y □(  x x=y →  X □  x (Xx ↔ ¬x=y))

The general form of the new contingentist comprehension principle is this:

30

Timothy Williamson

CompMC1 EA →  X □  x (Xx ↔ A) Here EA is the analogue of the condition  x x=y in (8) and (9), depending on the structure of the formula A. The same provisos apply as to CompM. The results of prefixing instances of CompMC1 with necessity operators and universal quantifiers of any type in any order also count as instances of CompMC1, as in (8) and (9). Care is needed in defining EA. We might try equating it with the conjunction of  x t=x for each individual constant or parameter t in A (if A contains no such terms, EA is a tautology). However, that definition does not fit the underlying motivation, on which contingencies of first-order being produce contingencies of higher-order being, so that we must also take account of higher-order constants or parameters in A. Indeed, if EA were so limited to first-order terms, we could still derive unconditional NNEM from CompMC1 as in section 1, for the relevant formula A is simply Yx, which lacks individual constants and parameters; thus the commitment to the non-contingency of second-order being would remain. To avoid it, EA must also have conjuncts analogous to  x t=x for each higher-order non-logical constant or parameter in A too, of the form described in section 1. When A is Yx, the one such expression is the second-order parameter Y; so EYx is  X □  x (Yx ↔ Xx). Thus the correct instantiation of CompMC1 for Yx is logically trivial: (10)

□  Y □(  X □  x (Yx ↔ Xx) →  X □  x (Xx ↔ Yx))

With the proper definition of EA, CompMC1 is a comprehension principle for second-order free modal logic. Is CompMC1 strong enough for mathematical applications of secondorder logic? We want to derive (5) as an instance of MI. The relevant formula A is Ran, which has no parameters but the first-order constant a and the second-order constant R, both non-logical. Thus the relevant instance of CompMC1 is (11): (11)

(  x a=x &  U □  x  y (Rxy ↔ Uxy)) →  X □  n (Xn ↔ Ran)

We can derive (12) from (11) by MI: (12)

(  x a=x &  U □  x  y (Rxy ↔ Uxy)) → ((Ra0 &  n (Ran → Rasn)) →  n Ran)

Metaphysics and Higher-Order Modal Logic

31

How to eliminate the antecedent of (12) to reach the unconditional (5), the intended application of mathematical induction? For a contingentist, the antecedent of (12) is a highly non-trivial assumption, especially when we necessitate (12) to reason about counterfactual situations. Such situations need not contain the relevant individual (a) or relation (R). For in applications of mathematics to natural or social science, the former need not be a mathematical object and the latter need not be a mathematical relation. The natural way to finesse the difficulty is by invoking being constraints on R, for if the absence of the relevant individual or relation falsifies Ra0, it thereby verifies the consequent of (12) vacuously. More specifically, we obtain (13) and (14) from first-order and second-order being constraints respectively: (13) (14)

Ra0 →  x a=x Ra0 →  U □  x  y (Rxy ↔ Uxy)

From (12)–(14), (5). Thus CompMC1 seems to be strong enough for mathematical applications. Unfortunately for the contingentist, the problems do not end there. For by the anti-haecceity principle (9), my anti-haecceity is such that everything would have it if there were not me. Although (9) alone does not imply that there would be my anti-haecceity in those circumstances, that does follow once we add the second-order being constraint for my antihaecceity analogous to (14). A more general version of the constraint is this: (15)

□  X □  x □(Xx →  Y □  x (Xx ↔ Yx))

But even in systems of higher-order S5 acceptable to contingentists, (15) and (9) yield: (16)

□  y □(  x ¬x=y →  X □  x (Xx ↔ ¬x=y))

This raises something close enough to the problem of anti-haecceities in section 2 to discredit that comprehension principle. It also fits badly with the motivation for CompMC1, the idea that the contingency of first-order being induces corresponding contingencies of higher-order being. The problem is general. For it is both plausible and technically desirable for properties (in a liberal sense) to be closed under the Boolean operation

32

Timothy Williamson

of negation, so that a formula defines a property if and only if its negation does: (17)

□(  X □  x (Xx ↔ A) ↔  X □  x (Xx ↔ ¬A))

The two formulas A and ¬A draw the same line; they differ only by marking different sides of it. Whatever metaphysical underpinnings one property needs, the other needs too. In terms of CompMC1, EA and E¬A are identical. But in the setting of higher-order S5 with the higher-order being constraint (15), (17) entails (18): (18)

◊  X □  x (Xx ↔ A) → □(  x x=x →  X □  x (Xx ↔ A))

This comes close enough to saying that no property has contingent being to make the antecedent EA in CompMC1 virtually pointless. The contingentist might draw the moral that even when the metaphysical precondition EA associated with a formula A obtains, so A determines a property, we still cannot expect A to determine what has that property in counterfactual circumstances in which the precondition fails. That suggests modifying the comprehension principle CompMC1 by adding the condition EA at a second point: CompMC2 EA →  X □(EA →  x (Xx ↔ A)) For example, if we make the anti-haecceity substitution of ¬x=y for A in CompMC2, the result is (19) instead of (9): (19)

□  y □(  x x=y →  X □(  x x=y →  x (Xx ↔ ¬x=y)))

This is harmless to the contingentist, even in the presence of the secondorder being constraint (15), because it allows that nothing has my antihaecceity when there is not me. Adding the condition at the second point does not permit the contingentist to drop it at the first, to give the comprehension principle  X □(EA →  x (Xx ↔ A)). For that still yields an antihaecceity principle that implies the unconditional being of a property that modally tracks the relevant individual enough to raise the problems for CompM discussed in section 2. CompMC2 does not make properties closed under negation, and so avoids consequences such as (18). For example, let F be a non-logical atomic monadic predicate, and consider these instances of CompM:

Metaphysics and Higher-Order Modal Logic

(20) (21)

X X

33

□  x (Xx ↔ Fx) □  x (Xx ↔ ¬Fx)

We can prove that even granted  x x=x and all instances of CompM− and the first-order and second-order being constraints as auxiliary assuimptions, CompMC2 does not make (20) entail (21). For the purposes of theoretical inquiry, however, the weakness of CompMC2 is a disadvantage. For example, think of an investigation into the necessary conditions for life to evolve. As a partial answer, a theorist suggests the absence of an atmosphere full of a chemical C. Let Lx and Fx abbreviate ‘Life evolves on planet x’ and ‘The atmosphere on x is full of chemical C’ respectively. At this level of abstraction, we can adequately formalize the theorist’s claim thus: (22)

□  x (Lx → ¬Fx)

For reasons explained in section 2, (22) meets the request for a necessary condition for life to evolve only if ¬Fx expresses such a condition, and so only if (21) holds. Thus CompMC2 does not deliver the required condition. For the contingentist, we may assume, the point of the restrictions in CompMC2 here is that it is contingent that there is such a chemical as C, and so contingent that there is such a condition as having an atmosphere full of C. Thus (20) is a contingent truth. The idea is that even when (20) is true, (21) is still false, because in the possible circumstances in which there is no such chemical as C, there is no such condition as not having an atmosphere full of C, so by the second-order being constraint nothing meets such a condition. All we have is a condition met by whatever does not have an atmosphere full of C when there is such a chemical as C. But the theorist may easily hold that such a restricted version of the condition of not having an atmosphere full of C is not necessary for life to evolve, because life can evolve even if there is no such chemical as C. Thus the second restriction in CompMC2 frustrates the theoretical purpose of framing inquiry in terms of conditions, as formalized in higher-order modal logic. Can the contingentist reply that the request for necessary conditions really just expected an answer of the form of (22), so that (21) is irrelevant? That is in effect to abandon higher-order modal logic as a framework for understanding such theoretical talk of conditions. Moreover, as noted in section 3, questions about the conditions for a phenomenon do not presuppose that our language is already expressively adequate for their answers.

34

Timothy Williamson

The process of inquiry may involve expanding its vocabulary. If the phrase ‘an answer of the form of (22)’ anticipates such expansions, which must be meaningful, its generality over their semantics already requires generality over something like conditions. Thus the need for (21) is serious. At this point, the contingentist might reconsider second-order being constraints such as (14) and (15), since they were assumed in the objections to the stronger comprehension principle CompMC1. Even that principle yields (21) only when (20) is assumed, but the contingentist may argue that we can postulate (21) only when there is such a chemical as C, in which case (20) is (contingently) true. Denying the second-order being constraint is very implausible. If concrete individuals would have a given property (such as my anti-haecceity) in counterfactual circumstances, what more is needed for there to be that property in those circumstances? Moreover, the denial creates technical difficulties. We saw that with CompMC1 as comprehension principle, MI yields only conclusions conditional on first-order and second-order being assumptions, as in (12), the obvious way of discharging which is by firstorder and second-order being constraints. The contingentist might seek the benefits of MI without invoking the second-order being constraint by an alternative strategy. Divide applications of MI into those to actual circumstances and those to counterfactual ones. In the former case, these contingentists need the actual truth of the relevant being assumption EA, where A is the defining formula for the relevant instance of MI. They might argue that EA has some sort of a priori status, as a precondition for the content of A. In the latter case, there is no corresponding argument for the truth of EA in the relevant counterfactual circumstances. However, to reason about those circumstances, they might use instead of the necessitation of MI a modalized version of MI with the necessitation inside the second-order quantifier: MI□

X

□((X0 &  n (Xn → Xsn)) →  n Xn)

To apply MI□ to counterfactual circumstances with a property defined by the formula A, one requires only the actual being of the property, and so only the actual truth of EA, which the contingentist might justify as in the former case. One does not expect the application of a standard non-modal mathematical principle to counterfactual circumstances to need more than the

Metaphysics and Higher-Order Modal Logic

35

necessitation of that principle. But a more serious problem for the divideand-conquer strategy is that MI□ is still too weak to cover all the required cases. Here is an example: (23)

□  x ((Rx0 &  n (Rxn → Rxsn)) →  n Rxn)

Here the relevant property-defining formula is Rxn, where x is the parameter and R the predicate constant. The corresponding being assumption ERxn is  y x=y &  X □  n (Xn ↔ Rxn). But for contingentists the first conjunct of ERxn may fail in the actual circumstances, because the quantifier governing the parameter x is inside the modal operator in (23), so we are talking about what individuals there would have been in counterfactual circumstances. Equally, the second conjunct of ERxn may fail in counterfactual circumstances, since the relation expressed by R may not be in those circumstances. For all the contingentist is entitled to assume, in each circumstance one or other failure occurs. Thus the being assumption ERxn may be impossible. If so, CompMC1 does not even yield the possibility of the relevant property. Of course, the contingentist might add examples like (23) as new axioms in their own right. But that is to give up a principal advantage of secondorder formulations of mathematical principles, which is that they capture the full range of intended generality, rather than simply collecting together those instances that happen to be expressible in the language. Similar problems arise in non-mathematical contexts. For example, the form of contingentism now in view is not entitled to (24): (24)

□  y ◊  X □  x (Xx ↔ Sxy)

For the contingentist cannot exclude the counterfactual possibility of an individual (y) incompossible with the relation actually expressed by S. The argument that in such cases the property at issue would have been empty tacitly relies on the second-order being constraint, which the contingentist in question has rejected. Without principles like (24), second-order quantifiers lack the generality we intended to capture in second-order modal logic. Thus rejecting the second-order being constraint solves some old problems for the contingentist only by making new ones. We saw in section 2 that the prospects are poor for reconciling contingentism with the strong comprehension principle CompM for higher-order modal logic. In sections 3 and 4, we have seen that replacing CompM with

36

Timothy Williamson

weaker comprehension principles creates a tangle of further problems for contingentism. By contrast, CompM fits nicely with necessitism. Thus, on present evidence, necessitism is better placed than contingentism to integrate with the sort of comprehension principle that higher-order modal logic needs if it is to serve as a framework for general systematic inquiry. Note * The arguments of this essay will be developed in more detail in my projected book Modal Logic as Metaphysics. My interest in the topic was reawakened by giving a talk on it to the conference on Analytic Ontology at the University of Innsbruck in 1997, subsequently published as ‘Bare Possibilia’ in a special issue of Erkenntnis dedicated to the conference proceedings and co-edited by Edmund Runggaldier. Not only do we share many philosophical values; we were fellow students of philosophy at Oxford in 1973-77, although he was several years senior to me. I applaud his important role in the revival of analytic philosophy in Central Europe, and dedicate this essay to him on the occasion of his 65th birthday.

On Metaphysical Possibility – in Classical and Analytic Philosophy OTTO MUCK, Innsbruck

Toward the end of the last century, Lowe’s idea gained attention that the aim of metaphysics is not conceptual analysis but an investigation of metaphysical possibility (Lowe 1998 and 2008). This, together with experience, should help to develop a better knowledge about reality as a whole. Metaphysical possibility is distinct both from logical and from physical possibility. So it was asked how this could be known. Lowe wants to pay attention to the essences, which are the essences of entities, the objects of intention of our concepts. Lowe’s approach was regarded as an innovation, and with regard to the development of analytical metaphysics, there is something to this. But familiar with the tradition of classical metaphysics, one might say that it was always an aim of general metaphysics to investigate the most basic structures of being without which reality cannot have existence. Furthermore, a closer knowledge of concrete reality was left to the branches of special metaphysics which try to understand special realms of reality by taking into account appropriate experience. However, within this tradition during the last century, the question of how to know what is metaphysically possible found different proposals, some of them even associated to the developments in analytical philosophy. It might be useful for the present discussion – in my opinion – to summarize some of the most interesting tendencies. 1. Metaphysical possibility according to E.J. Lowe First of all for Lowe, metaphysics is not simply an assembly of irrational opinions. Further, according to Lowe, the importance of metaphysics does not concern merely the consideration of particular aspects of reality, but it has to investigate the compatibility of these parts, that is, to consider reality as

38

Otto Muck

whole. Thus, the task of metaphysics cannot be fulfilled by the special sciences alone. For Lowe, the aim of metaphysics is not reached only by describing how we speak about these various realms in life or how we describe common beliefs about the world, both of which are intended by descriptive metaphysics. In addition, formal semantics, which develops interpretations and models for the conceptual consistency of formalized theories, is insufficient for metaphysics. Formal ontology, according to Lowe, does not care primarily for logical form as formal semantics does. It cares for ontological form. Knowing the nature or essence of a (possible) kind of being or entity cannot be reduced to knowing the meanings of words or understanding concepts and knowing logical relations between them. (Lowe 2008, 33).

The attention to essences reveals differences between entities in what regards their possibility to exist, and it thereby leads to different categories of entities. Lowe explained this in his Four-Categories-Ontology. The categories make explicit that metaphysical possibilities are the ultimate underpinnings for our judgments about the things which we meet in experience. To gain knowledge of the things which actually exist, experience is required, but the framework, which is used for this task, is to be clarified by investigating the metaphysical possibilities, especially the basic categories of being. In defending his approach, Lowe uses, among others, the following types of arguments: Considering the essences of the things we speak about should by no means be regarded as impossible. Even in semantics, we use an understanding of essence when one speaks of concepts and representations. Both are beings and have their respective essences. So, essences are also objects of the language of semantic investigations. It is interesting for me here that Lowe uses an argument which some would see as part of an operational analysis which makes explicit an understanding of a term, here of the meaning of “essence”, by pointing to the linguistic operation in which it is used. And, then, this is combined with an indirect argument in order to defend the possibility of using our understanding of essences, at least as an argument ad hominem.

On Metaphysical Possibility – in Classical and Analytic Philosophy

39

2. Questions and observations arising from Lowe’s intentions In order to resume the ideas of Lowe and discuss their merits, I will use the concept of a model. When using a model (e.g., a road map), its purpose and its specific advantages must be kept in mind – as well as the limits of its representation of reality. Call these the conditions of adequacy of an explication or an interpretation. Models, therefore, are usually only partial interpretations of the field which they should help to understand. Their adequacy depends on the purpose for which they are used. Therefore, in using a model one should be careful not to draw unjustified conclusions. For example, road maps provide reliable information about reality if I want to find my way. But it would be a misunderstanding to scold them for being unrealistic as they do not present the beauty of the landscape. When Lowe contrasts his own metaphysical approach to several trends in Analytic Philosophy, this does not mean (so I take it) that he would not appreciate the merits of Analytic Philosophy in the last century. But Lowe locates ontological questions beyond the boundaries of semantic models of modal logic. The task of investigating metaphysical possibility cannot be achieved by formal semantics alone. The semantic approach to systems of modal logic is useful to show the consistency of these systems by presenting abstract interpretations, which satisfy the formulas of the systems. These semantic models show the logical consistency of the systems. But the fulfilment of this minimal requirement for formal systems does not provide further ontological conditions for the beings about which we speak when we use concepts, which refer to objects of experience. For Lowe, the model of reality used in everyday life has its limits as well, and metaphysical analysis should help to transcend them. For him, descriptive metaphysics deals only with the common sense model of reality, as I would call it. This model sees reality only with regard to practical purposes instead of caring for reality as such. According to Lowe, this bears a similarity to neo-Kantianism and to the mistake of idealism: they are all satisfied with the way how we think about reality and do not ask about reality as it is. But is it true that the model of reality which we use in life, in our “lifeworld” (Lebenswelt), is unable to contribute to a better knowledge of reality? I want to point out some methodological elements which this model can contribute to develop metaphysics.

40

Otto Muck

2.1 The “Life-world” as the frame of reference for questions to be solved by metaphysics There is no doubt that our everyday model is imperfect in interpreting reality. Nevertheless, it represents certain traits of reality – otherwise it could not provide reliable information for many tasks in our life, which it obviously does. This leads to the further question of how much assistance our pre-philosophical language can provide for a critical appraisal of reality. The model which directs our everyday life, our “life-world”, should be regarded as the benchmark for the material adequacy of metaphysical proposals. Metaphysics is expected to assist us in solving questions that arise in everyday life. Are solutions to these questions supported by the considerations of metaphysics? If metaphysics should improve the model of reality which we are using in life, this should help to find answers to the questions arising there. Therefore, testing the sufficiency of metaphysical ideas has to consider the consequences for our life. This does not entail an uncritical acceptance of opinions of our common sense model of reality – this is not even the case in descriptive metaphysics according to Strawson (Löffler 2007, 123-130). I want to take the occasion of this paper to express my high estimation for Edmund Runggaldier’s way of dealing with philosophical problems. On the one hand, his many publications display his openness to the manifold proposals of how to handle philosophical problems, including formal ontology. On the other hand, he constantly considers how proposed ideas can be solutions for the field where the problems had their origin – our lifeworld. 2.2 The concepts which refer to beings with their essences have been produced in the formation of our life-world In order to have access to essences, we rely on the process which has led us to the formation of our everyday-model of reality. In that vein, Scott A. Shalkowski (2008, in response to Lowe 2008) raises an obvious question: How do we obtain concepts of the entities which contain those essential features which (according to Lowe) should be the proper object of attention in metaphysics? For Lowe, insight into essences is a condition a priori for empirical knowledge, which is formed a posteriori. But one should note that the formation of concepts is linked with the formation of experiential knowledge. So the “aprioricity” of essences for empirical knowledge

On Metaphysical Possibility – in Classical and Analytic Philosophy

41

should not be misunderstood in a temporal way. Therefore, the concepts used in the model of everyday knowledge (its need for critical improvement not withstanding) can be expected to give us sufficient acquaintance with the essences for the purpose of formal ontology. 2.3 Understanding of intended essences is manifested in relations between concepts But how tight is the link between ordinary language and reality? On the one hand, the practical interests of everyday life in the process of original formation of language seem to distort our relation to reality, and on the other hand, these interests must take into account the structures of reality. Is ordinary language thus distorting or rather revealing reality? Recently, Paul Symington (Symington 2010) has pointed out different views on this question in medieval philosophy. According to Symington, Thomas Aquinas considered the expression of our relation to the intended reality as strong enough to be helpful for determining differences in reality. This is done by reflecting on expressions in language. Modes of per se predication indicate how a predicate is related to its subject. These different modes reflect relations of the essences intended by the concepts. They are not merely a matter of our insufficient understanding of reality. Symington shows how, following this way, one can provide a reconstruction of Lowe’s four categories but at the same time overcome some difficulties which they seem to include (Symington 2010, 134-139). 2.4 The Life-world grounds our operations and interests for investigations in special realms We should not forget that all our attempts to develop a critical view of reality rely on ordinary language – it has at least the function of indicating problems and drawing our attention to ways of solving them. Methodical constructivism shows how this can be done, e.g., by an operational approach to logic and to the foundations of mathematics and the empirical sciences (Lorenzen 1969). This attention to the roots of specialized investigations and formulations helps to understand their power as well as their limits. It also highlights the relevance of our theoretical investigations for our relation to reality in everyday activity. C.F. Gethmann labels this approach of methodical construc-

42

Otto Muck

tivism as “phenomenology after the linguistic turn” (Gethmann 2010, 23). He shows how our insight into essences has its consequences in our practical life, especially in the performance of language. This use of ordinary language (i.e., to draw attention to our activity) can also be useful for the task Lowe is interested in, namely, the explication of what goes on when we gain insights regarding essences. In this way, insights which are operative in the performance of these activities could be explicated critically. This may help to avoid the danger of burdening the formulations of these insights with misleading associations. 2.5 The Life-world also uses an understanding of essences in their relevance for our active occupation with reality How does the practical engagement with reality in our life find consideration in the metaphysical quest for reality? This should not be neglected as an adequacy condition for metaphysical proposals. It seems to me that Lowe’s interest is unduly restricted to the static possibility of things with regard to existence and identification. But wouldn’t it also be a task of ontology to pay attention to the dynamic possibility of coming into existence? It seems that the experience of action (and of the causal interdependence with reality involved in it) reveals that there are presupposed a priori conditions for the possibility of coming or bringing to existence. Why not also treat them in ontology? Anyway, the dynamic aspect of coming into existence has several aspects which deserve attention, since misunderstanding them may lead to failures in action. (Examples are given by Aristotle, who paid much attention to the different aspects of explaining how beings come into existence, change and vanish.) So reality and being are not to be understood just as the actualisation of a possibility for existence. The existence of a being is essentially connected with the activity of beings, as our life proves. Therefore, the conditions of metaphysical possibility should also cover the conditions for coming into existence by action. 2.6 The Life-world is related to the personal world-view, which integrates our life One might ask whether descriptive metaphysics is restricted to the opinions and presuppositions present in the common understanding of language

On Metaphysical Possibility – in Classical and Analytic Philosophy

43

users. This seems to be the case in Strawson’s Individuals (Strawson 1959) and in Stephan Körner’s “immanent metaphysics” (Körner 1984). They both rather reflect the commonly shared system of rationality (Hudson 1991), not the concrete world-view of a person. Lowe would agree that metaphysics should not only be interested in one aspect of reality but in reality as a whole. Therefore, I think, metaphysics should help to clarify the personal outlook on life as it is used in the concrete actions and decisions of persons. This task goes beyond analyzing the beliefs immanent in the common understanding of language. It concerns beliefs of single persons. In the investigation into the logic of religious language, which started with A. Flew (Flew 1963, cf. Muck 1967), this stock of beliefs was sometimes called Weltanschauung (world-view). Körner would have labelled it as a person’s “transcendent metaphysics” (Körner 1984, ch.11 and 17). In our search for a critical account of the “reality” which a personal world-view intends to understand, the reference to commonly accepted structures is only of limited importance: The reason is that such common sense structures, when used by a person for leading her life, underlie personal evaluation. Furthermore, the common sense model extends only to beliefs which find concordance among people. It never has the universality of a person’s unlimited quest for reality which is exemplified in the questions this person could ask. The personal world-view depends on experience, including personal experience. Nevertheless, one may also ask for minimal a priori conditions for personal world-views, which are based on personal experiences of life. The discussion about personal world-views continues to shed light on formal structures and minimal requirements for a reasonable improvement of personal world-views, especially by way of dialogue (cf. Ferré 1961, Bochenski 1965 and 1994; Riordan 1991). Due to their basis in personal experience, personal world-views have an essentially personal element. Nevertheless (or just because of that!) a dialogue between world-views is possible and fruitful; it presupposes respect for the differences and involves the exchange of personal reasons and experiences. Therefore, the realm of admissible questions in metaphysics should not be limited a priori to questions with generally accepted answers. Metaphysics, therefore, should also direct our attention to the a priori conditions of forming and rationally improving personal world-views,

44

Otto Muck

which involves personal experience. In a similar vein, Karl Popper described this as the task of philosophy: We all have our philosophies, whether or not we are aware of this fact, and our philosophies are not worth very much. But the impact of our philosophies upon our actions and our lives is often devastating. This makes it necessary to try to improve our philosophies by criticism. This is the only apology for the continued existence of philosophy, which I am able to offer. (Popper 1973, 33)

3. Some tendencies among philosophers who try to renew traditional metaphysics Around the beginning of the 20th century, many philosophers who were interested in metaphysics welcomed the phenomenological movement as a “turn toward the things” instead of getting lost in critical reflection. This again raised the question whether such a turn could lead only to essential relations between intentional contents, or whether it was apt to discuss questions of existence. In his development of phenomenology, Martin Heidegger created the idea of a fundamental ontology. But from the viewpoint of logical empiricism, this was not an encouragement to do metaphysics! As an alternative way to achieve better knowledge about the things as they are, some philosophers favoured “inductive metaphysics”: they regarded metaphysics as the fruit of further hypotheses that reach beyond scientific hypotheses. Seen that way, metaphysics appears as a sort of very general empirical discipline. Against this approach (and inspired by Pierre Duhem and Henri Poincaré) Jacques Maritain emphasized the methodological distinction between metaphysics and empirical sciences – a distinction which had already been explicated by Thomas Aquinas, as Pacificus Borgmann pointed out in his critique of the scientism behind inductive metaphysics (Borgmann 1934). This distinction is associated with the difference between inductive and deductive arguments and corresponding types of explanation. The former start from experience and aim at describing its sufficient conditions, using general nomological statements. These statements permit forecasts and can hence be tested empirically. The latter start from experienced beings and apply to them some general principles which require that some other being exists. These necessary conditions make the existence of the observed being ontologically possible.

On Metaphysical Possibility – in Classical and Analytic Philosophy

45

Such an explanation derives its strength from an assumed general principle that states that the actuality of a being with special features presupposes the actuality of a being of a certain kind. Such principles are not conceptual necessities but express general truths concerning the world of human experience. This obviously raises a question: how do we get to know such principles? (Muck 1994, 47-48) The critical discussion of different forms of epistemology in the first decades of the last century lead some philosophers to see the foundation of human knowledge in an immediate evidence of necessary connections of the contents of our intentions. A much discussed example was the link between essences, referred to by ‘contingent being’ and ‘existing in virtue of a proportionate efficient cause’, as this was defended in the thirties, e.g., by Josef de Vries (de Vries 1937). An obvious difficulty of that approach was identified in the personal character of such immediate evidence. A further problem was seen in the adequate linguistic expression of it. So this procedure was challenged by hermeneutical considerations, pointing to the dependence of the verbal expression of the immediate insight on personal and cultural conditions. Analytic philosophy challenged it by the considerations concerning the problems connected with basic experiential sentences (see the debate in the Vienna Circle about the nature of protocol sentences) and of private language. To meet such problems, Joseph Maréchal’s (1878-1944) approach to develop metaphysics exploits the idea that the basic insights of metaphysics could already be operative in the development of human knowledge. In our reflection upon this process, these basic insights are to be articulated as conditions of the possibility of this knowledge. This bears a similarity to Shalkowski’s above mentioned consideration that Lowe’s approach should better rely on concepts developed in such a process. Although these concepts contain essences and are a priori insofar as they make experience possible, they are not explicitly formulated in advance of experience. They are grasped as conditions of metaphysical possibility of the things we can find in experience. Since Maréchal asks for the conditions of the possibility of human consciousness of objects, he finds an initial similarity between his and Kant’s method (Maréchal 1922-1944). But he tried to show how this transcendental approach can be further developed. According to Maréchal, Kant’s restriction of knowledge to the realm of appearances/phenomena can be

46

Otto Muck

transgressed towards the unlimited field of being. This field of being is relevant for our activity, for our contact with actual reality. Maréchal’s thought had an impact on the philosophies of Johannes Baptist Lotz (19031994), Karl Rahner (1904-1984), Emerich Coreth (1919-2006), Bernard Lonergan (1904-1984) and many others. Their work can be understood as emphasizing different aspects of this transcendental-realist approach. Coreth, e.g., starts from reflecting upon the activity of asking questions. He critically explicates the minimal conditions of the possibility of this performance. The meanings of the concepts involved in that explication are determined in relation to the performance of asking as such. He calls this procedure a “dialectic between concept and performance (Begriff and Vollzug)” (Coreth 1968). In that way, Coreth accomplishes an operational reinterpretation of classical metaphysics without relying on an obsolete cosmology. A comparative look at the procedures of these philosophers reveals some basic methodological elements (Muck 1968, 307-342) which have also been further explicated in the light of 20th century analytical philosophy (Muck 1969, Runggaldier 1990). Maréchal himself had already pointed to Aristotle’s defence of the legitimacy of the principle of non-contradiction. For Aristotle, this principle belongs to the basic structure of a cognizing subject. It is immediately evident and indispensable for anybody who speaks and communicates. So the content of such a sentence is (a) claimed to be evident and is (b) shown to be operative in any intellectual action. Furthermore, this fact is (c) reinforced in an indirect manner, reminding us that it is presupposed in any communication. Later on, element (b) was identified as operational analysis (Hoenen 1954) and (c) as indirect argumentation in the form of retorsion (Isaye 1954). Retorsive arguments show the inconsistency between the content of the negation of the principle and the performance of the very act of negating this principle (Muck 2009). Because such a principle expresses an operational structure, the reflection on this structure is the way to determine its meaning and its practical consequences. This reflection does not depend on a special language. In his book Insight (Lonergan 1957), the Canadian philosopher Bernard Lonergan (1904-1984) takes up and develops the approach of Maréchal in his own way (Muck 2010, 1424-1425). Lonergan investigated the details of the process leading to knowledge, and he did this in two steps: At the first stage (a phenomenological analysis of the process of knowing), he tried to elucidate what we are doing when we strive for knowledge. He observed a

On Metaphysical Possibility – in Classical and Analytic Philosophy

47

recurrent operative structure of our mental process in all different forms of knowing: starting from the experience of data, we try to understand them, and then we attempt to affirm what we have understood. For that purpose, we return to the data again until there are no further questions which are relevant for our understanding and our affirmations. In a second stage of his investigation, he asks what is attained when we engage in that activity. The answer to this question arises not from an uncritically assumed model of cognition, but from the immanent operative structure of knowing which is manifested in the activity of knowing. By a kind of retorsion, Lonergan argues that there is no knowledge without that recurrent operative structure of cognitional activity he had observed in his phenomenological analysis. (The same, by the way, holds also for the reflection upon this activity itself.) He then shows that one cannot in principle deny the objective validity of the results of that process. Doing this, he clarifies the core meaning of “objective validity” in terms of the operative structure itself. For Lonergan, the operative structures are not themselves knowledge, but they are the condition of the possibility of knowledge and of its organization. The expression of our awareness of these conditions might be dependent on language and culture, but the operative meaning of them has intercultural importance. The variety of different directions of investigation, which has led to the development of different sciences with their own principles, has its roots in this process, too. For Lonergan, these principles are heuristic structures for dealing with the various questions which life prompts. Those structures are involved in the acts of special questions and investigations as their presuppositions. Lonergan’s account of the operative structures of knowing and of its relationship to retorsive justification helped to make clear the logical structure of Maréchal’s transcendental method with its operative and retorsive elements. It also helped to make explicit and to justify the claim for objective validity underlying human knowledge. It does so by providing an operational criterion for affirming knowledge as valid, namely, its stability against all relevant questions (Muck 1969).

Can the indicated procedure help to improve our personal outlook on the world and the ways we speak about it? Is a useful combination of the indicated methodological elements possible which might be fruitful for the task of metaphysics?

48

Otto Muck

4. Suggestions for metaphysical analysis If metaphysics should assist us in coping better with questions that concern the beliefs which underlie the conduct of our life, then it seems natural to pay special attention to difficulties which appear as puzzles or antinomies. Their critical appraisal starts against the background of our previous cognitional practice, which was leading us to the personal model of reality with which we have lived so far; and against that background, we also discover those conflicting sentences which prompt our questions for truth. Let us start with the experience that we are sometimes confronted with opinions which seem to constitute reliable beliefs but oppose each other. Examples could be beliefs we hold, or experiences which seem to contradict well-founded expectations, or surprising statements of people whom we regard as reliable and serious in their views. Such experiences can lead to a process of critical appraisal of the conflicting beliefs. In my view, such processes can be understood as exemplifications of Maréchal’s and Lonergan’s ideas. The process of a critical appraisal of our often insufficient grasp of reality may start with the formulation of oppositions between beliefs both of which seem to be justified. My discussion of a simple example is intended to elucidate the structure of a process that could be called “integrating synthesis”. Lowe seems to intend something similar when he writes: Perhaps the most fundamental principle governing truth is the principle of noncontradiction, according to which a proposition cannot be both true and false – or, as it might more strictly be put, a proposition and its negation cannot be both true. (Lowe 2003, 121) […] reality is fundamentally one: that there is just one sum total of existence – one world – which is the same for all thinkers, places and times. And my suggestion is that, to the extent that we are committed to the unity of truth, at least in as much as this amounts to an unconditioned acceptance of the principle of non-contradiction, we are committed, in short, to realist metaphysics. Fortunately, this still leaves plenty of scope for many forms of pluralism [...].Our ontology will admit of multitudes within The One”. (Lowe 2003, 123)

A possible start for such a process may be situations where we ask which of two or more conflicting sentences might be true. In order to give an idea of the methodologically relevant considerations which can make insights explicit, I will make use of a simple, old and rather innocent example. (Wolfgang Stegmüller also used it to draw the attention to the way how we

On Metaphysical Possibility – in Classical and Analytic Philosophy

49

speak about “reality”.) It is the situation that a stick which is partly dipped into water looks buckled. Obvious questions like “is the stick really buckled?”, “is it true that it is buckled?” may arise (see 4.1 below). Our reaction could be that we try to solve the questions by touching the stick (4.2) and, in recalling some elements of optics, to understand that the stick is not really buckled (4.3), but only appears buckled (4.4). What is involved in this understanding of the situation and our reaction? Let me sketch its epistemological and ontological implications. 4.1 The challenge of opposite statements – the implicit ideal of reliable cognition Firstly, we may become aware of the fact that our surprise reveals our tacit conviction that it cannot be true that (under the same circumstances) the stick is buckled and is not buckled. Furthermore, this can be understood as the application of an ideal norm for reliable beliefs: according to this norm, a reliable belief is one for which no question which is relevant for this belief is open. (A note on my terminology: A “relevant question for belief p” is a question that admits of at least one possible answer that is incompatible with p. And a relevant question is “open” if it has not yet been ruled out that one of those incompatible answers is the reliable answer.) This norm might be seen as a postulate for our cognitive activities if we come across oppositions and wonder what should be accepted. Comparing this postulate with our actual cognitive practice in everyday life, we have to add specifications. We usually cannot check all relevant questions and decide whether some of them are still open. In practical life we have learned to restrict the relevant questions to those for which we have a prudent doubt which is based on our experience so far. Put in another way, for such questions there is a danger of error and not just the bare possibility of error. Let us call these questions “positively relevant questions for p”. This restriction is a matter of prudence: subjects who are unable to prudently distinguish between possibility and danger of an error would not have survived up to now. This epistemic quality of beliefs, which we use in practical life, was called hypothetical certitude (by Josef de Vries) and practical certitude (by Karl Popper [1973, 78]): For such beliefs, the relevant questions are me-

50

Otto Muck

thodically restricted to the positively relevant questions. This certitude is hypothetical, as it is dependent upon the realm of experience which a person could have taken into consideration. It is also dependent on the beliefs which are involved in our judgements about the conflicting statements – including our beliefs about sound ways of arguing. 4.2 Attention to the support of the statements. An operational explication of “reality” and “truth” The same hypothetical character also affects our beliefs in the model of reality which has guided our life so far and which had also been used as support of the conflicting statements. This can be seen as the reason for the limitations of this model. These limitations become manifest when we meet inconsistencies with other investigations or opinions – we then begin to develop a closer interest in reality by gaining a critical distance to the beliefs we use in practical life. We then pay critical attention to the support for our beliefs in order to find out whether these beliefs have perhaps transcended the limits of their hypothetical validity within the model we used. So we ask for the truth of our beliefs, and hence, we quest for reality and being and not merely our opinions about it. But what do we understand by “truth”, “reality” and “being” in such a context? We may first note the parallels between the use of “reality” or “being” and the use of “truth”. This makes us aware of a use of “true” which requires that the state of affairs represented by the proposition is real and not fictitious. Or, as Aristotle put it, a statement is true when it says what is of that which there is. Next, we may consider that the question for truth arose when relevant questions beyond the recognized limits turned up. This insinuates that the truth of a proposition is a minimal condition for the reliability of corresponding beliefs. When the support of a belief leads to practical certitude we may propose to say the following: within the (hypothetical) limits of the positively relevant questions, and for the practical purposes of ordinary life, we can trust in this belief as if it were true. Or in other words: within these limits, the model we use for our ordinary life can be taken as a model that matches with reality. But becoming aware of the limits and posing questions beyond these limits, e.g., questions for truth and reality, we distinguish between the reality according to the model and reality as it is. This corresponds to the

On Metaphysical Possibility – in Classical and Analytic Philosophy

51

distinction between positively relevant questions and all relevant questions. This distinction may at first glimpse appear as a merely epistemological matter. But it should also be seen in its ontological importance. In our buckled-stick-example, we distinguished between how “it appears” and what “really is the case”. This correlates with the difference between restricting the relevant questions to a limited range and considering all relevant questions without limitation. Hence, statements for which the relevant questions are not limited are intended as statements about what really is whereas statements about how things appear express that this holds only for a restricted realm of relevant questions. It is important to keep this distinction in mind because even when we only describe appearances, we use a language by which we usually intend to say what there is. So the restriction must be made explicit if it is not just obvious from the circumstances. If a restricted affirmation only intends to describe the content of appearance, this “brackets” the ontological value suggested by ordinary use and leaves it open to further specification. Since such a specification would consist in considering further relevant questions, this shows the ontological priority of the realm of being compared with that of appearance or other methodically restricted scopes of our claims. This should also be kept in mind when evaluating the familiar “realist” objection that a reflection on necessities of thinking cannot gain any knowledge about reality. But this objection relies on a common sense understanding of “reality” which is in need of critical improvement. The operational criterion of validity relies on the performance of activities that are also exercised in formulating the very objection, i.e,. by asking, affirming and reflecting on our activities. These activities do not only work within our every-day model of reality; they are also involved in critically reflecting it. Therefore, our understanding of “reality” should not be restricted to that model of reality which is to be improved. It should be formed according to the more general activity by which we are enabled to scrutinize the common sense model. And this understanding has been developed in the given example of an integrating synthesis.

52

Otto Muck

4.3 Understanding the initial problem – improving our experiential knowledge Hence, if we are intending to talk about “reality” or “being”, this involves statements with an unrestricted domain of relevant questions. Of course, such claims are again open to relevant questions, in the sense that they can be criticised and undergo further development. (I stress this since metaphysical statements, like any other statements, are in principle open to revision.) The example above suggested a cautious use of “reality”. Similar caution is called for regarding words like “being”, “appearance”, etc., which are used to formulate appropriate differentiations. They determine a horizon of questioning, but they do not yield the answers. Therefore, they can be called “heuristic concepts”. They would be misunderstood if they were associated with special features which are borrowed from an inadequate model of this realm. But they may imply elements which are a priori conditions under which something can possibly be an object within this horizon of questioning. But thereby it is not decided in advance what will be detected here in detail. Asking detailed questions is the way in which reality becomes present – or at least aspects of reality, corresponding to the horizon. So the procedure proposed here is not to be misunderstood as a kind of idealism, and it does not just lead to our construction of reality. What it shows is the way how reality becomes present in prompting and answering our relevant questions. This is the way how beings like us can become acquainted with a reality to which they stand in conscious and active relations. In our example, one statement of the initial opposition (namely the claim that the stick is buckled) is refuted by further experiments with the stick, e.g., by touching it, and by explaining the visual impression of a buckled stick via optical laws. As a consequence, the opposing statement “the stick is not buckled” can be reaffirmed, whereas the former one can be reaffirmed in a revised way: “it just appeared as if the stick was buckled”, thereby restricting the range of relevant questions. I would call this a critically positive interpretation: It is an interpretation since it is a reconstruction of the meaning of a statement. It is a positive one since it emphasises what can duly be affirmed in view of the given support. And it is critical, since it adjusts the meaning to the limits of supported validity. Thereby it

On Metaphysical Possibility – in Classical and Analytic Philosophy

53

improves our apprehension of reality. By this consideration an ontologically relevant differentiation became manifest. 4.4 Solution of the problem – integrating the criticism of its assumptions into our way to apprehension of reality The process by which critically positive interpretations are developed may be called an integrating synthesis. It elaborates the a priori conditions by which the limits of our knowledge can be understood, and it elucidates what “validity” means within these limits. It thereby develops elements of a basic structure of reality, and those elements in turn delineate what can be known in detail by further experience in the various realms of questioning. At the same time, this synthesis distinguishes a priori elements, which are necessary conditions for any representation of reality (call them pure a priori), from merely methodological or even subjective a priori conditions, which result both from particular personal interests or from a limitation for practical purposes dependent on previous experience. To identify limitations imposed by such special forms of a priori conditions is an important result of the process of integrating synthesis. Chief occasions for becoming aware of such limitations are oppositions which become apparent in a dialogue. This fact suggests that diligence in pursuing special ways of investigation must be combined with an evaluation of their results: Only such an evaluation can determine what they can contribute to our orientation in life. Interesting examples of attempts to finding an integrating synthesis of opposing views can be found, e.g., in the ongoing discussions about the relations between science and religion or between naturalist and non-naturalist worldviews (Jacobs 2008). To summarize, I have tried to take up Lowe’s suggestion to go beyond mere opinions about reality and to turn to reality, to the essence of beings, by investigating minimal conditions for the possibility of the existence of beings. Recalling similar endeavours in analytical and classical metaphysics, I tried to indicate ways how metaphysics should assist our understanding of the connection between our life-world and the underlying reality. I also suggested to make the understanding of essences explicit by means of analyzing activities which make use of these insights.

54

Otto Muck

5. References Bochenski, J.M. 1965 The Logic of Religion, New York: University Press. -

1994 “Religious Hypothesis Revisited”, in P.Weingartner (ed.), Scientific and religious belief (Philosophical Studies Series 59), Dordrecht: Kluwer, 143160.

Borgmann, P. 1934 “Gegenstand, Erfahrungsgrundlage und Methode der Metaphysik”, Franziskanische Studien 21, 80-103 and 125-150. Coreth, E. 1968 Metaphysics (English edition by J. Donceel, with a critique by B.J.F. Lonergan), New York: Herder and Herder. De Vries, J. 1937 Denken und Sein. Ein Aufbau der Erkenntnistheorie, Freiburg: Herder. Flew, A. (ed.) 1963 “Theology and Verification”, in A. Flew and A. MacIntyre (ed.), New Essays in Philosophical Theology, London: SCM Press Ltd. Gethmann, C.F. 2010 “Die Aktualität des Methodischen Denkens”, in: C.F. Gethmann and J. Mittelstraß (ed.), Paul Lorenzen zu Ehren (Konstanzer Universitätsreden), Konstanz: Universitätsverlag Konstanz, 15-37. Hoenen, P. 1954 “De definitione operativa”, Gregorianum 35, 371-405. Hudson, W.D. 1991 “The Rational System of Beliefs”, in A. Loades and L.D. Rue (ed.), Contemporary Classics in Philosophy of Religion, LaSalle: Open Court, 33-58. Isaye, G. 1954 “La justification critique par retorsion”, Revue philosophique de Louvain 52, 204-233. Jacobs, Ph. 2008 “An argument over ‘methodological naturalism’ at the Vatican observatory”, Heythrop Journal 49, 542-581. Körner, St. 1984 Metaphysics, its structure and function, Cambridge: Cambridge University Press. Le Poidevin, R. (ed.) 2008 Being: Developments in contemporary metaphysics, Cambridge: Cambridge University Press. Löffler W. 2007 “Über deskriptive und revisionäre Metaphysik”, in Lutz-Bachmann & Schmidt 2007, 114-131. Lonergan, B.J.F. 1957 Insight. A Study of Human Understanding, London etc.: Longmans, Green & Co. Lorenzen, P. 1969 Normative logic and ethics (BI Hochschultaschenbücher 236), Mannheim – Zürich: Bibliographisches Institut. Lowe, E.J. 1998 The Possibility of Metaphysics: Substance, Identity and Time, Oxford: Oxford University Press. -

2003 “Metaphysical Realism and the Unity of Truth”, in A. Bächli and K. Peters (ed.), Monism, Frankfurt: Ontos, 109-123.

On Metaphysical Possibility – in Classical and Analytic Philosophy -

55

2008 “Two notions of Being: Entity and Essence”, in R. Le Poidevin (ed.), Being: Developments in contemporary metaphysics, Cambridge: Cambridge University Press, 23-48.

Lutz-Bachmann, M. and Schmidt T.M. (eds.) 2007 Metaphysics Today – Problems and Prospects, Freiburg i. Br.: Alber. Maréchal, J. 1922-1947 Le point de depart de la métaphysique. Leçons sur le développement historique et théorique du problème de la connaissance (Museum Lessianum), Bruxelles: Desclée de Brouwer. Muck, O. 1967 “Zur Logik der Rede von Gott”, Zeitschrift für Katholische Theologie 89, 1-8. -

1968 The transcendental method (english ed. by W.D. Seidensticker), New York: Herder and Herder.

-

1969 “The Logical Structure of Transcendental Method”, International Philosophical Quarterly 9, 342-462.

-

1994 “Assumptions of a classical philosophy of God”, Milltown Studies 33, 37-50.

-

2009 “Operative Analyse und Retorsion” in G. Kamp and F. Thiele (eds.), Erkennen und Handeln (Festschrift für C.F. Gethmann), München: Wilhelm Fink, 19-39.

-

2010 “Transcendental method”, in ACPI Encyclopedia of Philosophy, Bangalore: Asian Trading Corporation, vol.II, 1419-1425.

Popper, K.R. 1973 Objective Knowledge, Oxford: Clarendon Press. Riordan, P. 1991 “Religion as Weltanschauung”, Aquinas 37, 519-534. Runggaldier, E. 1990 “The transcendental method of metaphysics of Coreth and Muck and its relation to Analytic Philosophy”, The Asian Journal of Philosophy 2, 57-69. Shalkowski, S.A. 2008 “Essence and Being”, in Le Poidevin 2008, 49-63. Strawson, P.F. 1959 Individuals. An Essay in Descriptive metaphysics, London: Methuen. Symington, P. 2010 On Determining What There is. The identity of Ontological Categories in Aquinas, Scotus and Lowe (EIDE Foundations of Ontology 2), Heusenstamm: Ontos.

Beyond the Ultimate Methodological Atomism, Methodological Holism, and Explanations LUDGER JANSEN, Rostock

1. Explanation and Reference In this paper I want to discuss two contrary claims about ontological restrictions in explanations: methodological holism and methodological atomism. Methodological holism is the claim that only holistic explanations can be accepted, i.e. such explanations, which refer only to wholes and their properties rather than to parts. Methodological atomism, on the other hand, claims that only atomistic explanations can be accepted; i.e. such explanations, which refer only to the smallest and thus most fundamental parts involved in a phenomenon. I will argue against both of these claims, thereby attacking their common tenet, namely that explanations have to refer to the ultimate, be it the ultimate micro-structure or the ultimate macro-structure. I will discuss how explanations without any of these restrictions may look like, arguing that not only common sense explanations fit to this liberal pattern, but, contrary to appearance, also our scientific explanations. That the smallest parts are most fundamental is in itself, of course, a substantial assumption. Robert A. Wilson has called this assumption “smallism”, and he describes it as a form of discrimination, namely as “discrimination in favour of the small, and so, against the not-so-small”: “Small things and their properties are seen to be ontologically prior to the larger things that they constitute, and this metaphysics drives both explanatory ideal and methodological perspective. The explanatory ideal is to discover the basic causal powers of particular small things, and the methodological perspective is that of some form of reductionism.” (Wilson 2004, 22)

In what follows, I use quite a broad concept of explanation. An explanation of an event or state of affairs E is a sentence or a speech act giving certain pieces of information, in the light of which E appears to be either normal, or probable, or necessary. By “holistic explanation” I mean an explanation

58

Ludger Jansen

which refers to properties of wholes rather than parts. In many fields of research there are debates between methodological holists and methodological atomists. For example, there is a lively debate about social explanations: Can we explain society by looking at the people (and their psychological apparatus) only? Aren’t sociological wholes like groups, institutions, organisations, movements and nations influential powers on their own and therefore indispensable in sociology? Similar questions arise in biology: Can we really explain life by splitting living beings into a lot of molecules? Don’t we have to look at an organism as a whole? To put it into the words of the biologist Paul Weiss: “[...] after all, organisms are not just heaps of molecules. At least, I cannot bring myself to feel like one. Can you?” (Weiss 1969, 400; cf. Wuketits 1989) Another example is provided by physics: Doesn’t quantum mechanics tell us that the world is one big whole, and anything in it is connected with any other thing in it? Isn’t this splitting-up into atoms and stuff simply out of date? And last, but not least, history: Do we understand history, if we know how some people in the middle ages felt or what Queen Victory did? Don’t we have to know the sense of all and what is the big end of all of history? Methodological holism, as I said, is a claim concerning explanations. It is the view that explanations have to be holistic to be acceptable. This is not by itself an ontological position. The methodological holist’s motivation, however, derives from ontology: We are to refer to wholes in our explanations, because they are the fundamental entities the world consists of, and thus the only entities that are causally efficacious – or so the argument goes. Thus, methodological holism is typically motivated by a kind of ontological holism, which is often characterized by – or based on – statements like the following (Bunge 1979, 39-40): (H1) (H2) (H3) (H4)

A whole precedes its parts. A whole acts on its parts. A whole is more than the sum of its parts. A whole emerges under the action of agents that transcend both the actions among the components and the environmental influences. (H5) A whole can not be explained by analysis. Mario Bunge, from whom this enumeration is taken, includes a sixth claim: “A whole is better than any of its parts.” (Bunge 1979, 40) As it

Beyond the Ultimate

59

stands, this statement needs a lot of analytic endeavour in order to make sense out of it. But even if it turns out to be true, I do not see how such a value judgement touches on the validity of explanations. Hence I leave this statement out in what follows. By combining any of these six statements we can construct different variations of ontological holism.1 I want to be noncommittal about how many wholes there are – a Spinozian type of holism would claim that there exists exactly one whole, namely Deus sive natura. But the arguments discussed in this paper apply also to types of holism allowing more than one whole. The aim of this paper is to show that none of these six statements yields any restriction to the entities our explanations refer to. None of these statements gives sufficient reason to prefer holistic explanations. In due course I will show that these statements are either false or have only a limited scope of application or fail to entail any restrictions for explanations. 2. Why methodological atomism cannot be practised But first I want to consider some motivations for a holistic position in methodology. Holism established itself as an opposition to what can be called “methodological atomism”. Methodological atomism holds that we understand, e.g., the movement of an object, if we can reduce this movement to the movements of the smallest and thus most fundamental parts of the object. Holism and atomism are not contradictory, but contrary positions: They may both be false, as also Bunge notes (1979, 41). And indeed I want to argue against both of them. Still, one of the best arguments for a holistic methodology is the impracticability of methodological atomism. I will now present some of the reasons, why atomism fails in practice. a. To begin with, methodological atomism presupposes the existence of smallest independent particles, and some may even hold that these minimal things are the only real things in the universe, all complex things being eliminable, reducible, or supervenient. This is: methodological atomism presupposes ontological atomism. And the latter might be false. As a matter of fact, what we today call “atom” in physics, is not a candidate for a smallest part which cannot be further divided. Whatever sub-atomic parti1

Röwer 1985, 6 holds even any of these claims to be a sufficient condition for holism.

60

Ludger Jansen

cles are discussed by today physics, it is questionable whether these particles are thought of as the smallest building blocks of the universe. In any case, physicists no longer believe in the independent existence of these sub-atomic particles: Quantum theory suggests that every point of the universe is potentially connected with any other point by Einstein-PodolskyRosen-correlations, which determine quantum phenomena happening at these points of space (cf. Esfeld 2002, chs. 7-8). Thus, modern physics gives us a reason no longer to speak about independently existing smallest particles. b. If there were smallest particles, there would be quite a lot. Carl Friedrich von Weizsäcker estimates that there are about 10120 “elements of reality”.2 These are far too many to calculate with. It should also be obvious that a map that displays all of these 10120 smallest objects would not be very useful. Moreover, it would not be possible to produce such a map within this universe. Even if we do not consider the whole universe: The number of the molecules in a certain amount of gas and the number of the cells in the brain are still quite large. A human body consists of something between 1013 and 1014 cells, and about 1011 of these are brain cells. Such numbers are still too big to be modelled by sets of equations or to be simulated by computers. And, as Stephen Clark puts it regarding history, even if “it were possible to analyse the behaviour of states wholly in terms of the behaviour of individuals, no historian has yet accomplished this analysis – for it is simply too complex” (Clark 1975, 140). c. The point comes out even stronger, if we think of the complicated and complex neural network existing between our brain cells. There may be so many different relations between the parts of a whole, that a complete survey is virtually impossible. Even if there are only few elements to be considered, the net of relations between them can be enormous. Even between two elements there can be a potentially infinite number of relations. A standard textbook example for this is the “intentional mirror” (Johansson 2

Von Weizsäcker 1974, 76 (“Elemente der Wirklichkeit”) . Cf. von Weizäcker 1974, 75-76: “Diese Urobjekte oder Urs haben das Grenzvolumen 10-39 cm3, innerhalb dessen für uns (wegen der Heisenbergschen Unschärferelation) keine eindeutigen JaNein-Entscheidungen über Vorgänge in dem Kleinstraum mehr möglich sind.” Drieschner calls von Weizsäcker’s Ur-hypothesis a “radical atomism” (1979, 183; 1992, 51). For the mathematical formalism cf. Drieschner 1979, 182-188.

Beyond the Ultimate

61

2004a, 276), which can be illustrated by, e.g., two persons A and B watching each other. Then we have: A is seeing that B is standing over there. B is seeing that A is standing over there. A is seeing that B is seeing A. B is seeing that A is seeing B. A is seeing that B is seeing that A is seeing B. B is seeing that A is seeing that B is seeing A. A is seeing that B is seeing that A is seeing that B is seeing A. B is seeing that A is seeing that B is seeing that A is seeing B. And so on, in infinity.3 We arrive at a potentially infinite number of contingent relations between only two elements. d. Here is another argument against atomism. In some cases it is quite useless to take small particles into focus. Russian scientist scrutinized Lenin’s brain closely, but without any results for a better understanding Russian history (Gregory 2007, ch. 3). Obviously, the neuronal structure of Lenin’s brain is of no help in explaining the October Revolution; these are just irrelevant facts for the historian. Speaking about irrelevant facts does, of course, posit that there is a distinction between relevant and irrelevant facts. Wilhelm Essler claims that as far as history is concerned there are indeed too many facts to be considered by the historian, but that there is no criterion to distinguish the important facts from the unimportant (Essler 1979, 159). On the other hand he thinks that this presents no difficulty for the natural sciences. However, I am afraid that we cannot ignore this problem for physics, chemistry and biology, either. It is true, scientist have a good routine in distinguishing between important and unimportant facts. But so do historians. There is no general criterion for distinguishing important facts from the unimportant. Essler suggests precision of measurement as a criterion. But this will not do, for which instrument is going to set the standards – yard stick or micrometer calliper? If we want to explain the chemical behaviour of a certain protein or the DNA, we have to take into account even the smallest twist or twill of the molecules. A nylon string has 3

Gilbert 1989, ch. 4 argues that such deliberations may even give rise to uncountably infinite numbers of relations. For a three-line recursive definition of the intentional mirror cf. Meggle 2002, 14-15.

62

Ludger Jansen

such twists and twills, too. But if we want to know something about its stability, these twists and twills seem to be quite unimportant.4 3. Why holism is not necessary It will now be my task to discuss the holistic statements (H1) to (H5). Up to now we have rehearsed several arguments, why an atomistic methodology is not possible. But if we cannot practise an atomistic methodology, methodological atomism must be false. For logical reasons, there cannot be a methodology which cannot be practiced, as much as it is an analytical truth that contradictory commands cannot be executed or that nobody can act according to norms prescribing something impossible. What, then, about methodological holism? It seems to be well motivated by the rejection of methodological atomism. We will now take a look at the statements by means of which I tried to characterize holism. The first task will be to clarify the meaning of the statements. I will do this by giving examples. Then our question will be: Do these claims impose any restriction on explanation? a. The first statement claims that wholes precede their parts. This can be understood in fundamentally different ways, three of which I will discuss here. The ambiguity of “precedes” has been noted since the time of Aristotle (Jansen 2002, ch. 6). For our purpose here it will be necessary to distinguish between temporal and logical priority, where the former may be priority in formation or in perception. We thus get the following claims: Formation Thesis: The coming-to-be of the whole temporally precedes the coming-to-be of the parts. Perception Thesis: The perception of the whole temporally precedes the perception of the parts. Logical Thesis: We cannot speak about the parts unless we (at least implicitly) speak about the whole. The formation thesis seems to be confirmed by many examples. In the ontogenesis of an animal there do not exist first head, legs and organs which are in a second move composed to a living being. Rather, the embryo can be considered as a whole which develops – by a process of differentiation – 4

I am indebted to Hans Primas for this example.

Beyond the Ultimate

63

legs, arms and organs. On the other hand, the formation thesis leads to strange results with other examples. Sociologists may regard groups as wholes. But of course, a school class is composed out of children who exist independently of and prior to their school class. And within the Gassendian model of gasses, a gas volume consists out of gas molecules that have existed prior to this very gas volume. These examples can also be brought up against the logical thesis of the whole. For people were able to talk about their children long before school classes were invented. Admittedly, the concept of gas molecules entered natural science much later than the concept of a gas volume, but it is not necessary that all gas molecules are members of gas volumes: Gas molecules can be solutes in, e.g., water, or they can be solitary molecules in the vast regions of interstellar space. We can distinguish these two groups of examples with the help of the concept of autonomous and non-autonomous parts. Something of kind F can be said to be autonomous, if its being F does not depend on being a part of some whole. If an F is autonomous, it may or may not be part of some whole; this does not matter to its being an F. Its being an F is not relative to a context. Hence, its being an F does not depend on any relations to other entities. Autonomous parts can thus exist independently of the whole they belong to: Even if a child is highly dependent on other people, it can exist without the school class in which it is being instructed. But a heart, which is a blood pumping organ, cannot normally exist without the body it belongs to. If a heart is to be transplanted to another human body, special liquids take care that the heart will be able to resume its function in the new body. Without a body, there cannot be a blood pumping organ, but only some lump of flesh. As Peter Simons puts it: It is only “a heart in the butcher’s sense, not the physiologist’s” (Simons 1987, 324). Thus, the formation thesis holds only for the special case of the existence of nonautonomous parts. And so does the logical thesis. The perception thesis is favoured, e.g., by Gestalt psychologists. Indeed, if we hear music, we do not hear ‘only’ a sequence of individual noises. We hear a ‘whole’: the tune. Similar things hold for visual perception. But this does not hold generally. Normal listeners cannot hear a symphony as a ‘whole’, rather than a sequence of different temporal parts of this composition. The human perceptive capacity is restricted: Even a rather small whole like, e.g., a human being cannot be perceived at once in his entire complexity, not to mention larger wholes up to the size of the universe. There are indeed many cases in which our perception of parts precedes

64

Ludger Jansen

even our concept of the whole. To draw an analogy, we can criticise Quine’s holism of language on quite similar grounds. As Fodor and Lepore say: “If [semantic] holism is true, then I can’t understand any of your language unless I can understand practically all of it. But then how, save in a single spasm of seamless cognition, could any language ever be learnt?” (Fodor/Lepore 1992, 9) We cannot learn a language as a whole. But we can learn words and grammatical rules: We can learn parts of a language (Sellars 1963). Hence, although there are some confirming cases, both the formation thesis and the perception thesis do not hold generally. Thus they give no reason for general restrictions on explanations. b. The second claim, (H2), is that the whole acts on its parts. My hand is a part of my body. Thus, if I raise my hand, I – the whole – obviously act on my hand – the part. Similar, if a certain gas volume (i.e. the whole) is being compressed, the molecules it contains (i.e. the parts) raise their velocity. Indeed, changes in wholes can cause changes in their parts. Although this is obviously true, no restriction on explanation follows: Besides the whole-part explanation for the discussed phenomena there exist alternative explanations based on part-part interaction. The movements of my hand may also be explained by certain processes in brain cells and nerves, i.e. in other parts of me. The higher velocity of the molecules may be explained by the higher frequency of collisions with other molecules within this gas volume. Unless we have independent reasons to regard these explanations as inferior to the holistic explanations, we cannot discard them as invalid. c. What about the claim that a whole is more than the sum of its parts? This claim has been dubbed the ‘principle of supersummativity’. It has provoked fierce criticism. One of the critics is Moritz Schlick. For him, the difference between ‘sum’ and ‘whole’ was based on a different mode of representation of one and the same object, rather than an ontological difference between two groups of entities.5 My reading of this ‘principle of supersummativity’ is along this line:6 5

Cf. Schlick 1935; Simons 1987, 324. Schlick stresses on the vagueness of the distinction between “mere sums” and “wholes”. My (H3*) further down is meant as an attempt of disambiguation – but I will show in the following that it does not help to draw a clearcut distinction. On Schlick’s relationship to Gestalt psychology cf. now Kluck 2008. 6 Cf. Bunge 1979, 40; Hassenstein 1972, 32. At least, this rendering is in harmony with the classical account of supersummativity in Ehrenfels’ explanation of ‘Gestalt-

Beyond the Ultimate

65

H3* Wholes have properties which are different in genus from the properties of their parts. What is at stake here are determinable properties rather than determinate properties (Johnson 1921, 173-175), or attributes rather than their values (Simons 1987, 343): Trivially, a collection of molecules has a different mass than each molecule taken singly. But these are only different determinate properties of the same determinable property, or different values of the same attribute, i.e. of the attribute mass. In fact, the mass of a gas volume is composed out of the individual masses of the molecules. With respect to its mass, that is, a gas volume is indeed only the sum of its parts. We can explain the over-all mass by recurring to the parts and their masses. In the case of the gas volume as well as in the case of the single molecules it is the same kind of property, i.e. the property mass, even if the mass of the volume is higher than the mass of any molecule. But we also ascribe properties like temperature and pressure to the gas volume, which we cannot ascribe to single molecules (Johansson 2004b). These are kinds of properties which the whole possesses, but not the parts. Nevertheless these properties can be explained by recurring to the parts, as these properties of the whole are causally connected with the parts. Temperature and pressure of the gas are determined by the impulse of the individual molecules. The properties of the whole and its parts may be separated categorically but are connected causally (Hassenstein 1972). Although the parts do not have a property like temperature, the temperature of the gas can be explained by recurring to the parts. Hence, the existence of such emerging properties imposes no restriction on the selection of the entities we refer to in explanations.7 d. The fourth holistic statement directly contradicts this solution, claiming that wholes do not emerge through the actions among the components and the environmental influences only. This means, that there are properties of the whole, which do not depend only on the properties of the parts or of the environment. Let’s discuss this using a famous historical example. qualitäten’, who takes a musical melody (or “Tongestalt”) as his paradigm. Cf. Ehrenfels 1890 (transl. Smith, p. 85). 7 Mass seems to be only a matter of quantity. Then the interpretation of Simons 1987, 359-360 fits in: Supersummativity is fulfilled “so long as the Gestalt in question is richer in structure than simply that of having so many parts.”

66

Ludger Jansen

At the turn of the century, the German biologist Hans Driesch (18671941) experimented with sea urchins. He divided a yet undifferentiated sea urchin’s germ and observed that both halves developed to integral sea urchins, which were only a bit smaller than usual. As Driesch couldn’t explain this phenomenon by help of physical or chemical laws, he postulated a non-physical cause. This non-physical (and therefore immaterial) cause was meant to be the missing “whole-making factor”.8 Driesch explained the development of the germs, which could not be explained otherwise by the biology of his days, by means of this “Faktor E”, for “entelechy” (Hassenstein 1981, 60-61). Though Driesch borrowed the word “entelechy” from Aristotle, he did not take over the concept: In fact, Driesch transformed Aristotle’s final cause into an efficient cause.9 Thus, Driesch’s vitalism rests on confusion between different kinds of causal explanations. The problem of vitalism can be disposed of by a clear cut conceptual distinction between (efficient) causality and teleology. Given this conceptual distinction, there is no contradiction between physical or chemical causality on the one hand and biological teleology on the other.10 A technical analogy can show that teleological principles of a whole can indeed have their causal origin in the parts and their interactions, because from a cybernetic point of view teleological behaviour is not a proprium of biological systems. A thermostat is a decent example for teleological behaviour: It rules about cooling or heating in order to keep the temperature at the target value. But in order to explain a thermostat, no immaterial “Faktor E” is necessary. It is sufficient, to point out the individual parts of this mechanism and their interactions. Similarly, we can causally explain certain teleological processes within living beings by means of 8

Driesch’s explanation is ‘better’ than one may suspect at first glance: It has not only “explanatory appeal”, but also “explanatory power”, as the heuristic power of his hypothesis in biology indicates, which has not even been denied by adversaries of Driesch’s vitalism. Cf. Hassenstein 1981, 68; Hartmann 1948, 186. For the distinction between explanatory power and appeal cf. Rapoport 1972. 9 For the deviating Aristotelian concept of entelechy cf. e.g. Chen 1975, Kullmann 1979, Jansen 2002, 95-98. 10 Cf. Hassenstein 1981, 62: „Die Vitalisten [...] übersahen, daß die Finalität der Lebenserscheinungen auf einer speziellen Form der Kausalität beruht (z.B. auf Regelprozessen) und daß deren Verständnis die Anerkennung der Finalität nicht aufhebt, sondern untermauert.”

Beyond the Ultimate

67

physical or chemical laws, without depriving these processes of their teleological character. The two modes of explanation do not exclude each other, but have their own merits and their own limits.11 e. The fifth holistic statement said that wholes cannot be explained analytically. This statement is obviously true, if ‘analysis’ here means only the splitting into parts, neglecting their interactions (Bunge 1979, 40). But if we consider parts plus the net of relations and interactions between them, then we can indeed successfully analyse wholes, as the discussion so far has pointed out. No wonder, the standard definitions of systems refer both to the elements and the relations between them.12 These “relational parts” have to be taken into account for successful explanations. 4. Parts and wholes in explanations Let us summarize the discussion so far by rehearsing the argument: Atomism and holism restrict our explanations either to the smallest parts or the wholes themselves. Methodological atomism, on the one hand, cannot be practised; thus it is false. This is a sound argument, as methodological atomism (like any other methodology) is a promise in disguise. The promise is that if you follow recipe X you will have explanatory success. But if it is impossible to act according to recipe X, be it for logical or other reasons, the recipe must deceive all hopes connected with it. It cannot possibly lead you to whatever is the promise connected with it. In particular, it cannot lead to explanatory success. Thus, methodological atomism is false. Methodological holism, on the other hand, is not necessary. Thus we do not have any reason to dispose of with a wide range of explanations hitherto successfully accepted. Therefore, we may refer both to wholes and their parts in explanations, as we may consider it to be appropriate.13

11

Cf. for this e.g. the passages on evolution in Popper 1976, ch. 37. Feedback mechanisms (like mechanic speed regulators) are also discussed in Johansson 2004a, 249253. 12 Cf., e.g.,, Bertalanffy 1972, 18 („eine Menge [...] von Elementen, zwischen denen Wechselbeziehung bestehen”). For a lexicographic survey of the word „system” in the history of ideas cf. Steinbacher 1990. 13 Bunge 1979, 41 calls a similar position “systemism”. Röwer 1985 argues for such a position by means of a case study demonstrating the explanatory power of the different

68

Ludger Jansen

Then, how do explanations look like that are neither bound to atomism nor to holism? Well, such explanations are really nothing special. Not only common sense explanations, but also scientific explanations refer to both wholes and their parts. Here is an example: Consider someone working on a computer with a graphic user interface. He will ‘close windows’, ‘move arrows’, and so on. Even if he is quite computer illiterate otherwise, he may be successful in using the computer for his purposes and most probably he will be able to teach other people to do this work, too. However, problems may occur, which can only be solved by someone knowing how to translate ‘moving an arrow’ into the language of the operation system. Hence, the ‘holistic’ explanations on the level of the graphic user interface are appropriate for the computer illiterate, and for certain purposes. For other purposes we have to proceed to another level: to the operation system, or further into the computer’s innermost functioning: the assembler code, the binary digits, the transistors and maybe to the atomic structures of the silicon chips. We deal with a multitude of levels of explanations. Out of these we choose one level according to our knowledge, our curiosity, or our purposes. Maybe we aim at solving technological problems, like in the computer example. But there is a wide range of non-technological aims of scientific research, especially if there is such a thing like a natural pursuit of knowledge as an end in itself for human beings. It is important to notice that such a hierarchy does not imply reductionism. ‘Closing a window’, for example, may have quite a lot of possible translations into the machine code level. Quite similar examples can be found in science: Instruments like clocks, respiratory systems and even enzymatic proteins can have quite different structures. When we consider examples from biology, we can make another observation: We can ascend from the individual living beings up to species, ecological systems, and so on. We have “natural hierarchies of natural wholes”, as Peter Simons puts it (Simons 1987, 332; italics in the original). It is not necessary that all these hierarchies bottom out with atoms and their substructure. A linguist may design a hierarchy with a ‘semiotic’ bottom of morphemes and phonemes, and a sociologist (like Luhmann 1987) may develop a social system theory and take acts of communication as the smallest elements. It is also approaches. I avoid using a positive name for my position in this paper, because it is essentially a negative position: Neither atomism nor holism is true.

Beyond the Ultimate

69

important to note, that there are not only different levels to explain the same phenomenon, but as we have already seen, also different aspects for explanations, e.g. the causal and the teleological aspect, which we may both consistently use. Not only can we refer to different levels in different explanations, but often we refer to different levels in one and the same explanation. In classical dynamics a central elastic impact of two spheres is normally explained in the following way: The kinetic energy of the sphere initially moving is transposed to the sphere initially at rest. This is a statement about the parts (the two spheres) rather than about the whole of the physical system of this impact. But for the whole explanation something is wanting, namely the statement that the amount of energy remains the same within closed systems. Hence, we do not only refer to properties of the parts, but to a property of the whole system at the same time. In this paper I argued for a negative thesis: There are no reasons to restrict the entities we refer to in explanations to the ultimate. I presented reasons for refusing methodological atomism and demonstrated that typical teachings of holism do not justify a restriction to the choice of entities to be referred to in explanations. Finally, I showed that commonly accepted scientific and common sense explanations do indeed refer both to wholes and parts, in some cases even at the same time. Thus philosophical reasoning as well as scientific practice and common sense advocate a liberal view of reference to wholes and parts in explanations. This is a good thing, and for several reasons: First, ultimate things may not be cognitively accessible to us; or our language may be inadequate to talk about such ultimate things. But second, maybe there are neither ultimate atoms nor ultimate wholes. Maybe they just do not exist. And the possibility of explanation, as well as the possibility of ontology as a philosophical discipline, should not be contingent on such details of the causal architecture of the universe.14

14

Earlier versions of this paper have been presented at conferences in Leeds and Herzberg ob Aarau. I would like to thank the participants for helpful comments and challenging questions. I am especially indebted to Manfred Drack, Michael Esfeld, Niels Grewe, Ingvar Johansson, Hans Primas, Niko Strobach and Marcus Willaschek, who read and commented on various earlier versions of this paper. The paper was completed under the auspices of the ExCell project funded through the ExzellenzFörderprogramm Mecklenburg-Vorpommern (EFP.MV).

70

Ludger Jansen

References Bertalanffy, L. von 1972 “Vorläufer und Begründer der Systemtheorie”, in: R. Kurzrock (ed.), Systemtheorie, Berlin: Colloquium, 17-28. Bunge, M. 1979 Treatise on Basic Philosophy 4: Ontology 2: A World of Systems, Dordrecht, Boston, London: Reidel. Chen, C.-H. 1958 “The relation between the terms ENERGEIA and ENTELECHEIA in the philosophy of Aristotle”, in: Classical Quarterly NS 8, 12-17. Clark, S.R.L. 1975 Aristotle’s Man. Speculations upon Aristotelian Anthropology, Oxford: Clarendon Press. Drieschner, M. 1979 Voraussage – Wahrscheinlichkeit – Objekt. Über die begrifflichen Grundlagen der Quantenmechanik, Berlin, Heidelberg, New York: Springer. -

1992, Carl Friedrich von Weiszäcker zur Einführung, Hamburg: Junius.

Ehrenfels, Chr. von 1890 “Über ‚Gestaltqualitäten’ ”, in: Vierteljahresschrift für wissenschaftliche Philosophie 14, 249-292; translated into English by B. Smith: On ‚Gestalt Qualities’, in: B. Smith, (ed.), Foundations of Gestalt Theory, München, Wien: Philosophia 1988, 82-117. Esfeld, M. 2002 Holismus in der Philosophie des Geistes und in der Philosophie der Physik, Frankfurt M.: Suhrkamp. Essler,

W.K. 1979 Wissenschaftstheorie Freiburg/München: Alber.

IV.

Erklärung

und

Kausalität,

Fodor, J.A. and Lepore, E. 1992 Holism. A Shopper’s Guide, Oxford: Blackwell. Gilbert, M. 1989 On Social Facts, London/New York: Routledge 1989. Gregory, P.R. 2007 Lenin’s Brain and Other Tales from the Secret Soviet Archives, Stanford CA: Hoover Institute Hartmann, M. 1948 Die philosophischen Grundlagen der Naturwissenschaft, Jena: Fischer. Hassenstein, B. 1972 “Element und System – geschlossene und offene Systeme”, in: R. Kurzrock (ed.), Systemtheorie, Berlin: Colloquium, 29-38. -

1981 “Biologische Teleonomie”, in: Neue Hefte für Philosophie 20, 60-71.

Jansen, L. 2002 Tun und Können. Aristoteles’ Theorie dispositionaler Eigenschaften im neunten Buch der Metaphysik, Frankfurt M.: Hänsel-Hohenhausen. Johansson, I. 2004a Ontological Investigations. An Inquiry into the Categories of Nature, Man and Society, 2nd extended edition, Frankfurt M.: Ontos. -

2004b “The Ontology of Temperature”, in: Ursus Philosophicus. Essays dedicated to Björn Haglund on his sixtieth birthday (= Philosophical Communications. Web Series 32), 115-124, online at: http://www.phil.gu.se/posters/festskrift/festskrift_johansson.pdf.

Beyond the Ultimate

71

Johnson, W.E. 1921 Logic. Part I, Cambridge: Cambridge University Press. Kluck, St. 2008 Gestaltpsychologie und Wiener Kreis. Stationen einer bedeutsamen Beziehung, Freiburg/B., München: Alber. Kullmann, W. 1979 Die Teleologie in der aristotelischen Biologie. Aristoteles als Zoologe, Embryologe und Genetiker, Heidelberg: Carl Winter (= Sitzungsberichte der Heidelberger Akademie der Wissenschaften, Philos.-hist. Klasse, 1979; 2. Abh.). Meggle, G. 2002 „Gemeinsamer Glaube und Gemeinsames Wissen“, in: Mark Siebel (Hg.), Kommunikatives Verstehen, Leipzig: Leipziger Universitäts-Verlag 2002 (= Leipziger Schriften zur Philosophie 16), 12-20. Popper, K.R. 1976 Unended Quest. An Intellectual Autobiography, Glasgow: Fontana. Rapoport, A. 1972 “Explanatory power and explanatory appeal of theories”, in: Synthese 24, 321-342. Röwer, H.-G. 1985 Holismus und Elementarismus in der Systemtheorie, Frankfurt/M., Bern, New York: Peter Lang. Schlick, M. 1935 “Über den Begriff der Ganzheit”, in: Wissenschaftlicher Jahresbericht der Philosophischen Gesellschaft an der Universität zu Wien für das Vereinsjahr 1933/34 und 1934/35, repr. in: M. Schlick., Gesetz, Kausalität und Wahrscheinlichkeit, Wien: Gerold, 1948, 56-74. Sellars, W. 1963 “Empiricism and the Philosophy of Mind”, in: W. Sellars, Science, Perception and Reality, London, New York: Routledge & Kegan Paul, 127196. Simons, P. 1987 Parts. A Study in Ontology, Oxford: Clarendon Press. Steinbacher, K. 1990 Art. “System/Systemtheorie”, in: H.J. Sandkühler (ed.), Europäische Enzyklopädie zu Philosophie und Wissenschaften, Hamburg: Meiner, Bd. 4, 500-506. Weizsäcker, C.F. von 1974 Die Einheit der Natur, München: dtv. Weiss, P.A. 1969 “The Living System. Determinism Stratified”, in: Studium Generale 22, 361-400. Wilson, R.A. 2004 Boundaries of the Mind. The Individual in the Fragile Sciences: Cognition, Cambridge: Cambridge University Press. Wuketits, F.M. 1989 “Organisms, Vital Forces, and Machines: Classical Controversies and the Contemporary Discussion ‘Reductionism vs. Holism’ ”, in: P. Hoyningen-Huene and F. Wuketits (eds.), Reductionism and Systems Theory in the Life Sciences, Dordrecht, Boston, London: Kluwer, 3-28.

A Plea for a Naturalistic Ontology JOHANNES L. BRANDL, Salzburg

1. Introduction In philosophy, Naturalism is understood to be a paradigm that confers a scientific direction on our thinking. 1 In conjunction with this, there are in the foreground conceptual and methodical constraints that a naturalistic orientation implies. Thus, it is characteristic of naturalists’ self-understanding to reject as unscientific any inquiries whose presuppositions exceed the bounds of empirical knowledge. Inquiries of that sort are to be found in all realms of philosophy. Traditionally, it was the questions of metaphysics, ethics and theology that attracted the criticism of the Naturalists. Nowadays, this debate has shifted a bit, and it is issues in epistemology and the philosophy of mind to which the attention of both the detractors and champions of Naturalism are directed.2 What has not changed is the vehemence with which the viability of the naturalistic paradigm is being discussed. Here, world-view opposites collide and polarize opinions. Proponents of Naturalism see in it – as Quine once aptly remarked – an “object of loyalty,” whereas its opponents have made of it an “object of obloquy” (Quine 1995, 251). In this context, the loyal Naturalist views himself in the role of an enlightened human being of modern times, who does not believe in esoteric powers or paranormal phenomena. For his adversary, in contrast, there is concealed behind this enlightened posture an uncritical, and ultimately unphilosophical, scientism that is not shared even by the scientists themselves. A substantive reckoning with the naturalist paradigm hardly seems possible in view of such a deep conceptual gulf. Indeed, it seems altogether unnecessary. For what is the point of conducting a serious debate, when one has the choice between a Naturalism that cannot be challenged with rational arguments 1

I am grateful to Edmund Runggaldier for inspiring me to think about Philosophical Naturalism in this way. 2 Cf. Villanueva (1993), Wagner (1993), French et. al. (1994), Quitterer and Runggaldier (1999), Keil (2000), Craig and Moreland (2000).

74

Johannes L. Brandl

and a scientism that is self-contradictory? Geert Keil and Herbert Schnädelbach arrive at this skeptical conclusion in their futile efforts to endow the Naturlism-debate with a philosophically interesting content: “Much that passes for Naturalism,” so they summarize, does not deserve this name, “whereas that which does deserve it does not merit subscribing to” (Keil and Schnädelbach 2000, 44, transl. AS).3 My objective in the sequel will be to sketch a picture of Naturalism that makes a substantive engagement with this paradigm both possible and meaningful. In doing so, I have no desire to invent any new type of Naturalism, but only to show how some of the familiar naturalistic slogans can be interpreted in such a way that they cannot be frivolously dismissed ad acta.4 A properly understood Naturalism, I aim to show, qualifies neither as an “object of loyalty” that one can accept untested, nor as an “object of defamation” that can be easily ridiculed. Rightly understood, so my thesis runs, Naturalism poses a mandate to ontology, namely, the mandate – in view of the diversity and complexity of scientific knowledge – to make plausible how all of reality can be uniformly constituted. The two slogans that I shall consider in greater detail are (1) the denial of a “First Philosophy” and (2) the idea of a “Unified Science.” Together, these slogans make up the methodological core of Naturalism. On the ontological side, in contrast, we encounter two other slogans, namely (3) the thesis “Everything is part of nature” and (4) the principle “There is nothing new under the sun.” As concerns the proper interpretation of such sayings, I shall have to rest content with a few indications. I shall also not be able to discuss in this regard to what extent methodical and ontological issues can be separated at all, nor what other aspects a comprehensive Naturalismconcept would have to take into account. To my way of thinking, what is decisive is the proper interpretations of the methodological core, for it is 3

W.L. Craig and J.P. Moreland also argue in a similar veign by setting Naturalism before the following dilemma: “Either Core Scientism is not itself a thesis included in or warranted by the (hypothetically completed) empirical sciences, or [...] it becomes harmlessly stipulative in nature, with no firm basis in what actually goes by the name “empirical science” [...] Given that dilemma, Philosophical Scientism becomes philosophically innocuous.” (Craig and Moreland 2000, 13f) 4 On this point I differ from those who plead for a “naive” or “innocuous” interpretation of the naturalist standpoint, as, e.g., Almeder (1998) and Hornsby (1997). Kitcher (1992) also tends in this direction.

A Plea for a Naturalistic Ontology

75

from this that the mandate to ontology emerges. This is especially so with regard to the domain of philosophy of mind, which I have in my sights here as the primary target of application. 2. Quine as Anti-Fundamentalist The fact that naturalist positions play such a significant role in contemporary philosophy is attributable not only, but primarily, to the influence of Quine. Quine is the “Piper-Naturalist” toward which both adherents and critics of the naturalist paradigm are oriented. While this focus on Quine has the advantage of being able to refer to a concretely spelled-out philosophical position when speaking about “Naturalism,” this also involves the danger of an insular exegesis of this concept. Not everything for which Quine has achieved fame must also therefore be regarded as part of the naturalist paradigm. This even holds for the central theses of Quine’s philosophy: the critique of the analytic-synthetic dichotomy, the favoring of extensional over intensional languages, the theses of the indeterminacy of translation and of ontological relativity, or the program of ontological austerity that, apart from four-dimensional individuals, recognizes only sets of such individuals as constituents of reality. Since Quine’s philosophy forms a finely spun web that draws its strength from the crosslinks between individual knots, revisions of central spots obviously have far-reaching consequences in the entire system of his philosophy.5 However, that need not be a reason for staving off such revisions with the argument that they would make questionable Quine’s Naturalism as a whole. With this outlook, I now wish to fathom the meaning of that slogan with which Quine is fond of giving expression to his naturalist position: namely, with the rejection of a First Philosophy. To begin with, I shall plumb the range of interpretations that allows for stronger and weaker readings. I shall then – in contradistinction to Quine – declare in favor of a relatively weak, though not trivial, interpretation of this slogan. The rejection of a “First Philosophy,” I propose, is a clear and unequivocal repudiation of philosophical fundamentalism in the theory of knowledge. 5

The significance of these crosslinks in Quine’s “web of belief” is stressed by, among others, Gibson (1982), Hookway (1988), Hylton (1994) and Decock and Horsten (2000).

76

Johannes L. Brandl

Quine himself has conferred the predicate “naturalist” on his position in a certain way only in hindsight – not until several years following the publication of Word and Object (1960) – in the now classic paper “Epistemology Naturalized”, as well as in the John Dewey Lecture on “Ontological Relativity”.6 In them, Quine programmatically proclaims that, following the debacle of empiristically-reductionist programs, theory of knowledge must be “naturalized,” i.e., it “simply falls into place as a chapter of psychology and hence of natural science.“ (Quine 1969a, 82). He finds support for this idea in John Dewey, with whom, as he says, he feels himself to have a bond “concerning Naturalism.” It is in this connection that he decrees the end of a First Philosophy: With Dewey I hold that knowledge, mind, and meaning are part of the same world that they have to do with, and that they are to be studied in the same empirical spirit that animates natural science. There is no place for a prior philosophy. (Quine 1969b, 26)

Quine names here three concrete themes – knowledge, mind and meaning – which should be investigated not under a “philosophical” rubric, but rather under a “natural-scientific” one. This is not to be understood as a thematic constriction, since Quine later extends this disposition to all realms of philosophy. He then sets the naturalist stance on a par with the “recognition that it is within science itself, and not in some prior philosophy, that reality is to be identified and described” (Quine 1981, 21). What Quine is therefore keen on pointing out is that knowledge, mind and meaning do not merit any special treatment, but rather must be dealt with as part of the reality in which we live. This becomes particularly apparent in his answer to an objection raised by Putnam in the Schilpp volume: “I admit to naturalism, and even glory in it. This means banishing the dream of a first philosophy and pursuing philosophy rather as part of one’s system of the world, continuous with the rest of science” (Quine 1986, 430). Nonetheless, there is still room for interpretation here. One could understand this rejection of First Philosophy as saying that philosophy should no longer concern itself at all with those ontological questions which, according to Aristotle, belong to prima philosophia. Since Quine did, however, 6

Both papers first appeared in the collection of essays Quine (1969). In this connection, see also Quine’s brief sketch on the history of empricism in Quine (1981a).

A Plea for a Naturalistic Ontology

77

continue to deal with such questions, and we do not wish to allege any kind of inconsistency on his part, we must assume that Quine did not want to exclude some particular themes of traditional metaphysics, but rather certain methods by which these issues have thus far been addressed. This is a first, important clarification which continues to leave some matters open however, as we shall presently see. The decisive question at this point is in which “grade of naturalistic commitment” one might have to engage in this context, as Christopher Hookway has formulated it (see Hookway 2000). In parallel to Quine’s “Three Grades of Modal Involvement” (1966), Hookway sets up a threetiered scale. The principle that philosophical and empirico-scientific theories must be consistent with each other, and should be interdisciplinarily linked, corresponds to a first grade of commitment. Secondly, a Naturalist can commit to the principle that the well-established explanatory models of physics are also to be applied in biology, and hence to human capacaties – including our rational faculties. However, one arrives at the third grade of naturalistic commitment only when the possiblility of there being any other cognitive methods for the realm of rationality is explicitly excluded, methods that philosophy in particular is supposed to contribute. Now as far as Quine’s orientation is concerned, Hookway maintains that Quine inclines rhetorically to a radical position, hence to a third degree commitment on this scale. But behind this rhetorical facade there is often concealed only a weaker Naturalism of the second, or even of the first, grade (see Hookway 2000, 37ff). I consider this diagnosis as fundamentally on target, and might refine it a bit. There are two indicators, in my view, that the strength of Quine’s Naturalism is subject to vacillations. I have already mentioned one of these indicators: on the one hand, Quine formulates his naturalist theses by reference to the domain that can be circumscribed with the concept of rationality; on the other hand, he espouses a standpoint that does not allow for any restrictions of content. One can therefore distinguish between a local and a global Naturalism. Secondly, Quine adopts an ambivalent stance vis-à-vis the humanistic and social sciences. Since these sciences too are through and through empirically founded, they are due the same respect as the natural sciences. It is clear, nonetheless, that Quine wants to grant the natural sciences a methodological privilege, resting on the premise that only the

78

Johannes L. Brandl

natural sciences embody that form of empiricism which matters to Quine.7 Consequently, one can additionally distinguish in Quine between a liberal and a restrictive Naturalism. By permuting these possibilities, one arrives at a spectrum of four possible positions. This now extends from a local Naturalism, which is also liberal, all the way to a global Naturalism, which is also restrictive. Having staked out in this fashion the playing field for possible interpretations of Quine’s slogan, I would now plead for the second weakest interpretive option – namely, for an exposition in the sense of a local, but restrictive Naturalism. In doing so, I shall delineate the domain to which these restrictions apply even more narrowly than Quine, and shall disconnect it from the issue of how a theory of rationality is to be developed. To that end, I take my point of departure from a different quotation of Quine, which points in that direction: My point in the characterizations of naturalism that I quoted is just that the most we can reasonably seek in support of an inventory and description of reality is testability of its observable consequences in the time-honored hypothetico-deductive way – whereof more anon. (Quine 1995, 252)

Quine speaks here once again as “global Naturalist” for whom every kind of cognitive effort contributes toward offering “an inventory and a description of reality.” What is more important, however, is that here Quine stipulates precisely the method according to which this inventory is to proceed: Only those theories are admissible which are verifiable by “observable consequences.” It is of course “intersubjective observation” that is meant by “observation” in this context. With this it also becomes clear what Quine wants to foremost exclude here: knowledge whose empirical basis is subjective. A mentalistic, introspective psychology can make no contribution to a description of reality if its results are not verifiable on the basis of intersubjectively accessible data. In the case of mental experiences this means: on the basis of intersubjectively observable conduct of subjects. It 7

The problem that Quine faces here is that the boundaries between the so-called “soft” and “hard” sciences are fluid. Quine would therefore prefer, as he himself says, to forgo drawing demarcation lines: “Demarcation is not my purpose”. (Quine 1995, 252). But without such demarcations it is impossible to distinguish the empiricist disposition of the natural sciences from the empiricist disposition of other sciences, and thus stipulate priority for the former.

A Plea for a Naturalistic Ontology

79

is therefore a behaviorist principle that Quine allows here to infiltrate into his Naturalism.8 It would appear superfluous to expend any further thought on behaviorism after it has lost its scientific justification in psychology. Yet it is not quite so. Ad acta one can only dismiss the claim that the domain of psychological facts is identical with the domain of intersubjectively observable conduct. One must also not discuss any further whether the domain of subjective experience should be ranked as an ontologically second-rate domain. At issue are plainly and simply facts that are a part of the reality in which we live and which call for scientific explanation. Despite this, I believe there is something about Quine’s commitment to behaviorism that should not be overlooked. It seems to me that the operative point to be sought here is not whether or not intersubjectively verifiable observations are the sole legitimate source of knowledge. The pertinent question is whether there is some other subjective, cognitive method that is even more foundational, and therefore in a certain way higher-ranking. With this we first get on the track of the idea of a First Philosophy, which Quine puts in question. What he denies is that philosophy can employ a method – in the form of subjective cognition – which confers on it a privileged status vis-à-vis all other sciences that make use of the familiar hypothetico-deductive method. It is precisely this claim that epistemologists in the tradition of Descartes have made by appeal to inner experience. They saw in it a cognitive method capable of providing an empirical foundation for all of science. I therefore submit that Quine’s slogan be understood as giving expression to an anti-fundamentalist stance, without thereby denying to inner perception every cognitive function whatsoever. How is this to be reconciled? Toward this end, it is requisite – in my view – to restrict the domain within which the intersubjective methods of the empirical sciences can lay claim to a cognitive monopoly. To my way of thinking, this monopoly can be substantiated when it comes to the elemenary structure of reality. Fun8

This declaration undoubtedly attains the third grade on Hookway’s scale of naturalistic commitment. For precluding inner perception from being any kind of selfsufficient source of knowledge also means “that the philosophical needs that prompt us to think about mind, cognition and rationality can be met entirely by research in the special sciences” (Hookway 2000, 38).

80

Johannes L. Brandl

damentalist theories of the Cartesian ilk assume, in contrast, that the subjective access to the elements of consciousness is simultaneously an access to the elementary constituents of reality. Consequently, they advance the bold conjecture that the entire structure of science can be erected on the slim basis of subjective knowledge. Inner experience therefore loses its supposed relevance as the foundation of all science, if it does not give us access to elementary cognitions in this sense. With this background, the anti-fundamentalist thesis of Naturalism can now be formulated somewhat more precisely, as follows: The Anti-Fundamentalist Thesis: The elementary constituents of reality – including elementary mental events – are only knowable with the aid of theories that are only verifiable on the basis of intersubjectively verifiable observational data.

This thesis undermines the pretensions of a First Philosophy which attempts to derive a special relevance for subjective knowledge from the evidence of inner perception. Philosophy is no foundational science that investigates the conditions of all possible knowledge, and wants in this fashion to procure the presuppositions on which all sciences ultimately depend. A Naturalist who rejects this does not thereby deny that inner perception is an autonomous source of knowledge. What it offers us, however, is no elementary knowledge concerning the structure of reality, but rather insights into highly complex processes even when subjectively these might appear to us as simple. This complexity implies that we also rely on intersubjectively accessible observational data as evidential basis when we investigate seemingly elementary mental facts. 3. Carnap and the Program of “Unified Science” The rejection of a philosophical knowledge of elementary facts – which cannot be explained by the empirical sciences, and therefore must be presupposed by them – is the key methodological part of the naturalist paradigm. It is a form of local Naturalism that declares the methods of the empirical sciences to be indispensable in a certain domain – namely, in the exploration of the elementary structure of reality. At the same time it is a restrictive Naturalism, because it denies to philosophical knowledge a cognitive privilege within this domain.

A Plea for a Naturalistic Ontology

81

How is this restricted cognitive privilege of the natural sciences now to be justified, and what further consequences follow from it? From a pragmatic point of view, it would be obvious to point to the success of the natural sciences. To be sure, the appraisal of this success is necessarily a double-edged sword, since every advance in science opens the door to new and as yet unsolved problems. We need to approach the present issue therefore from a different direction. Perhaps the cognitive privilege of the natural sciences, as pertains to the elementary structure of reality, possesses an ontological legitimation in that its object-domain is more fundamental than the domain of all other sciences. Can such a justification be reconciled with the principles of Naturalism? To answer this question, I shall now consider in greater detail the program of Unified Science as developed in the 1930s by exponents of the Vienna Circle, and by Rudolf Carnap in particular. Also here I shall call attention to the fact that there are various interpretations of this program. Viewed historically, what is at issue here is a progam for the semantic unification of science. In contrast I shall propose a reading that rests on the idea of a ontological unity of reality. As I shall elucidate, the advantage of this interpretation for the Naturalist consists in the idea of the unity of science – so understood – being thoroughly compatible with the semantic autonomy of individual sciences. To begin with, I describe the program of Unified Science in its original form, based on Carnap’s early papers in which he advocated a rather naive semantic reductionism. I will then mention important modifications in Carnap’s view that can be found in his later works.9 In view of these modifications, I will then raise the question what remains of the Unity Thesis if one accepts the autonomy of the individual sciences. This is not only a challenge for Carnap, as we shall see, but also for Quine’s preference for extensional languages as the unifying bond in science.

9

By the early works I mean, in particular, the essays “Die physikalische Sprache als Universalsprache der Wissenschaft [The Physical Language as Universal Language of Science]” (1931) and “Psychologie in physikalischer Sprache [Psychology in Physical Language]” (1932). The later works I shall take into account are foremost “Logical Foundations of the Unity of Science” (1938) and the corresponding passages in Carnap’s Autobiography as well as his Reply to Feigl in the Schilpp volume (Schilpp 1963, 50-53 and 882-886).

82

Johannes L. Brandl

The conviction that the natural sciences possess a methodological privilege vis-à-vis all other sciences, was a firm component of the scientific world-view of the Vienna Circle. Still, there was a considerable divergence of opinions as to what exactly this privilege consists in and how it is to be legitimated. Otto Neurath espoused in this connection a scientificsociological and “encyclopedic” standpoint, while Carnap argued in a scientific-systematic fashion (see Mormann 2000, 122). In retrospect it looks like Neurath rendered better service to the naturalist standpoint than Carnap; but it was Carnap’s perspective that stamped the Unified Science program with its outward brand. Carnap’s starting point was that the unification of science is to be conceived as a semantic unification of the language of science. This was supposed to lead in consequence to a methodological unification of science, and ultimately to its unification in terms of content (see Carnap 1938, reprint 422). For this reason Carnap strived to show that a natural-scientific language – namely the physical language – is the universal language of science. The following passage from Carnap’s Autobiography makes clear how he understood this task: In our discussions, chiefly under the influence of Neurath, the principle of the unity of science became one of the main tenets of our general philosophical conception. This principle says that the different branches of empirical science are separated only for the practical reason of division of labor, but are fundamentally merely parts of one comprehensive unified science. (...) In contrast to this customary view, Neurath maintained the monistic conception that everything that occurs is a part of nature, i.e., of the physical world. I proposed to make this thesis more precise by transforming it into a thesis concerning language, namely, the thesis that the total language encompassing all knowledge can be constructed on a physicalistic basis. (Carnap 1963, 52)

With reference to this portrayal, from today’s vantage point one could say that Neurath suggests an ontological grounding for the unity of science: The boundaries between the individual sciences can only be pragmatically drawn because “everything that happens is a part of nature.” In particular, there is no ontological reason that would justify seeing a divide between the natural sciences and the humanistic and social sciences. Carnap wants to make this thesis more “precise,” as he says by reconstructing it into a

A Plea for a Naturalistic Ontology

83

semantic thesis that pertains to the total language of science.10 In the course of this “precisioning”-process, however, Carnap arrived at a conception that is markedly stronger than the Principle of Unified Science as understood by Neurath. Carnap initially tackled this project with the aid of translations or by proffering rules of translation. All scientific propositions were supposed to be translatable into a physical language, whereby physical language is characterized by two features: It is intersubjective, and it contains only quantitative determinations (i.e., predicates for quantifiable magnitudes) or translations of such determinations (Carnap 1931, 441). How Carnap envisaged these translations is illustrated by the examples he employs. So, e.g., the psychological proposition “Mr. A has toothaches” is supposed to be translatable into a proposition of the form “ Mr. A’s body is in a state that is characterized by measurable properties of its micro-structure (e.g., neuro-physiological properties) as well as by measurable reactions to certain stimuli (e.g., pain tolerance)” (see Carnap 1932, 112ff). Yet Carnap himself soon recognized the naivete of his project. One reason for this was that it employed a translation concept containing verificationist presuppositions that were too strong. According to it, two propositions were supposed to be translatable into each other precisely when the same propositions of the protocol language are derivable from them. This is why in later works we find in Carnap markedly weaker criteria for correct translations: logical equivalence of expressions, intensional isomorphism, but also weakened verificationist criteria, such as equality in degree of confirmation (see Carnap 1947, and Schilpp 1963, 883). But the question persists whether the adduced examples of translations satisfy these reduced standards.11

10

That such reconstructions were needed was beyond question for Carnap: “The question of the unity of science is meant here as a problem of the logic of science, not of ontology. We do not ask: ‘Is the world one?’, ‘Are all events fundamentally of one kind?’, ‘Are the so-called mental processes really physical processes or not?’ [...]. It seems doubtful whether we can find any theoretical content in such philosophical questions as discussed by monism, dualism, and pluralism.” (Carnap 1938, reprint 413) 11 In his later analysis of belief-sentences, Carnap treats terms like “belief” or “anger” as theoretical terms and admits that sentences containing such terms cannot be

84

Johannes L. Brandl

Secondly, Carnap must have discerned that theoretical terms like, e.g., “electrical charge” are not definable in an observational language. Thus, if theoretical terms are involved, it is impossible to base a translation of psychological propositions on a definition of psychological concepts in terms of physical concepts. For this reason Carnap switched to speaking of a “reduction” of psychological propositions into propositions of physical language, and no longer spoke of translations. In doing so, he continued to understand reduction as a semantic procedure by means of which entire theories were ultimately supposed to be reducible to more basic theories. As tools for such a “semantic reduction” – or, better put: explication – Carnap considers conditional propositions of the following sort: “If certain conditions are fulfilled, then the following holds: Mr. A has toothaches precisely when Mr. A satisfies such-and-such physiological and behavioral criteria.” The accompanying optimistic assumption was that criteria can be found which unequivocally regulate the application of the predicate “toothaches” under the presupposed conditions (see Carnap 1938, reprint 418ff). This procedure of explication is far more subtle than the simple idea of a translation with which Carnap originally attempted to realize the program of Unified Science. But nothing vital has changed about setting the goal itself: as before, a semantic procedure was supposed to prepare the ground for a unification of science, both methodically and in terms of content. As applied to psychology, this meant that everything that can be known by means of an introspective method would also have to be knowable by application of behavioristic (or physiological) methods. Beyond this, Carnap also hoped to be able to show that a derivation of psychological laws from physical laws is in principle possible (see Carnap 1938, reprint 421f). Carnap’s explication-procedure remained just as hopeless as far as these demands were concerned as his original translation project. To be sure, the translation method poses problems that the explication-method avoids, such as, e.g., the problem that most of the concepts of science are clusterconcepts that cannot be defined by means of necessary and sufficient conditions. But the objections that target both procedures remain. To these belongs foremost the objection that psychological concepts are not definable or explicable in terms of physical concepts without a vicious circle. If, deduced from sentences of the “language of observables”, “but at best inferred with high probability” (Carnap 1947, 230).

A Plea for a Naturalistic Ontology

85

e.g., one attempts to adduce the conditions under which the expression ‘toothache’ is applicable, one cannot avoid getting started from presuppositions which in turn can only be described by way of psychological concepts, e.g., from the presupposition that a test-subject does not intend to feign pains, and that he or she believes that medications can alleviate pains, etc. Hence, also Carnap’s refined procedure of semantic reduction cannot show how psychological propositions can be traced back to purely physical propositions, but rather only how we can explicate them with the aid of propositions which, aside from physical concepts, also contain psychological ones.12 Quine was the first to recognize in full sharpness this problem of circularity. His attack on the “second dogma of empiricism” explicitly targets Carnap’s reductionism. Interestingly enough, Quine criticizes there not Carnap’s attempt to establish the physical language as the universal language of science, but rather his earlier enterprise in Der logische Aufbau der Welt (1928) to reduce all empirical propositions to a language of sensory data. However, since his critique is independent of the choice of a specific reduction-basis, it can also be carried over without curtailments onto Carnap’s new physicalist program. It is also noteworthy that in his critique Quine already distinguishes sharply between the “radical reductionism,” that above I termed as “naive,” and the “subtle reductionism” in Carnap’s later works. For Carnap radical reductionism was already done with, since Carnap himself no longer adhered to it.13 It is on Carnap’s later reductionism that Quine sets his sights: 12

A proposal for avoiding such circularities introduced by David Lewis offers here a possible way out, in that one begins from a complete psychological theory of a person, transforms this theory into a Ramsey-sentence, and in this fashion advances a noncircular condition of application for every psychological term. But since what must be involved is a theory that contains an enormous array of information, this procedure is surely to be understood as an alternative but not as a rescue of Carnap’s sematic reductionism. See Lewis (1972). 13 “Radical reductionism, conceived now with statements as units, sets itself the task of specifying a sense-datum language and showing how to translate the rest of significant discourse, statement by statement, into it. Carnap embarked on this project in the Aufbau.” (Quine 1951, 36). One could certainly question whether Quine here gives a plausible interpretation of Carnap’s Aufbau. More recent interpretations show that Carnap’s theory of constitution cannot be understood as easily as the exposition of the old empiricist program – as Quine assumed it could (see Mormann 2000, 83ff).

86

Johannes L. Brandl But the dogma of reductionism has, in a subtler and more tenuous form, continued to influence the thought of empiricists. The notion lingers that to each statement, or each synthetic statement, there is associated a unique range of possible sensory events such that the occurrence of any of them would add to the likelihood of truth of the statement. (Quine 1951, 38)

Quine does not object here to the verification theory of meaning that surfaces once again, but to the atomism bound up with it, i.e., the assumption “that each statement, taken in isolation from its fellows, can admit of confirmation or infirmation at all” (Quine 1951, 38). It is in this that the “dogma of reductionism” consists for him, which cannot be justified by the praxis of science. This holistic objection is intimately bound up with the problem that psychological predicates cannot be explicated free of a vicious circle. If an explication free of circularity were possible, it would show how a single proposition can also be confirmed in isolation on the basis of observational data (or sensory experiences). Conversely, the fact that such confirmations are only possible holistically adduces a reason to doubt that the search for explications free of circularity will be successful. These two arguments taken together constitute the principal reason why Carnap’s semantic reductionism, even in its subtler form, is today considered as having failed. What now does this mean for the thesis of Unified Science, to the “making precise” of which Carnap developed his program? It means that one can continue to pursue this idea only if one liberates it from the reductionist dogma. I have already indicated how such an understanding of the unity thesis might look by alluding to Neurath: It could consist in the empirical sciences’ having a unified domain of objects. The thesis of Unified Science would then have to be interpreted ontologically, not semantically or methodologically. Needless to say, one cannot draw from this the conclusion that no constraints at all can be imposed on the language of science and on scientific methodology. But such constraints are only acceptable insofar as they can be ontologically explained. It must be shown that the particular objects investigated within a science call for a language and methodology suited to these objects. It is a matter of concern in this connection that the concept of “ontological uniformity” can be understood not absolutely, but rather only relatively. Indeed, object domains are not straightforwardly uniform or non-uniform, but can be regarded in various respects as ontologically differentiated. For this reason one can subsequently only speak of a unity

A Plea for a Naturalistic Ontology

87

within science in the sense that there is one specific aspect with regard to which the various object domains appear uniform. I wish to make these abstract deliberations more vivid by way of an example. As already mentioned, Quine defends the principle of extensionality as a principle which, in his opinion, a scientific language ought to satisfy (see Quine 1994). One could therefore expect that Quine’s attempt to show how all scientific propositions can be translated into propositions of an extensional language, or how all concepts can be explicated in terms of concepts that are expressible in such a language. Were Quine to proceed in this fashion, he would clearly expose himself to the same objections that he leveled against Carnap’s semantic reductionism. But how else can Quine back up his preference for extensional languages? In my view, he is only left with an ontological argument. Quine’s appeal must be that the ontology of scientific theories consists exclusively of four-dimensional individuals and sets of such individuals. An extensional language is sufficient for describing such entities, and therefore we can forgo intensional modes of expression in science. This argument stands and falls with the assumption that reality consists only of four-dimensional individuals and their sets. Were Quine to also accept properties as a separate ontological category, it would pull the carpet from under his extensionalism. For predicates can be co-extensional even if they designate (or express) different properties. Likewise, the truth of a proposition does not depend only on whether an object belongs to the extension of a predicate. Thus, it can be true, e.g., that Oedipus believes himself to be embracing his beloved, but does not believe himself to be embracing his mother, even though the predicates “my beloved” and “my mother” refer in this case to the same person. Now Quine can certainly defend his extensionalism by conceding that reality can be regarded from a point of view which makes the application of non-extensional languages superfluous. We need only relinquish treating Oedipus’ convictions as something relevant, and confine ourselves instead to what Oedipus actually does. It then turns out that the embrace of his mother is identical with the embrace of his beloved. But what has been gained with this? The uniform describability of reality would in this case be a trade-off, in that it is constrained to a particular perspective of worlddescription. And what is this constraint supposed to justify? To my mind, a consistent Naturalist has to forge a different path here. The possibility of being able to regard objects from various perspectives

88

Johannes L. Brandl

implies that we must also come to terms with a correlative diversity in our ontology. The question “What is there?” is resolved in the sciences themselves; it cannot be decided through the philosophical project of Unified Science. But then what still remains of this project? It is quite surprising, but certainly no accident, that one can find the answer to this question also in Carnap. It is the idea that reality has a unitary “structure,” tied to the idea that this unitariness can be demonstrated by means of a system of constitution. This seems to me to be for Naturalism the fruitful kernel from Carnap’s philosophy, which can also be rediscovered in the program of Unified Science. The interpretation of the notion of constitution is a topic onto itself, into which I cannot go any further here. I therefore propose, without further substantiation, a formulation of this notion, which, in my view, is consonant with the spirit of Naturalism: Thesis of the Unitary Constitution of Reality: The elementary constituents of reality – including the elementary mental events – form the basis of a unified ontological order, in which more complex unities and their properties are constituted by means of less complex unities and their properties.

This thesis undermines every attempt to enrich reality with events or entities whose existence remains puzzling to us. To these belong magical or esoteric phenomena, as well as objects of a transcendental order that do not have any kind of constitutive relationship to the elementary constituents of reality. To be sure, we do not know with certainty what these elementary constituents of reality are, and therefore every attempt to establish the constitutive unity of reality must remain a preliminary attempt. This, however, is part of how the Naturalists understand themselves. Whoever expects philosophy – especially ontology – to identify the elementary constituents of the world would contradict the anti-fundamentalist thesis of Naturalism. 4. Conclusion In this paper, I have tried to provide Naturalism with both an epistemological and ontological lynchpin. In my estimation, both theses are accessible to a substantive critique and need not be object of polemical debates. The theses differ radically, however, in the claims inherent in them. The anti-

A Plea for a Naturalistic Ontology

89

fundamentalist thesis is, methodically considered, an appeal to philosophical restraint. Since it is not the task of philosophy to come up with theories that are intersubjectively verifiable on the basis of experience, neither can its mission be to indulge in a program whose goal is to found science by means of cognitions pertaining to the elementary constituents of the world. On the other hand, the thesis of the unitary constitution of reality is a challenge to philosophy, especially to ontology. Given the enormous diversity of object domains that are scientifically investigated today, it is by no means settled that this thesis is cogent. The substantive discussion should revolve here around the question of which fundamental problems loom when one claims that such complex entities as, e.g., human persons are capable of being constituted by means of processes or entities with “subpersonal” complexity. That is a formidable task for a naturalistically inspired ontology. It requires the kind of honesty that Russell demanded in philosophy of mathematics: “The method of ‘postulating’ what we want has many advantages; they are the same as the advantages of theft over honest toil. Let us leave them to others and proceed with our honest toil.” (Russell 1919, 71).14 References Almeder, R. 1998 Harmless Naturalism. The Limits of Science and the Nature of Philosophy, La Salle: Open Court. Carnap, R. 1928 Der logische Aufbau der Welt. Berlin: Weltkreis Verlag. -

1931 “Die physikalische Sprache als Universalsprache der Wissenschaft”, Erkenntnis 2, 432-465.

-

1932 “Psychologie in physikalischer Sprache”, Erkenntnis 3, 107-142.

-

1938 “Logical Foundations of the Unity of Science”, International Encyclopedia of Unified Science I, reprint 1949 in Feigl, H. and Sellars W. (eds.) Readings in Philosophical Analysis, New York: Appleton-Century-Crofts, 408-422.

-

1947 Meaning and Necessity. A Study in Semantics and Modal Logic, Chicago: University of Chicago Press.

14

The present text has been translated from German into English by Arthur Szylewicz.

90

Johannes L. Brandl

Craig, W.L. and Moreland, J.P. (eds.) 2000 Naturalism. A Critical Analysis, London: Routledge. Decock, L. and Horsten, L. (eds.) 2000 Quine. Naturalized Epistemology, Perceptual Knowledge and Ontology, Amsterdam: Rodopi. French, P.A., Uehling, T.E. Jr. and Wettstein, H.K. (eds.) 1994 “Philosophical Naturalism”, Midwest Studies in Philosophy 19, Notre Dame: University of Notre Dame Press. Gibson, R.F. 1982 The Philosophy of W.V. Quine, Tampa: University Presses of Florida. Hookway, C. 1988 Quine, Cambridge: Polity Press. -

2000 “Naturalism and Rationality”, in L. Decock and L. Horsten (eds.) Quine. Naturalized Epistemology, Perceptual Knowledge, and Ontology, Amsterdam: Rodopi, 35-55.

Hornsby, J. 1997 Simple Mindedness. In Defense of Naive Naturalism in the Philosophy of Mind, Cambridge, MA.: Harvard University Press. Hylton, P. 1994 “Quine’s Naturalism”, in P. French et. al. (eds.), Midwest Studies in Philosophy, Notre Dame: University of Notre Dame Press, 261-282. Keil, G. and Schnädelbach, H. (eds.) 2000 Naturalismus. Philosophische Beiträge. Frankfurt: Suhrkamp. Kitcher, P. 1992 “The Naturalists Return”, Philosophical Review 101, 53-114. Lewis, D. 1972 “Psychophysical and Theoretical Identification”, Australasian Journal of Philosophy 50, 249-258. Mormann, T. 2000 Rudolf Carnap, München: Verlag C. H. Beck. Quine, W.V. 1951 “Two Dogmas of Empiricism”, Philosophical Review 60, reprinted in W.V. Quine, From a Logical Point of View, Cambridge Mass.: Harvard University Press, 20-46. -

1960 Word and Object, Cambridge: MIT Press.

-

1966 “Three Grades of Modal Involvement”, in W.V. Quine, The Ways of Paradox and Other Essays, New York: Random House, 151-176.

-

1969 Ontological Relativity and Other Essays, New York: Columbia.

-

1969a “Epistemology Naturalized”, in W.V. Quine, Ontological Relativity and Other Essays, New York: Columbia, 69-90

-

1969b “Ontological Relativity”, in W.V. Quine, Ontological Relativity and Other Essays, New York: Columbia, 26-68.

-

1986 “Replies to Putnam”, in L.E. Hahn and P.A. Schilpp (eds.) The Philosophy of W.V. Quine, La Salle: Open Court, 427-431.

-

1981 Theories and Things, Cambridge, MA: Harvard University Press.

A Plea for a Naturalistic Ontology

91

-

1981a “Five Milestones of Empiricism”, in W.V. Quine, Theories and Things, Cambridge, MA: Harvard University Press, 67-72.

-

1994 “Promoting Extensionality”, Synthese 98, 143-151.

-

1995 “Naturalism; or, Living within One’s Means”, Dialectica 49, 251-263

Quitterer, J. and Runggaldier, E. (eds.) 1999 Der Neue Naturalismus – Eine Herausforderung an das christliche Menschenbild, Stuttgart: Kohlhammer. Russell, B. 1919 Introduction to Mathematical Philosophy, London: Routledge. Schilpp, P.A. 1963 The Philosophy of Rudolf Carnap, La Salle: Open Court. Villanueva, E. (ed.) 1993 Naturalism and Normativity, Philosophical Issues, Atascadero: Ridgeview. Wagner, S.J. and Warner, R. (ed.) 1993 Naturalism: A Critical Appraisal, Notre Dame: University of Notre Dame.

Ockham on the Nature of Concepts GERHARD LEIBOLD, Innsbruck

1. Introductory remarks It is well known that Ockham went through a remarkable development in his treatment of the nature of concepts (i. e. conceptus universales, intentiones animae, passiones animae).1 This is one of the few areas in which Ockham, in different periods of his writing, oscillates between different views in his philosophical opinions. It is important to note that such uncertainties are limited to questions about the nature of concepts, or, in another terminology, to the ontological status of concepts, but not to Ockham’s unwavering and consistent criticism of realistic positions of different shades of various realistic positions. Among the authors who took up this question is Vladimir Richter (1975, 1998). He discussed chronological problems, the time of writing, and the authenticity of Ockham’s nonpolitical works. Richter based his thoughts on the investigations of Boehner (1942, 1946, 1951, 1958), Gál (1967), Brampton (1963), and Weisheipl (1968). Following Richter, I first recapitulate the results of previous research. In a second step, I expose Ockham’s different positions regarding the nature of concepts. 2. Results of previous research In previous research, there is a high degree of widespread consensus about the development of Ockham’s ideas on the nature of concepts.2 Most scholars agree that one can distinguish four phases3:

1

I am grateful to my admired collegue Christoph Jäger (Innsbruck), who was a great help for me to finish this contribution. The remaining mistakes are my own responsibility. 2 G. Leff (1975), pp. 78-94, has suggested some corrections to Boehners chronology. 3 I essentially follow Richter’s representation in: Richter/Leibold (1998), pp. 106110.

94

Gerhard Leibold

(i) During the time of lecturing about the Sentences (Oxford 1317-1319) Ockham defended the so called fictum theory. He treats this theory most detailed in Quaestio 8 of the 2nd distinction of the Scriptum in primum librum Sententiarum Ordinatio (OTh II, pp. 266-292). (ii) During the time of commenting on the logical works and the Physics of Aristotle (1320-1324) he took up a “disjunctive”4 position between the fictum theory and the undifferentiated qualitas theory. This position is extensivly presented in the early redaction of the disgression about the problem of universals at the beginning of the Expositio in Perihermenias (OPh II; §§ 9-10, pp. 363-371). (iii) Another phase is marked by Ockham’s preference of the intellectio theory (cf. Gál 1967). This position is considered in the late redaction of the digression about the problem of universals in the Prooemium of the Perihermenias-commentary (OPh II, § 6, pp. 351-358), in several additions of the Scriptum in I librum Sententiarum (cf. Boehner, 1951, pp. 305-316 and 1958, pp. 96-110) and in the Summa logicae (OPh I, p. 1, c. 15, p. 53). (iv) A radicalization of all previous positions is to be found in the Quodlibeta (Quodlibet IV, q. 35, OTh IX, pp. 469-474) and in the Quaestiones in libros Physicorum (qq. 1-7, OPh VI, pp. 397-412). In these texts the fictum theory is firmly rejected. 3. The fictum theory in the Ordinatio Ockham’s earliest view of the nature of concepts in the Ordinatio is the socalled fictum theory.5 In this account, he treats concepts as part of the problem of universals. At the beginning of quaestio 8, four opinions are discussed, but dropped in favour of the fictum theory, which is called “probable”. The first of these

4

With such words this position is characterized by his brethren and critic Walter Chatton in the so-called Reportatio (probably 1322-1323): „aut significat fictum, quale frequenter ponis, saltem sub disiunctione, et aliquando tenuisti absolute“. (All references in Richter/Leibold (1998), p. 106). 5 Ord. I, d. 2, q. 8 (OTh I, pp. 266-292). Cf. to this passage: G. Leff (1975), pp. 9498, and M. McCord Adams (1987), pp. 73-105.

Ockham on the Nature of Concepts

95

four opinions is the intellectio theory6, the second is the species theory.7 According to the third opinion the universal is a real thing which follows the act of the intellect.8 According to the fourth opinion the universal is nothing from nature, but exists only by virtue of convention.9 Ockham himself prefers the fictum theory. “So one can probably say that the universal is nothing real that has a subjective being neither within nor outside the soul, but has only an objective being in the soul. And it is a fictum that has such a being in the objective being as a thing outside the soul in the subjective being.“10 As a fictum, a concept inheres in the soul. It is an object of the mind that lacks independent existence within or outside the soul. Already in this passage it is evident that Ockham uses the concepts ‘subjective’ and ‘objective’ in a sense quite different from modern use. Following Aristotle, Ockham uses ‘subjective’ to denote something real outside the soul. ‘Objective’, by contrast, are things within the soul, something the act of the intellect “puts before itself” as an image in the soul. A fictum emerges in the soul through imagination (“fingere”) on the part of the intellect, when the latter sees something outside and causes a similar thing to occur within the soul. If the intellect had a productive power, as it has an imaging power, it would create such a thing in the subjective being outside the soul; a thing that would differ only in number from the thing in the soul.11 A frequently cited example is that of a builder who sees a real 6

Ibid., pp. 267 s.: « Prima opinio posset esse quod universale est conceptus mentis, et quod ille conceptus est realiter ipsa intellectio, ita quod tunc universale non esset nisi intellectio confusa rei, ... » 7 Ibid., p. 269: „Secunda opinio posset esse quod universale est species aliqua …“ 8 Ibid., p. 269: „Alia posset esse opinio quod est aliqua vera res sequens actum intellectus, quae esset similitudo rei … » 9 Ibid., p. 271: „Quarta posset esse opinio quod nihil est universale ex natura sua sed tantum ex institutione … « 10 Ibid., pp. 271 s.: „Ideo potest aliter dici probabiliter quod universale non est aliquid reale habens esse subiectivum, nec in anima nec extra animam, sed tantum habet esse obiectivum in anima, et est quoddam fictum habens esse tale in esse obiectivum quale habet res extra animam in esse subiectivo.“ 11 Ibid., p. 272: „Et hoc per istum modum quod intellectus videns aliquam rem extra animam fingit consimilem rem in mente, ita quod si haberet virtutem productivam sicut habet virtutem fictivam, talem rem in esse subiectivo – numero distinctam a priori – produceret extra. »

96

Gerhard Leibold

house. In his soul, a "mapping" appears that resembles the house outside the soul and then in turn causes a house like this to emerge in reality.12 Ockham also calls this fictum a universal because it is an image (“exemplar”) that refers equally to all the individual things outside the soul, and can stand for other things outside because of its similarity in the objective existence of things outside the soul.13 So, Ockham's reasoning can be formulated as follows14: (i) We can think of objects of various sorts that either do not really exist or cannot exist, for example of chimeras. But we also think of objects of logic or merely possible objects that God could have created but did not. (ii) If we think of something then it has some sort of existence, which is either objective existence or real existence. (iii) Objects of the kind described in (i) have objective existence. (iv) Mental acts or qualities, and the mind itself, have real existence. (v) Therefore, objects of the kind described in (i) are distinct from mental acts, mental qualities, and the mind itself. Adhering to the above analysis, it is quite natural to extend this idea to things that have real existence, so that really existing things also have an objective existence in the mind, provided they are thoughts. In other words, what can be thought has objective existence.15

12

Ibid. Ibid.: „Et illud potest vocari universale, quia est exemplar et indifferenter respiciens omnia singularia extra, et propter istam similitudinem in esse obiectivo potest supponere pro rebus extra …“ 14 Cf. McCord Adams (1987), pp. 76 s. 15 McCord Adams (1987, pp. 96-102) points to a modern objection of W.V.O. Quine. He expresses general suspicion of, and scorn for, the doctrine that whatever is thought of has some sort of being – if not real, then unreal: ‚From a Logical Point of View (Cambridge, MA, 1953), p. 1 s.: „This ist the old Platonic riddle of non-being. Nonbeing must in some sense be, otherwise what is it that there is not? This tangled doctrine might be nicknamed Plato’s beard: historically it has proved tough, frequently dulling the edge of Occam’s razor. It is some such line of thought that leads philosophers like McX to impute being where they might otherwise be content to recognize that there is nothing.” 13

Ockham on the Nature of Concepts

97

Ockham's theory of objective existence analyzes the intellectual act into (i) a really existing mental act and (ii) an objectively existing object. But according to Ockham the objective existence of something doesn’t reduce to the actual existence of something else. Concepts are thus ficta that only exist as objects of the acts of knowledge in the soul. They are not real things, but can stand for individual things outside the soul. From a historical perspective, we should draw attention to the Augustinian influence on Ockham’s fictum theory (cf. Leff, 1975, pp. 9698). Ockham refers to various books of De Trinitate of Augustine and – firstly – takes over the notion that the mind forms images of things of which it has knowledge but which it does not actually encounter. These images are derived from experiences of real things. The fictum – secondly – does not stand for itself but for the things outside the soul. Thirdly, Augustine treats all terms as representations of real things, calling them variously “image”, “likeness”, “picture”, “phantasm”, “species”. Finally, ficta reside in memory (memoria). They therefore constitute potential knowledge. Qua concepts, ficta differ in their nature from the nature of the things they represent. They are common because of their similarity to real things. On the other hand, ficta have to be distinguished from logical terms, which also constitute universals (cf. Leff, 1975, pp. 97 s.). Logical terms are results of convention, not of nature. Ficta are abstracted in a natural way from real individuals. This is essentially Ockham's discussion of the fictum. In relation to the four different opinions that were mentioned at the beginning of this passage there can be little doubt that he favours this fictum theory and prefers it over the other possible explanations of the nature of concepts. It is important to note however that Ockham never identifies himself with the fictum theory: He never calls it more than “probabilis”. 3.1 Objections to the fictum theory Ultimately, Ockham abandoned the fictum theory. Probably, Ockham’s brethren Walter Chatton strongly influenced this change, as for example G. Gál has convincingly argued.16 Chatton seems to have convinced Ockham 16

Cf. Gál (1967), pp. 191-212. Cf. also: McCord Adams (1987), pp. 102-105.

98

Gerhard Leibold

that the objective existence theory violates the principle of parsimony, better known as Ockham’s "razor". According to Ockham's early expositions of the nature of concepts, objective existence pertains only to what is thought of by someone. Ockham does consider Robert Grosseteste’s objection that in conseqence there must be as many universals as intellects that think these of.17 However, Ockham brushes this objection aside, replying that he does not care whether or not a figmentum, or concept, is varied as the intellects are varied, so long as the concepts signify equivalently.18 Chatton however, insists that Ockham must make the assumption that there are as many universals as intellects. He maintains that if the objective existence of a thing in the soul does not logically presuppose the real existence of a cognitive act, then it cannot presuppose the real existence of a cognitive act generally (Gál, 1967, p. 202, lin. 2-12). In the Quodlibeta – be they a work of Ockham or not – there is an argument that agrees with Chatton.19 Even so, however, it is true that positing, in addition to really existent acts of the intellect, a coordinate realm of objectively existing entities violates Ockham's principle of parsimony. As long as Ockham maintains the (first) premise, these objectively existing entities serve no necessary theoretical function. If we think of something, then it has some sort of existence (either actually or objectively). Thus, objectively existing entities are required, for we sometimes think of things as having properties that no real things can have. Marilyn McCord Adams (1987, p. 104) says, convincingly, that “Chatton argues that objectively existent entities are superfluous because they do not fill the theoretical role, while the real things posited by the theory can

17

Ordinatio I, d. 2, q. 8 (OTh II, p. 284): „Si dicatur quod secundum Lincolniensem universale non est figmentum: Similiter tunc tot essent universalia quot sunt intellectus.“ 18 Ibid. (OTh II, pp. 284 s.): „ … sive figmentum vel conceptus varietur ad variationem intellectuum sive non, non curo ad praesens –, dicerent quod genus generalissimum substantiae vel est unum simpliciter non variatum vel est unum per aequivalentiam, … sed tamen est eadem per aequivalentiam, ut dicunt et bene …” Cf. again McCord Adams (1987), p. 103. 19 Quodlibeta IV, q. 35 (OTh IX, pp. 472 s.). Regarding the problem of the authenticity of the Quodlibeta cf. Richter/Leibold (1998), pp. 110-117.

Ockham on the Nature of Concepts

99

do so, unassisted by such objectively existing entities.” 20 Chatton concludes that if the real existence of an intellective act is logically independent of the existence of real things, then it must be logically independent of the objective existence of anything. McCord Adams continues (ibid.): “His (i. e. Ockham's) second premise is: ‘It is a contradiction to suppose that there is an act of intellect unless something - namely, its term – is understood’ (Gál, 1967, p. 202, lin. 2224). Necessarily, every act of thought has an object. If it is not necessary that an act of thought is directed towards some objectively existing entity ... , then that the thought is directed towards the objectively existent entity and towards one rather than another, cannot be what accounts for its being a thought of something and of this rather than that.” Chatton’s view is this: “If we think of a round square or a chimera, it is not necessary that there be (in some sense) anything that is a round square or a chimera” (McCord Adams, 1987, p. 105). He writes: “The mind can imagine (“fingere”) many things, such as a golden mountain, a chimera, etc. But this is nothing other than for it to have an act of understanding by virtue of which I could judge that its nature would be of this sort if it existed” (Gál, 1967, p. 207, lin. 2529). Whatever one may think about the authenticity of the Quodlibeta, the upshot is that in this work the distinction between objective and real existence is given up and concepts are identified with real existent acts of intellect. 4. The “disjunctive” position in the Expositio in librum Perihermenias As mentioned above, Ockham has taken a “disjunctive” position between the (undifferentiated) qualitas theory and the fictum theory in the previous redaction of his Expositio in librum Perihermenias (OPh II, § 9, pp. 363369). After previously having rejected the views of the reales 21, he describes the qualitas theory as follows: “There can be another opinion I hold to be probable, namely, that the passions of the soul are qualities of the 20

Cf. Gál, 1967, p. 202, lin. 21 s. (OPh II, § 8, p. 363): „Sed istam opinionem quantum ad hoc quod ponit esse aliquas res extra praeter singulares exsistentes in eis reputo omnino absurdam et destruentem totam philosophiam Aristotelis et omnem scientiam et omnem veritatem et rationem ...” 21

100

Gerhard Leibold

mind existing subjectively in the mind, so truly and realy how the white in the wall or the cold in the water.”22 Ockham supports this theory by references to Averroes and Avicenna, and discusses it in the form of objections and answers. Finally, it is said that this theory is acceptable in two ways: as intellectio theory or as so-called "terminating" qualitas theory.23 Ockham is not committed to any of these theories. In the third place, Ockham mentions the fictum theory and describes it similarly to the way he portrays it in the Ordinatio.24 At the end of his digression on the question of the nature of concepts in the Expositio in librum Perihermenias, Ockham emphasizes his disjunctive position regarding the qualitas theory and the fictum theory and repeats his rejection of the realistic accounts : “So I hold these last opinions to be probable. Which is true, however, and which is false, zealous people should discuss. But this is for me absolutely certain that neither passions of the soul nor universals are something outside the soul and of the essence of individual things, whether they are grasped or not.”25

22

Ibid., (OPh II, § 9, p. 363): „Alia potest esse opinio quam reputo probabilem quod passiones animae sunt quaedam qualitates mentis exsistentes subiective in mente ita vere et ita realiter sicut albedo exsistit in pariete vel frigus in aqua.” 23 Ibid., (OPh II, § 9, p. 369): „Sic igitur potest teneri quod passiones animae et omnia praedicabilia et subicibilia et universaliter omnia universalia … sunt qualitates mentis reales, … et hoc quia sunt intellectiones, secundum unam opinionem, vel quaedam qualitates animae terminantes tamquam obiecta intellectiones animae. Et secundum istam opinionem potest probabiliter teneri quod omnis propositio in mente quae non componitur ex signis ad placitum institutis componitur ex realibus qualitatibus mentis ...” 24 Ibid., (OPh II, § 10, p. 370): „Tertia opinio principalis posset esse …quod passiones animae de quibus loquitur hic Philosophus et propositiones et syllogismi et universalia omnia non sunt nisi quaedam ficta in anima habentia tantum esse obiectivum, hoc est esse cognitum, nullibi exsistentia realiter. Quae ficta vocari possunt pro quanto non sunt entia realia. Et tunc suponenda aliqua entia rationis distincta universaliter contra omnia entia realia et contra omnia exsistentia in praedicamentis.” 25 Ibid., (OPh II, § 10, p. 371): „Sic igitur istas ultimas opiniones reputo probabiles. Quae tamen sit vera et quae falsa discutiant studiosi. Hoc tamen apud me omnino certum est quod nec passiones animae nec universalia sunt aliqua ex anima et de essentia rerum singularium sive sint conceptae sive non conceptae.”

Ockham on the Nature of Concepts

101

5. Ockham’s preference of the intellectio theory This phase of Ockham’s development is represented in the later redaction of the digression about the question of the nature of concepts in the Perihermenias-Commentary, in several additions of the Ordinatio, and in the Summa logicae. The most detailed exposition is again the presentation in the Prooemium of the Expositio in librum Perihermenias. After a brief rejection of the terminated qualitas theory (OPh II, § 4, pp. 349 s.) and the species theory (OPh II, § 5, pp. 350 s.), Ockham begins his remarks about the intellectio theory, which he describes as "more probable" than all other theories.26 This is followed by a discussion of this theory, finally Ockham formulates a clear position in favor of the intellectio theory.27 The fictum theory is again treated and objections are considered (OPh II, § 7, pp. 359-361). Finally, it is said that the qualitas theory is certainly preferable to the fictum theory .28 Moreover, the intellectio theory is preferred to the qualitas theories (see Richter & Leibold 1998, pp. 108 s.). A similar position is taken in the Summa logicae (OPh I, p.1, § 12, pp. 41-44). The fictum theory, the terminating qualitas theory, and the intellectio theory are mentioned. With Okham’s "novacula" the author argues against the first two theories, and indicates a preference for the intellectio theory.29

26

Ibid., (OPh II, § 6, p. 351): „Alia posset esse opinio quod passio animae est ipse actus intelligendi. Et quia ista opinio videtur mihi probabilior de omnibus opinionibus quae ponunt istas passiones esse subiective et realiter in anima tamquam veras qualitates ipsius, ideo circa istam opinionem primo ponam modum ponendi probabiliorem, si debeat poni.” 27 Ibid., (OPh II, § 6, p. 358): „Sic igitur, qui vellet, posset tenere istam opinionem: quod passiones animae de quibus loquitur Philosophus, sunt intellectiones, quae est opinio probabilis et concordat in ista conclusione communi cum praecedentibus quod passiones animae sunt verae qualitates mentis.“ 28 Ibid., (OPh II, § 7, p. 361): „Igitur propter istas rationes magis debet poni quod tales passiones animae de quibus loquitur hic Philosophus, sunt qualitates mentis quam quod sunt talia idola sive ficta.“ 29 (OPh I, p. 1, § 12, pp. 42 s.): „Et pro istis est ratio ista quia ‚frustra fit per plura quod potest fieri per pauciora’. Omnia autem quae salvantur ponendo aliquid distinctum ab actu intelligendi possunt salvari sine tali distincto, eo quod supponere pro alio

102

Gerhard Leibold

6. The radicalization of the intellectio theory It has frequently30 been noticed that in the Quodlibeta and the Quaestiones in libros Physicorum Ockham radicalizes various of his views, among them also his position regarding the nature of the terms.31 In these works the fictum theory, for example, is sharply rejected. If a fictum – it is said here – is known, then the appropriate thing outside the intellect is not known. So we would know only the fictum but not the real thing to which a fictum is related, which is absurd.32 However, the authenticity of these works is disputed. For the Quaestiones my own researches suggest that – in a strict sense of authorship – Ockham is not the author of this work. For the Quodlibeta – in my view – the situation is similar.33 7. Conclusions It is important to distinguish four phases in Ockham’s treatment of the nature of concepts, as Richter (1998, p. 110) has done in perfect clarity. Richter refers to Boehner34, who has stressed that the intellectio theory et significare aliud ita quod potest competere actui intelligendi sicut alii signo. Igitur praeter actum intelligendi non oportet aliquid aliud ponere.” 30 For example Hochstetter (1927), p. 10: „Die Härte und Unbedenklichkeit der Thesen in den Quodlibeta legt es m. E. nahe, sie vom Sentenzenkommentar räumlich und zeitlich stärker abzurücken, als das bei ihrer Verlegung in die Oxforder Periode möglich wäre. Auch fehlt in ihnen – wie übrigens auch in der Logik – das Schwankende, Tastende in seinen Behauptungen, das in entscheidenden Lehren die Sentenzen noch zeigen. Das alles spräche dafür, vorläufig die Zeit ihrer Ausarbeitung in die Avignoner oder Münchener Periode zu legen.“ Cf. also Richter/Leibold (1998), p. 110. 31 For example Quodlibet IV (OTh IX, q. 19 and q. 35, pp. 395-398 and 469-474). 32 Quodlibet IV, q. 35 (OTh IX, p. 473): “Praeterea tale fictum impedit cognitionem rei. Ergo non est ponendum propter cognitionem ... Ergo si illud fictum intelligitur, tunc res extra non intelligitur ... quod est absurdum.“ 33 It is incomprehensible to me why M. McCord Adams, in her opus magnum of 1987, in a nearly naive manner assumes the authenticity of the Quodlibeta and the Quaestiones in libros Physicorum. 34 Boehner (1958, p. 172): „It seems to be more natural to suppose that Ockham, when he was shaken in his belief in the fictum-theory and had become more inclined towards the qualitas theory, did not yet have a clear idea as to the nature of this quality. But he soon realized that, in accordance with his whole system and with the principle of economy, the intellectio theory could best satisfy him.”

Ockham on the Nature of Concepts

103

corresponds very well to Ockham’s linguistic principles, especially to his principle of economy – with good reasons, I think. In the Quodlibeta and the Quaestiones in libros Physicorum (plus ultimately in the Summa logicae as well) the intellectio theory is portrayed as the only viable position. But on the personal and direct authorship of these works one must cast doubt – despite the arguments of the editors. The disjunctive position regarding the qualitas theory (which Ockham ascribes to Aristotle) and the early fictum theory turns out to be Ockham's original conception. Literature Works of Ockham: Guillelmi de Ockham Scriptum in libros Sententiarum Ordinatio. Editio St. Bonaventure, N.Y., 1967 and 1970 (OTh I and II). Guillelmi de Ockham Quodlibeta septem. Editio St. Bonaventure, N.Y., 1980 (OTh IX). Guillelmi de Ockham Summa logicae. Editio St. Bonaventure, N.Y., 1974 (OPh I). Guillelmi de Ockham Expositio in librum Perihermenias Aristotelis. Ed. St. Bonaventure, N.Y., 1978 (OPh II). Guillelmi de Ockham Quaestiones in libros Physicorum Aristotelis, Ed. St. Bonaventure, N.Y., 1984 (OPh VI).

Literature of research: Boehner, Ph., (1942), The Text Tradition of Ockham’s Ordinatio, in: The New Scholasticism 16, pp. 203-241. -

(1946), The Realistic Conceptualism of William Ockham, in: Traditio 4, pp. 307-335.

-

(1951), The Relative Date of Ockham’s Commentary on the Sentences, in: Franciscan Studies 11, Comm. Vol, pp. 305-316. All these contributions reprinted in:

-

(1958), Collected Articles on Ockham, St. Bonaventure, N.Y., pp. 96-110, 110-127, 156-174.

Brampton, C. K., (1963), The Probable Order of Ockham’s Non-polemical Works, in: Traditio 19 pp. 469-483. Gál, G., (1967), Gualteri de Chatton et Guillelmi de Ockham Controversia de natura conceptus universalis, in: Franciscan Studies 27, pp. 191-212.

104

Gerhard Leibold

Hochstetter, E., (1927), Studien zur Metaphysik und Erkenntnislehre Wilhelms von Ockham, Berlin and Leipzig. Leff, G., (1975), William of Ockham. The Metamorphosis of Scholastic Discourse, Manchester. McCord Adams, M., (1987), William Ockham, Vol. 1, Notre Dame. Quine, W. V. O., (1953), From a Logical Point of View, Cambridge, MA. Richter, V. (1975), Zu Ockhams Entwicklung in der Universalienfrage. Bemerkungen im Zusammenhang mit dem Problem der Chronologie, Abfassungszeit und Authentizität von Ockhams nichtpolitischen Schriften, in: Philosophisches Jahrbuch 82, pp. 177-188. Reprinted in: Richter, V. / Leibold, G. (1998), Unterwegs zum historischen Ockham (Mediaevalia Oenipontana, Vol. 1), Innsbruck, pp. 105-117.

What is Aristotelian Metaphysics? FRIEDO RICKEN, Munich

It is well known that the term “Metaphysics”, under which fourteen of his writings have been passed on, does not go back to Aristotle himself; it is only documented for the 1st century B.C.. The traditional ordering is a product of Andronicus of Rhodes (first half of the 1st century B.C.). Aristotle himself refers to these writings as “first philosophy” (De motu an. 700b8). What does he understand by this science which was later named “metaphysics”? What questions does it deal with? By which methods are they supposed to be answered? Book I (Α) determines “the science we are searching for” (983a21) as “wisdom”, i.e. a science of the first principles and causes (982a2). Book IV (Γ) begins as follows: “There is a science which investigates being as being and the attributes which belong to this in virtue of its own” (1003a21f.). Book VI (Ε) declares as its task: “We are seeking the principles and the causes of the things that are, and obviously of them qua being” (1025b3f.). According to Book XII (Λ), it is about the ousia, the essence or substance. “The subject of our inquiry is the ousia, for the principles and the causes we are seeking are those of the ousiai” (1069a18f.). I. What connects these apparently diverse questions? The key is in the sentence “The being is predicated in manifold senses, but in relation to one and one particular nature and not by a mere equivocation” (IV 2,1003a33f.). The word “being”, so Aristotle claims against Parmenides, is used in multiple senses. But this is not a mere equivocation, i.e. the usage of the same word without a common ground in the respective things (cf. Categories 1a1-6), but the usage of the same word rests on the fact that everything which is called being stands in a relation to one and the same nature. Metaphysics is a semantic investigation, it asks for the manifold use of the word “being”, and at the same time it is an investigation into reality, into that what is, since it asks for the real relations which are the basis for the manifold uses of the word. Metaphysics draws a framework

106

Friedo Ricken

which comprises the whole of reality, it distinguishes various realms of reality within that framework and asks for the relations between them. The sentence is elucidated by two examples (1003a34-b4). The word “healthy” is used in various senses, but each of theses uses implies a relation to one and the same nature: health. Everything we call healthy stands in a real relation to health – be it that it preserves or effects health, be it that it is a symptom of health, be it that it is capable of health, i.e. be healthy or become healthy. What we call medical (iatrikon) is in relation to the science or art of the physician, namely medicine. A medical man is one who has command of this art; the medicine or medicament and the medical treatment are products of this art. In the same way, “being” is predicated in manifold senses, but each of these uses implies a relation to one principle. What is thus predicated in manifold senses, but in relation to one principle, is the object of one science; in that way, for everything healthy the science of medicine is competent. “Hence it is clear that it is also the task of one science to study the being qua being” (1003b16f.). The paramount and primary object of this science is the “First” or the principle “on which the other depends and by virtue of which it gets its name” (1003b16f.). What is this paramount object of the science of the being qua being? In this passage, Aristotle only gives a hypothetical answer: “If that is the ousia, then the philosopher must have the principles and causes of the ousiai” (1003b17-19). Book VII (Ζ) once again takes up the question of the relation between the various senses of “being”. Here Aristotle starts out with the various forms of the simple predicative statement; such a statement can denote what something is, how it is, how big it is, where it is, what it does, what it sustains etc. (cf. Topics I 9; Categories 4). He refers to Δ 7, where he has distinguished the various senses of “being”. “There are several senses in which a thing may be said to ‘be’, as we pointed out previously in our book on the various senses of words; for in one sense the ‘being’ meant is ‘what a thing is’ or a ‘this’, and in another sense it means a quality or quantity or one of the other things that are predicated as these are” (1028a10-13). For Aristotle, the answer to the question which of these various senses is primary goes without saying, it is “evident” (phaneron, 1028a14). The being in the first sense is “that what it is” (to ti estin)“ (1028a14). If ‘being’ is used in this sense, it denotes the ousia. On being asked ‘What is it?’ we answer, e.g., ‘it is a man’ or ‘it is a god’, but surely not by statements like ‘it is white’ or ‘it is three yards long’. Being-white or being-three-yards-

What is Aristotelian Metaphysics?

107

long is only called being because it is a quality or a quantity of a being which is being in the mode of what it is. This intuition is confirmed by a semantical observation. Aristotle asks: Does the infinitive denote a being at all? “And so one might even raise the question whether the words ‘to walk’, ‘to be healthy’, ‘to sit’ imply that each of these things is existent [...]; for none of them is either self-subistent or capable of being separated from substance” (1028a20-24). If we want to denote a being, we must use the participle and talk about a going thing, a sitting thing or something being healthy. The reason why we do denote a being by that grammatical form is the following: its use expresses that there is a certain thing which goes, sits or is healthy; “but that is the ousia and the individual” (1028a27). Hence it is clear, so Aristotle concludes, that all the other forms of being “are” by virtue of the ousia, “so that the being in the primary sense (and not the being-something, but the simply being) seems to be the ousia” (1028a29f.). The word “the first” is used in various senses; the ousia is the first in threefold respect. (a) With respect to time: Quality, quantity etc. may change, but the ousia remains. The former can only be in the ousia and not separated from it, whereas the ousia can exist separately from its particular and changing properties. (b) Conceptually: The concept of the other modes of being cannot be defined without the concept of the ousia; qualities, quantities etc. are modes of being or determinations of the ousia. (c) With respect to our knowledge: We think that we know something best if we know what is is. We know more about a human being if we know what it is than if we just knew how tall it is, where it lives etc. “The question which was raised of old and is raised now and always”, Aristotle summarizes the result of these considerations, “and is always the subject of doubt, viz. the question what being is, means nothing other than the question what the ousia is” (1028b2-4). The word ousia is used in various senses; the four most important ones are “the essence (to ti ên einai) and the universal and the genus […] and as the fourth of those the substratum (hypokeimenon)“ (Z 3,1028b33-36). Aristotle adresses this fourth sense first. “The substratum is that of which everything else is predicated, while it is itself not predicated of anything else; hence we must first determine this” (1028b34-1029a1). By the substrate (of a statue) we mean the matter (the ore), the form (morphê; the shape of the depicted person) and the whole of matter and form, i.e. the statue. Aristotle puts up a hypothetical ranking between these three senses: “If the form (eidos) is prior to mat-

108

Friedo Ricken

ter (hylê) and more being, then it must also be prior to the compound of both, for the same reason” (1029a5-7). By that, the aim of further investigation is determined: it shall be shown that the form (eidos) or the essence (to ti ên einai) is the ousia in the proper sense. As a provisional result, Aristotle notes: “We have now outlined what the ousia is, namely that it is that which is not predicated of a substratum, but that it is that of which the other is predicated” (1029a7-9). The ousia is determined by its position in the statement. Metaphysics Z 17 wants to answer the question of the ousia from another starting point. “Let us state what, i.e. what kind of thing, substance should be said to be, taking once more another starting-point. Since the ousia is a principle and a cause, we must start out from that” (1041a6-10). That the ousia is a cause has been shown in the analysis of generation and corruption (Z 7-9). “Everything that comes to be comes by something and out of something and becomes something” (1032a13f.). Aristotle distinguishes between a generation by nature, a generation by art, and a spontaneous generation. In the case of natural generation, that by what something is generated, is the specific identical eidos in another individual of the same species, “because a man generates a man” (1032a25); the cause of generation by art is the eidos or essence (to ti ên einai) in the soul of the artisan (cf. 1032a32-b2). What the ousia is, shall be clarified by showing in what sense and in what mode the ousia is a cause. “The ‘why’ is always inquired into by asking why one thing does attach to some other” (1041a10f.). “Why is that, e.g. bricks and stones, a house?” (1041a26f.). By asking in that way, one asks for the cause. One asks “why matter is something, e.g. why is that here a house? Because what it is to be a house is present” (1041b5f.). This cause, so the general answer, is the essence, that what it is. In artifacts, e.g. in a house or a bed, the essence is the purpose; they are produced for a purpose and are here for a purpose; in organisms it is “the first mover” (1041a30), the principle of life or the soul. “Therefore we ask for the cause of the matter (but this is the eidos), by what it is something; and that is the ousia” (1041b7-9). A house is composed of bricks and stones, and the next step in the determination of the ousia is Aristotle’s distinction of two possibilities how a multiplicity of parts can make up a unity; examples are the “heap” (1041b12) and the syllable; both are a unity from the sum of its parts. The syllable is not identical with the vowels and consonants out of which it is

What is Aristotelian Metaphysics?

109

formed; ba is not the same as b and a, “nor is the flesh the same as fire and earth; because after decomposition one is not there anymore, i.e. flesh and the syllable, but the letters are there, and fire and earth. Hence the syllable is something, it is not only its elements, vowels and consonants, but also something else, and the flesh is not only fire and earth or warm and cold, but also something else“ (1041b13-19). But what is that other thing? Aristotle gives a negative answer first. It is not an element, and it does not consist of elements; hence it belongs to another ontological realm. If it were an element, we would have only augmented the number of elements by one; but the question for the cause why the elements are arranged in a certain way would not be answered by that. Were we to add another element which was supposed to be the cause of order, we would be confronted with same question again, and so on in infinitum. If we answered that the other thing consists of elements, then it would have to consist of more than one element, and hence we would be confronted with the same question as in the case of the flesh and the syllable. What positive answer can be given? The other thing exerts a causality and is, hence, active and real. It is the cause why the elements are not a heap, but an ordered whole – syllable, flesh or house. It is that by which something is what it is – syllable, flesh or house –, whereas the elements are only that from which it is. “But this is the ousia of each thing, for it is the primary cause of being” (1041b27f.). II. The question “what being is means nothing else but the question what the ousia is” (1028b2-4). The second book of De anima asks: “What is the soul?“ (II 1,412a5). Aristotle presupposes that one of the categories of being is the ousia, and in the ousia he distinguishes between the matter, the form (morphê, eidos) and the whole out of matter and form. Matter is potentiality (dynamis), form reality (energeia), “and that in a twofold sense, like science (epistêmê) and contemplation (theôrein)“ (412a10f.). On being asked what falls under the concept of ousia, one would primarily mention bodies, and among them in turn “natural” bodies, which have the origin of motion and rest in themselves. Among them we have to distinguish between those having life in them and those like earth, fire, air, water or sun, moon, stars, which do not have life (cf. Metaph.VII 2,1028b8-13). “But by life we mean nourishment, growth and decay by itself” (412a14f.). Hence,

110

Friedo Ricken

every natural body which shares life is an ousia, and morever – as this characterisation indicates – a “composite” ousia (412a16). This composite ousia is a body of a certain quality: a body bearing life. The ontological analysis of this whole starts from its linguistic description: a certain quality is predicated of a subject. From that, Aristotle concludes that the soul is not the body. The body is the subject (hypokeimenon) of which a certain quality is predicated; it is not the predicated quality. To this body it belongs that it shares life. Of this body it is true that it shares life. By that, the other ingredient of the composite ousia, the whole out of matter and form, is mentioned: It is the ability to nurture itself and to perform the processes of growth and decay out of itself, namely the soul. “Necessarily, then, the soul is ousia as the form (eidos) of a natural body which potentially has life. But the ousia is perfecting reality (entelecheia)” (412a19-21). We may elucidate this definition using one of the examples from Metaphysics Z 17. Bricks and stones can be arranged in a way that they make up a house which protects from the rigours of the weather; they are potentially a house. The order which arranges bricks and stones to a house makes them apt to serve this purpose; it connects them to an object of higher order, and in this sense it is a perfection of those materials. The natural bodies earth, water, air, fire can be connected in a way that they form an organism which performs by itself the processes of nourishment, growth and decay. In this sense they potentially have life. The cause which connects them in this way is the soul or the form (eidos). Unlike the form of the house, the arrangement of bricks and stones which is created by the builder and which is a static principle, the soul is an active, dynamic principle which performs this arrangement by itself in the processes of nourishment and growth. It is a “perfecting reality” in the sense that it brings the organism to full development from the seed or the embryo. That is why the soul is an ousia in a more proper sense than the form of a house. Its reality consists not only in an arrangement of elements that enables these elements to a performance to which they were otherwise unable – rather, the soul itself causes this arrangement which enables to higher performances. The soul is “cause” (aitia; De an.II 4,415b8) of the living body in a threefold sense: (a) It is the origin of motion (hothen hê kinêsis; 415b10), i.e. the cause of the processes of growth and decay, the changes in qualities and location of the organism. (b) It is the “for-what” or final cause (hou heneka; 415b10f.), “since all natural bodies are instruments of the soul, the

What is Aristotelian Metaphysics?

111

bodies of plants as well as those of animals; this shows that they are all means to the soul as end” (415b18-20). (c) The soul is cause of the living body in the sense of ousia, “because the ousia is for everything the cause of being, but for the living thing the life (zên) is its being; and the principle and cause of life is the soul” (415b12-14). The order of the elements in an organism, the construction of its organs, and the shape of its parts are such that they enable the whole, i.e. the organism, to perform the processes of life. But the claim that the soul is eidos and ousia, cause of the being, does not only mean that it is the form in a comprehensive sense, which gives the natural body its life (in the twofold sense: as potentiality and realisation of this potentiality) and its being. The soul, i.e. the eidos and ousia, is also the origin of motion in the sense that it constitutes the natural body to an object that has the form and the ability for the processes of life. It constitutes the natural body in a way that it is an instrument for the soul’s activities, and it preserves the body. It is a “perfecting reality” (entelecheia), i.e. a moving principle which guides the natural body to its destination (telos), the fully developed abilities in the service of the soul. It is the cause of being not only in the sense of a static form which makes an object be what it is, but also a dynamic principle which determines the process of development directed to the full unfolding of the abilities. Aristotle explains this causality with the example of growth and nourishment. He rejects the reductionist thesis that growth would be explainable by the natural movement of the elements (De an II 4,415b27-416a18). Empedocles explained the growth of plants by the natural motion of earth and fire. According to him, plants grow downwards (i.e. strike their roots deeper and deeper into the ground) “because earth naturally moves like that, and upwards, because fire moves likewise” (415b29-416a2). In contrast, Aristotle objects: (a) Up and down in an individual organism is not the same as up and down in the cosmos. If growth goes back to the upward movement of the warm within the organisms, then this movement cannot be explained by the natural movement of the elements in the cosmos. (b) The natural movement of the elements in the cosmos cannot explain the unity of the organism. “What is it that connects fire and earth, which move in opposite directions? For they will be torn apart, unless there is something to prevent it. While, if there is, then this is the soul and the cause of growth and nourishment” (416a6-9). (c) The causality of fire in nourishment and growth is beyond dispute. Nevertheless, fire can be only one

112

Friedo Ricken

cause and not the cause; rather, there needs to be a cause which limits the fire to the extent which is necessary for growth and nourishment. “Because the growth of fire reaches to infinity, as long as there is fuel; in all natural objects, however, there is a limit and a measure of size and growth. But they go back to the soul and not to fire, and rather to the form (logos) and not to matter” (416a15-18). What the organism takes is “nutriment” (trophê) and “means for growth” (auxêtikon). In so far as the organism falls under the category of quantity, it is means for growth; in so far as it is an individual and an ousia, it is nutriment, “for it preserves the ousia, and the organism exists as long as it is nourished” (416b14f.). Aristotle compares the relation between organism and nutriment with the relation between carpenter and timber. The active, formative cause is the carpenter and not the timber. The same is true of the relation between organism and nutriment. “The nutriment is acted upon by that which is nourished, but not that by the nutriment” (416a34f.). As the carpenter treats the timber, the organism digests the nutriment. Contrary to Empedocles’ assumption, the elements do not act upon the organism according to their own laws of motion; to the contrary, the effective cause is the organism which in the process of digestion directs the elements in a way that is useful to its processes of life. The nourishing soul is the organism’s power “which is able to preserve that, which has it, as such a one” (416b18f.). Although the active principle in nourishment is the organism and not the nutriment, the organism is in need of nutriment. The carpenter is not acted upon by the timber, but he can only act as a carpenter if he has timber; in that sense the timber causes that he “passes from inaction to activity” (416b2f.). Likewise, the nourishing soul is in need of nutriment. By its activity, it preserves the organism as individual ousia, but without nutriment it cannot perform this activity. That is why the organism cannot be without nutriment (cf. 416b19f.) Aristotle illustrates the relation between the causality of the soul and the causality of the elements with another metaphor: “But that by which it [the organism, F.R.] nourishes is twofold, just like that by which he [the steersman] steers, the hand and the rudder, the one causing motion and being moved, the other only causing motion” (416b25c-27, Ross’ text). The activity of the nourishing soul is the adaptation of the taken nutriment to the organism, viz. the digestion. This process is only possible through warmth; hence, every living thing has warmth. But like the rudder by the hand of

What is Aristotelian Metaphysics?

113

the steersman, so warmth must be guided by the soul; it may not follow its natural motion, but it must be guided by the laws of the organism. III. “Now the soul is cause and principle of the living body” (415b8). But ‘life’ is used in manifold senses, and we call something living even if only one of the following is present: reason, perception‚ local motion, nourishment, growth and decay (cf. 413a21-25). We have considered the causality of the soul in the processes of nourishment and growth and do now turn to local movement. The actions of the human being include bodily movements; hence the question for the causality of the soul for local movement is at the same time the question for the ontology of human action. “What is it that causes the local movement of the animal?” (De an.III 9, 432b8). Aristotle gives a negative answer first (it is neither the faculty of nourishment nor perception nor reason) and adds his positive thesis: “Now obviously these two cause motion: appetency or intelligence (nous), if one regards imagination (phantasia) as a form of thinking (noêsis). For many follow their imagination against their knowledge, and in other animals there is neither thinking nor reasoning” (433a9-12). The concepts of intelligence and appetency and the relation between them will now be determined more closely. Every appetency is directed to an end. It is cause and starting point (archê) of the practical reasoning which asks for the means and their order, by which the desired end can be realised. The final point of the practical reasoning, i.e. the insight of the first step to be done here and now, is origin (archê) of the action. Hence, if we ask for the cause of local movement, we have to distinguish: (a) the desired thing (orekton). That is the first cause, and it moves without itself being moved. It moves by being imagined or being grasped by reason. (b) the appetitive faculty. It is “moving and moved” (433b16); it is moved by the desired thing and in turn moves the animal. The desired, the first cause of local movement, is characterised in two respects. It is „either the good or the apparently good (phainomenon agathon); but not any good, but only the good realisable in action (prakton agathon). Realisable in action is only that which can also be otherwise” (433a28-30). Only possible states of affairs can be objects of appetition (nobody can strive for impossible or necessary things), and among them

114

Friedo Ricken

only those which the agent can realise in its action (cf. Nicomachean Ethics III 5, 1112a18-31). Only that can be strived for which appears as good, i.e. which someone considers good. What is considered good can be good or not, i.e. it can be a true or apparent good (cf. Nic.Eth. III 6); only if it is a true good, then the striving is right (cf. Nic.Eth. VI 2, 1139a24). Action takes place in the realm of contingent being, which can be so or otherwise (433a30; Nic. Eth. VI 2, 1139a7f.); here the human being acts as one of the causes. Aristotle distinguishes a twofold causality. The unmoved moving cause is the end or what is strived for, it is determined by the judgement; what is strived for is something that someone considers good. The moved and moving origin of the action, “origin of movement and not end” (1139a31f.), is the decision, and origin of the decision is “the striving and the reasoning that is directed to an end” (1139a32f.). “Hence the decision is either striving reason or thinking striving, and such an origin is the human being“ (1139b4f.). IV. The question “what being is means nothing else than the question what the ousia is“ (1028b2-4). We saw that the ousia in the first and proper sense is the form of a whole. The beginning of Metaphysics Ζ 17 (where Aristotle proves this) however indicates that this is not the final and ultimate answer to the question for the ousia and the being. “As what and how qualified one should describe the ousia, we want to say once again, this time proceeding from another starting point; for perhaps from this we shall get a clear view also of that substance which exists separated from sensible substances” (1041a6-9). What is this “separated” ousia? Why does the question for the being as being lead to that ousia? How can it be shown that a separated ousia exists? Metaphysics Ε 1 starts out from the question for the “principles and causes of the being things, and obviously of them qua being” (1025b3f.). Aristotle distinguishes three disciplines of theoretical philosophy: mathematics, physics and theology (theologikê; 1026a19), and he evaluates the ontological rank of their domains of objects according to the criteria of subsistence, whether they exist “separately” (chôriston), and immovability (akinêton). The objects of physics, the organisms and the celestial bodies, stand alone in themselves, but they are subject to change, at least local motion. Whether the objects of mathematics are immovable and exist sepa-

What is Aristotelian Metaphysics?

115

rately, is “unclear”; it is “clear”, however, that mathematics considers some of them qua immovable and qua separate” (1026a9f.). The “first” philosophy deals with the “separate and immovable” (1026a16); its rank results from the rank of its domain of objects. But does not the first philosophy, as claimed at the beginning of this chapter, deal with being qua being? What, hence, is the object of the first philosophy: the universal, the being qua being, or a certain genus of beings, the separate and immovable ousia? Aristotle answers: If there is no other ousia beyond the changeable ousia, then physics would be the first philosophy. But if there is an immovable ousia, then the science which has it as its domain enjoys higher rank, and hence it is the first philosophy. This science of the first being is as such universal, i.e. science of the being qua being, “and universal in this way, because it is first” (1026a30f.). The word ‘being’ is used in multiple senses because everything that is called being stands in a relation to a First one; the distinct and primary object of this science is the First or the principle „on which the other things depend and in virtue of which they get their name” (1003b16f.). How is the question of whether physics or theology are the first philosophy to be settled? Metaphysics Ε 1 sketches an argument. The first philosophy deals with the „separate and the immovable. Now, of necessity all causes must be eternal, especially those ones, because they are causes for the visible of the divine things” (1026a16-18). The existence of the separate and immovable ousia is proved from the eternal motion of the stars; theology rests on premises from the Aristotelian world-view. These relations are elaborated in book Λ (XII). The book deals with the ousia. “The consideration is on the ousia, since the principles and causes of the ousiai are being sought” (Λ 1,1069a18f.). Aristotle distinguishes three ousiai, two of them perceptible and subject to change, of which one (plants and animals) is perishable, the other one (celestial bodies) eternal and only subject to local motion, and an immovable eternal ousia (cf. 1069a30-33). He wants to show that “there must be of necessity an eternal immovable ousia” (Λ 6,1071b4f.). We must constrain ourselves to point out the premises which this proof presupposes. Metaphysics as philosophical theology, as developed in book Λ , rests on the thesis of the eternity of the cosmos and the motion and on the Aristotelian theory of time. “For the ousiai are the first of the beings, and if they are all perishable, then everything is perishable. But it is impossible that motion either emerges or perishes (for it was

116

Friedo Ricken

always there), and likewise time. For it is not possible that there is a before and after if there is no time” (1071b5-9). An argument that does not presuppose these premises can be found in Θ (IX) 8. The thesis of this chapter is “actuality (energeia) is prior to potency (dynamis)” (1049b5). Aristotle distinguishes three senses of “prior”: namely, prior with respect to the concept (logos), to time (chronos) and to being (ousia). The crucial argument why actuality is prior according to being than potentiality, goes as follows: “The eternal is prior according to being than the perishable; but nothing that exists potentially (dynamei), is eternal” (1050b6-8). Every potency, Aristotle argues, is at the same time a potency for the contradictory opposite. “Hence; what is capable (dynaton) of being, may either be or not be; hence the same thing is capable both of being and of not being. But what is capable of not being may possibly not be. But what may possibly not be, is perishable, either in the full sense or in the precise sense in which it is said it possibly may not be, i.e. in respect either of place or quantity or quality; in the full sense perishable is what may not be in respect of the ousia. Hence, nothing of that which is in the full sense imperishable is in the full sense potentially existent [...]; all imperishable things, then, exist actually. Nor can anything which is of necessity exist potentially; yet these things are the First; for if these did not exist, nothing would exist” (1050b11-19). So, is physics or theology the first philosophy? “Nature is in the same genus as potency; for it is a principle of movement, not however in something else but in the thing itself qua itself” (1049b8-10). The organisms are “in potency”, i.e. they can be or not be. What can be or not be, is neither eternal nor necessary. Why is that what can be and not be? It can only be because there is something that cannot not be. This necessarily being is the First that is sought by the science of the being qua being. V. What is Aristotelian metaphysics? The “first philosophy” asks for the unity of the various meanings of ‘being’ and the relation between the domains of objects of the various sciences, which all talk about that what is. The semantic question is a question for a ranking in being. What is in the first and proper sense? Which dependencies exist between the various realms of beings? The being in the primary sense is the ousia, the individual standing in itself, at which we can point and of which we can ask “what is it?” It

What is Aristotelian Metaphysics?

117

must be distinguished between man-made artifacts and organisms, the latter standing in the full sense in themselves because they preserve themselves. The ousia is the “first part” of the “whole (holon)” which is “the universe (to pan)” ist (1069a19f.); it is the cause of the processes in the universe. Aristotle develops the ontological concepts and distinctions which we need to understand the phenomenon of the organism; in that way he achieves a differentiated concept of causality. To what activities an organism is able and how it performs them depends on its form; local movement of animals and humans, e.g., is performed in different ways. The ontological analysis of the perceptual, changable ousia brings two characteristics of Aristiotelian metaphysics to light: anti-reductionism and anti-determinism. (a) The ontological concepts apt for the description of parts are not sufficient to grasp the whole. The whole belongs to a higher order which cannot be explained by the parts. The activities and the causality exerted by the whole cannot be reduced to those exerted by the parts. The laws governing the activities of the parts are modified in the sense that the motion of the parts is now steered by the whole. The parts may make up the whole, but in order to really make up the whole, a cause belonging to another ontological order is required. (b) The processes in the cosmos do not constitute a single, closed causal nexus. The First in the cosmos is the ousia; every ousia is an origin of movement, the activity of each ousia is not determined by its elements, but by its specific form. This is also true of man; he is “origin” of his actions as “striving reason or thinking striving”. The Aristotelian ontology is a framework in which the human being can understand itself as an agent. The Aristotelian ontology of the perceptible, changable ousia can be regarded as a sort of snapshot of the scala naturae. The dimension of time is missing; Aristotle presupposes the constancy of the biological species and does not think of evolution. He asks: which ontological concepts are required to understand the present phenomena? Does this mean that his ontology is made obsolete by evolutionary biology? – Every explanation presupposes that one knows the explanandum. Aristotle develops the ontological concepts necessary to grasp the present phenomenon of the scala naturae in a differentiated way. By that he formulates the task which an evolutionary explanation has to fulfill, and he delivers criteria to judge whether an evolutionary explanation has come up to this task. Aristotelian metaphysics as theologia naturalis poses the question: How can it be explained that that which can and can not be really is? The ques-

118

Friedo Ricken

tion for the principles and causes of the being qua being is not yet answered by the insight that the natural ousia (which can and can not be) is primary. If the necessarily being were not to exist – that is the crucial step now –, then nothing would exist. Thomas Aquinas takes up this argument in his “Third Way”, ex possibili et necessario (S.th. I q.2a.3). It does neither rest on particular cosmological premises nor does it require epistemological presuppositions; it just consists in modal-ontological reasoning. We notice that something is the case which could also not be the case, and we ask why it is the case and not not the case. The being in the first and proper sense, the principle of the being qua being, can only be the being which excludes the possibility of non-being in any respect, and that is the unmoved mover; it cannot be otherwise in any respect, because it is pure actuality and hence necessary (cf. Λ 7,1072b4-13).* * Translated by Winfried Löffler

2. Ontology at Work

Experience of Change and Change of Experience E.J. LOWE, Durham

Edmund Runggaldier has long included amongst his philosophical interests problems concerning time, persistence, and the nature of human persons and their experience. With this in mind, it seemed to me appropriate to offer something dealing with all of these topics as my tribute to his wonderful qualities as a philosopher, educator, and friend – even though he may well disagree with many of the things that I am about to say. Do we ever, strictly speaking, experience change – and if so, how is that possible? To answer this question, we need to inquire into what change is and how change is related to the passage of time. We also need to inquire into the nature of experience, especially sense-experience. In this paper, I shall be defending, inter alia, the following claims, which I have numbered for ease of reference. Where relevant, I shall cite earlier work of mine, most of it fairly recent, expanding on some of these claims. But other claims are new to this paper. At various points I shall raise supplementary questions which generate several of the claims that I shall be making. I shall mark these questions ‘SQ’, followed by a number. (1) The passage of time consists in the coming into existence and passing out of existence of entities of various sorts. All real change is ultimately existence-change. What are sometimes thought of as being different species of change, such as qualitative change, relational change, and substantial change are really just existence-changes involving different kinds of entity – qualities, relations and substances, respectively. However, as we shall see, this view requires us to think of qualities and relations as particulars, not just as universals. (See further Lowe 2006a and Lowe 2009.) (2) The concept of existence is not, contrary to widespread dogma, expressed by the misnamed ‘existential quantifier’, ‘’, which ought instead to be called the particular quantifier and has no existential import. ‘Exists’ is a first-level, not a second-level, predicate. But existence is not a property of entities, if by a ‘property’ is meant an entity either in the category of universals or in the category of so-called tropes or (as I prefer to call them)

122

E.J. Lowe

modes. Indeed, existence isn’t any kind of entity, any more than identity is. The concept of existence is primitive and indefinable, like the concept of identity. And like the concept of identity, it is a formal ontological concept. But this doesn’t mean that existence-claims lack objective truth and falsehood: quite the contrary. (See further Lowe 2003, Lowe 2006a, and Lowe 2009.) SQ1. So, what is wrong with the standard, quantificational interpretation of ‘exists’, according to which this is a second-level predicate? Answer: (3) Many things are wrong with the standard intepretation, but one way to highlight its inadequacy is this. According to this view, an existence statement, such as ‘Tigers exist’, effectively means something like ‘Tigerhood has at least one instance’, which predicates the second-level property of having at least one instance of the first-level property being a tiger. But what is it to say that a first-level property ‘has at least one instance’, if not precisely to say that such an instance exists? It had better mean that, for the proposal to work: but if it does, then the analysis already presupposes the notion of existence and doesn’t in any sense explain it. And that the particular quantifier (as it should rightly be called) does not necessarily carry existential import is easy to show, by way of examples. For instance, the Red Queen in Lewis Carroll’s Through the Looking-Glass claims to think of six impossible things before breakfast. Plainly, these are not six existent things. ‘At least one x’, like ‘Exactly six xs’, expresses quantity, but not existence. (See further Lowe 2003, Lowe 2006a and Lowe 2009.) (4) When something ceases to exist, it ceases to exist absolutely: it ceases to be a part of reality as a whole. The idea that something which existed in the past, but which no longer exists, merely exists at a nonpresent temporal location is entirely misguided, along with the attendant notion that time is a dimension in which reality is extended, akin to the three dimensions of space. A consequence of this is that there are no ‘crosstemporal relations’, conceived as external relations, in the way that there plausibly are external spatial relations, such as distance relations. This is because external relations can obtain only between co-existent relata, and hence not between items that do exist and ones that have ceased absolutely to exist. I might be taller than Julius Caesar, because being taller than is an internal relation, but I am not at some temporal ‘distance’ from him. (See further Lowe 1998, 101–105, Lowe 2006a, Lowe 2006b and Lowe 2009.)

Experience of Change and Change of Experience

123

SQ2. So, if all change is existence-change, can we experience something’s coming into or going out of existence, and if so, how? For example, we have implied that a qualitative change, such as a leaf’s turning from green to brown as it burns in a flame, consists in the going out of existence of one entity, the leaf’s green trope or mode, and the coming into existence of another, its brown trope or mode. But what, if anything, is it to experience such a change? Preliminary answer: (5) We don’t need to speak here of an event – the leaf’s changing colour from green to brown – and regard that as an object of our experience. Indeed, we shouldn’t reify events at all. Strictly speaking, no such entity as an ‘event’ exists. That helps to explain why we feel uncomfortable about saying that events ‘exist’, and say instead only that they ‘occur’. But to say that an event has occurred is just to say that one or more things has come into or gone out of existence. The things in question are not, of course, events – rather, they are bona fide entities, such as modes or substances. For instance, a death occurs whenever a living substance goes out of existence, which it does when certain of its essential qualities, or certain of the essential relations between its parts, cease to exist. SQ3. So how can we experience change if not by experiencing events – there being, strictly speaking, no such entities? At this point we need to invoke a theory of perception: (6) In my view, any adequate theory of perception must be a causal theory. I reject the objections to such a theory that are commonly raised by, for instance, adherents of so-called disjunctive theories of perception. A perceiving subject, S, can perceive an entity, e, only if S has some perceptual experience, p, elements of which are suitably causally related to e. To cut a long story short, the ‘elements’ in question will be sensory or qualitative elements of the experience: in a word, sense qualia. (See further Lowe 2008b and Lowe 2008c.) (7) A sense quale is a species of mental trope or mode, examples being a particular sensed taste, odour or colour, such as the tang of a lemon, or the acridity of some wood-smoke or the redness of a cherry. The tang of a lemon is not, of course, literally a mode or trope of the lemon: it is not one of the lemon’s particular qualities. Rather, it is a mode or trope of the perceiving subject: it is a mental mode or trope. Thus, when S perceives the lemon’s sensible qualities, one of those qualities – its acidity – causes a mental trope of S’s to come into existence: a sense quale describable as a

124

E.J. Lowe

‘tang’. S perceives the lemon’s acidity by experiencing the tang that it causes. And it is most important to note here that the sensible, or perceptible, qualities of a thing, such as a lemon’s acidity or its yellowness, must be regarded as particular qualities of that thing – modes or tropes – not universals. For only particular qualities can stand in causal relations to perceiving subjects, in the manner required for sensible qualities to be perceptible, in accordance with a causal theory of perception. (See further Lowe 2008a.) (8) S does not perceive the tang, of course, in the sense that S perceives the lemon’s acidity: rather, S has the tang, since it is a mental mode or trope of S. S ‘experiences’ the tang only in the sense that the tang is one element in the combination of mental modes or tropes that makes up the qualitative content of S’s perceptual experience at the time. Perceptual ‘experiences’ have as their qualitative content combinations of co-existing sense qualia, of which a particular tang may be one element amongst many. Experiences may also have another kind of content – intentional or representational content – but that is not my concern here. (See further Lowe 2008c.) SQ4. So, what happens when, say, the lemon loses its acidity, because an alkaline substance is added to it, or when its acidity is masked, because sugar is added to it? Answer: (9) Then, of course, that acidity no longer causes S to experience a tang. The tang that it formerly caused ceases to exist, and is replaced by a different sense quale, this being caused instead by the lemon’s newly acquired neutrality, or its newly acquired sweetness. So, existence-changes in the qualities of a perceived object may result in existence-changes in the perceiving subject’s sense qualia. I say ‘result in’, not ‘cause’, since what is strictly and literally caused is, first, the existence of the tang, which then ceases to exist when its sustaining cause, the acidity of the lemon, ceases to exist, and second, the existence of a successor quale by the lemon’s newly acquired neutrality or sweetness. Thus, existence-changes in the qualities of perceived objects typically give rise to corresponding existence-changes in the sense qualia of perceiving subjects. (10) And this, ultimately, is all that the experience of qualitative change can amount to: it amounts to the qualitative change of experience. In other words, we may be said to experience things undergoing changes of their qualities precisely and only to the extent that such changes result in our

Experience of Change and Change of Experience

125

own experiences changing, in respect of their qualitative constituents, that is, in virtue of the coming into or going out of existence of certain of our sense qualia. (Note, incidentally – although I don’t have space to expand upon this here – that causation need not and should not be thought of as involving external cross-temporal relations: for effects are simultaneous with their immediate causes and the rest is just a matter of persistence and thus of identity, an internal relation.) (11) As was remarked at the outset, the passage of time, quite generally, consists in the coming into and going out of existence of various entities. Without such existence-changes, there would simply be no passage of time. There cannot be time without real change. Because time passes uninterruptedly – or so we assume – existence-change must be continual, without ‘gaps’. There can be no period of time during which everything continues to exist. Equally, though, there can be no period of time during which nothing continues to exist. The passage and unity of time depend upon both change and persistence, with every change having some entities persisting through it and all persistence being accompanied by some change. (See further Lowe 2006a and Lowe 2009.) (12) But this view requires us to include modes or tropes in our ontology, not just properties and relations conceived as universals. For when an object (or ‘substance’) undergoes a change with respect to the universals that it exemplifies, neither the object nor the universals in question undergo an existence-change. Such a change, then, would be a real change in the absence of any existence-change, if we had to describe it purely in terms of an object first exemplifying one universal and then exemplifying another. But once we appreciate that an object can do this only by first possessing a mode of the first universal and then possessing a mode of the second, with these modes successively coming into and going out of existence, then we see that this kind of change does after all involve existencechange. So here is a powerful additional reason for including modes or tropes in one’s ontology: they enable one to give a simple and unitary account of change and the passage of time. (See further Lowe 2006c.) (13) We can distinguish between ‘objective’ and ‘subjective’ temporal passage in the following way. Objective temporal passage simply consists in universal, ubiquitous existence-change: the continual coming into and going out of existence of entities of various sorts, quite independently of any perceiving subject. Subjective temporal passage consists in the existence-changes involved in a particular subject’s experiences, such as the

126

E.J. Lowe

successive coming into and going out of existence of my sense qualia. Objective time cannot ‘stand still’, but perhaps subjective time can: it will do so for me if none of the constituents of my experience, such as my sense qualia, undergoes existence-change during some period of objective time. For this reason, too, subjective time may seem to pass ‘more quickly’ or ‘more slowly’ than objective time. SQ5. Does this view of time conflict, to its detriment, with the special and general theories of relativity? Probably not, for reasons that I am about to give. (14) First, it is not uncontroversial to assume that the theory of relativity is genuinely a theory about time. (This is not settled by the fact that its mathematical equations include a parameter, t, whose values are empirically determined by clock-readings.) If the theory commits us to regarding time as a dimension in which reality is extended, or at least as somehow contributing to a four-dimensional spacetime manifold, then it may be better to say that it really amounts to a denial of the reality of time. Now, some philosophers say that time is, indeed, some kind of illusion – although I am not at all persuaded that it is, nor do I really see how it could be. It is no use saying, for instance, that time merely appears to pass, because even if it does as much as that, then at least subjective passage occurs: but this consists in the coming into existence and going out of existence of experiential elements, such as sense qualia – and if at least these entities undergo existence-change, then objective passage is also real. The appearance of passage can only consist in the passage of appearance, and so entails the reality of objective passage and hence the reality of time. (15) Second, my theory of temporal passage implies that the sum total of reality changes as time passes. By contrast, the theory of relativity assumes (or is standardly taken to assume) that reality is eternally one and the same: so it effectively denies the reality of change, treating it as if it were nothing more than mere variation along a dimension. (On this view, a growing apple no more literally changes in size than a river changes in breadth as it nears the sea.) The objection that the theory of relativity shows that simultaneity, and thus pastness and futurity, are ‘frame-relative’ is ineffectual here, if it is meant to demonstrate that we are committed to some bizarre and unacceptable notion of ‘relative existence’, whereby what exists for one observer doesn’t for another. Our theory is, in a perfectly good sense, committed to a ‘relative’ conception of existence, but not a bizarre one, nor

Experience of Change and Change of Experience

127

a purely subjective or mind-dependent one. On our theory, Julius Caesar, who no longer exists, is no longer part of the sum total of reality: what existed ‘for him’ – that is, co-existed with him – does not exist ‘for us’. But this does not make our theory a version of ‘relativism’ in any objectionable sense, comparable with the crazy view that the earth is flat for some people but round for others, owing to their different cultures and beliefs. SQ6. But, returning to the issue of the experience of change: if the experience of change consists ultimately in the change of experience, does this mean that we cannot really experience motion, say: is it that we can only experience a succession of different ‘location tropes’ of an object, somewhat in the way in which motion seems to be experienced in the cinema, even though nothing literally moves on the cinema screen? Answer: (16) No, we are not committed to this unattractive view, if we understand properly the nature of motion. Objects may possess motion tropes or modes at an instant. The difference between a moving object and an otherwise qualitatively indistinguishable stationary object at a certain instant consists in the fact that the former object possesses, while the latter does not, a motion mode or trope at that instant. Such a mode or trope is to be thought of as being a particular directed quality or, more accurately, quantity: in effect, a ‘vector’ property of the object. The object’s continuing possession of this property causally explains its successive possession of different ‘location tropes’: but the object’s motion does not consist in its successive possession of those tropes. A perceiving subject may perceive an object’s motion tropes, since these may cause appropriate sense qualia in the subject’s experience (see below). Moreover, when an object is moving at a high velocity, the subject may simultaneously experience a number of different qualia relating to several of the object’s successive ‘location tropes’, caused by its continuing motion. This is why fast-moving objects can appear to have a ‘smeared out’ location as they rapidly cross our visual field. However, the ‘smeared-out-ness’ thus experienced is only a symptom of the object’s high velocity, not an experience or perception of it. (See further Lowe 2002, 242–4, 300–304.) SQ7. But are there really distinctive sense qualia – motion qualia, as we might call them – that are caused by the motion tropes of objects and thereby render the latter sensible qualities that subjects can perceive? Or

128

E.J. Lowe

are the only sense qualia associated with movement location qualia, caused by a moving object’s successive location tropes? Answer: (17) Plausibly, there are distinctive motion qualia, not just successive location qualia. Evidence for this is provided by the well-known waterfall illusion. If one stares at a waterfall for a while and then looks away at a static scene, one seems to see upward movement, even though no object appears to be changing its location. (The illusion is rather like an afterimage: just as the colour of an afterimage is the complementary colour of the colour first seen, so here the downward movement first seen is replaced by an illusory upward movement.) Movement and location appear to be processed separately by the human visual system, as is shown by the fact that brain-damaged subjects who cannot locate objects visually can sometimes still detect movement visually. Indeed, even normal subjects exhibit an ability to detect movement visually very easily in circumstances in which location is hard or impossible to detect visually – for instance, in very poor light, or in the periphery of their visual field. SQ8. Would we experience change and temporal passage if we had no sense-memory? (18) Perhaps not – or, at any rate, we might not be in any manner aware of time passing, or believe that it passes, if we possessed no sense-memory (by which I mean experiential memory of our own past sense qualia). But that is no reason at all to think that we ‘really’ only ever experience the world as it is at an instant and that all the rest is just a matter of what we remember of our immediately past experiences. The so-called ‘specious present’ in fact embraces sense qualia whose (distal, or non-immediate) external causes are often non-coexistent – that is, objectively nonsimultaneous – modes or tropes of external objects. In that sense, we can and often do experience at one subjective time how the world external to us is at a succession of different objective times. The ‘smeared-out-ness’ phenomenon described a moment ago is an example of this. That is to say, some of the qualia included in a single perceptual experience have as their (distal) causes modes or tropes some of which have already ceased to exist before others of them have come into existence. This fact is forcefully brought to our attention when, for example, we simultaneously see and hear a distant person strike a hammer on a stone, when the hammer seems already to have finished moving even as we hear it strike.

Experience of Change and Change of Experience

129

(19) Note that wherever, above, I have spoken about ‘time’ and ‘times’, I should be understood to have done so entirely without reificatory intent. Time is not an entity, to be included in our ontology, nor are particular times. All talk about time and times has as its ultimate ontological ground the existence-change of bona fide entities, such as substances and modes. As mentioned earlier, there are not even any external cross-temporal relations, so that even a ‘relational’ theory of time, along with a ‘substantival’ one, is ruled out. But, most emphatically, this does not amount to an endorsement of any sort of idealist theory of time, least of all a transcendental idealist one like Kant’s. For existence-change is mind-independent and real. (See further Lowe 2009.) SQ9. Should the foregoing theory be regarded as a ‘tensed’ theory, or ‘Atheory’ or ‘dynamic’ theory of time? Is it a variety of presentism? (20) All of these ways of characterizing the theory would be misleading. Tense is really a feature of language, rather than of the world. Calling a theory of time an ‘A-theory’ endorses a distinction of McTaggart’s (the Aseries/B-series distinction) that is certainly questionable, for all sorts of reasons, not least because it takes a reificatory view of events and also because it presumes the reality of external cross-temporal relations. Describing a theory of time as ‘dynamic’, as opposed to ‘static’, is to invoke a metaphor without explaining its underlying ontological basis. Presentism, understood as the view that only the present time, or only the present ‘moment’ of time, is real and that past and future times are mere abstracta, such as maximal consistent sets of present-tensed propositions, takes a reificatory view of times and is mistaken in its presumption that real crosstemporal relations must somehow be accommodated by an adequate theory of time. The theory advanced here is what it is, and not another thing (to echo Bishop Butler). Its key idea is that all talk of time has, as its ontological underpinnings, the reality of absolute existence-change: the absolute coming into and going out of existence of entities of various kinds. In the absence of such change, there would be no temporal duration, no temporal passage, and no temporal experience. Temporal experience, however, requires not only the reality of mind-independent existence-change, it requires also existence-change in the minds of experiencing subjects – in

130

E.J. Lowe

their experiences. There can be no experience of change without change of experience.1 (See further Lowe 2006a and Lowe 2009.) References Lowe, E.J. 1998 The Possibility of Metaphysics: Substance, Identity, and Time, Oxford: Clarendon Press. Lowe, E.J. 2002 A Survey of Metaphysics, Oxford: Oxford University Press. Lowe, E.J. 2003 “Identity, Individuality and Unity”, Philosophy 78, 321-36. Lowe, E.J. 2006a “How Real is Substantial Change?”, The Monist 89, 275-93. Lowe, E.J. 2006b “Endurantism versus Perdurantism and the Nature of Time”, Rivista di Filosofia Neoscolastica 4, 713-27. Lowe, E.J. 2006c The Four-Category Ontology: A Metaphysical Foundation for Natural Science, Oxford: Clarendon Press. Lowe, E.J. 2008a “Tropes and Perception”, in S. Gozzano and F. Orilia (eds.), Tropes, Universals and the Philosophy of Mind, Frankfurt: Ontos Verlag, 175-92. Lowe, E.J. 2008b “Against Disjunctivism”, in A. Haddock and F. Macpherson (eds.), Disjunctivism: Perception, Action, Knowledge, Oxford: Clarendon Press, 95111. Lowe, E.J. 2008c “Illusions and Hallucinations as Evidence for Sense Data”, in E. Wright (ed.), The Case for Qualia, Cambridge, MA: MIT Press, 59-72. Lowe, E.J. 2009 “Serious Endurantism and the Strong Unity of Human Persons”, in L. Honnefelder, E. Runggaldier and B. Schick (eds.), Unity and Time in Metaphysics, Berlin: Walter de Gruyter, 67-82.

1

I am grateful for comments received when I presented an earlier version of this paper in a Workshop on Time and Experience held at the University of Geneva in December 2008.

On the Unifying Role of Laws of Nature PAUL WEINGARTNER, Salzburg

“Perfections which in creatures are received, divided and multiplied pre-exist in God unitedly and simply.”1

I may paraphrase this quotation and apply it to laws of nature: Properties which in things of this world are divided and multiplied are unified in laws of nature.

The paper is divided into the following sections: 1) Introduction 2) Symmetric Laws Unifying Asymmetric Phenomena 3) Unification by Causal Relations represented by Laws of Nature 4) Unifying Inertial Systems (SR) 5) Unifying any Moved Systems (GR) 6) Unifications of Forces. Unified Theory 1. Introduction 1.1 Different meanings of the expression “laws of nature” L1 The “law” as it “is” in the thought of the inventor or creator or discoverer L2 The “law” as it “is” in the things which are ordered by it (i.e. as a structure in things) L3 The “law” as a law statement formulated in some scientific language L4 The “law” as an ideal true law w.r.t. which the laws known at present in the sense of L3 are approximations. 1.2 Differences between L2 and L3 -

1

Law statements are true or false. Structures of objects are not.

Thomas Aquinas (STh) I,13,4.

132

-

Paul Weingartner

Law statements can be tested, confirmed, refuted, revised… etc. Neither of those holds for structures. Law statements can be nearer to the truth than others. Structures cannot. Law statements can contain sentence-negations and thus can express negative facts (“there is no perpetuum mobile”). Structures of objects cannot contain sentence-negations.

1.3 Genuine Laws G1 Genuine laws (GL) describe or order a large domain of objects together with their properties and relations. G2 GL describe the conserved and invariant properties of the objects, but they abstract from the objects as individuals. G3 GL are invariant with respect to certain changes of parameters. Specifically they are Space-Time invariant. G4 GL hold either in all cases of application (i.e. without exception) or in most cases. G5 GL are good approximations to the true law, i.e. they have a high degree of confirmation and of informative content. G6 GL belong to a system of laws which makes up the central part of a theory. G7 GL refer (ultimately) to objective reality.2 2. Symmetric Laws – Unifying Asymmetric Phenomena Every law of nature expresses a unifying symmetry (invariance). “It is not necessary to look deeper into the situation to realise that laws of nature could not exist without principles of invariance.”3

This was already understood by the Greek Ideal of Science which was implicit in the doctrine of the Presocratic Philosophers, but explicit in Plato and Aristotle: To describe and explain the objects of the world which are visible, observable, concrete, particular, changing, material and contingent by principles (laws) which are non-visible, unobservable, abstract, univer2

These conditions are elaborated in detail in Mittelstaedt/Weingartner (2005, LNt), ch. 2. 3 Wigner (1967, SRf), p. 29.

On the Unifying Role of Laws of Nature

133

sal, non-changing, immaterial and necessary. Hence every law is something which does not change relative to something which changes and the things which change are unified by the law. We can distinguish four groups of unifying symmetries (invariances): Continuous Space-Time Symmetries, Permutational Symmetry, Discrete Symmetries, Gauge Symmetries. 2.1 Continuous Space-Time Symmetries “The paradigm for symmetries of nature is of course the group of symmetries of space and time.”4 These symmetries tell us that the laws of nature do not change while time passes and they are not different at different places in space. More accurately, there are several continuous space-time symmetries: (a) Translation-symmetry in space. This leads to three conservation principles of momentum. Unobservable: absolute place. (b) Rotation-symmetry in space. This leads to three conservation principles of angular momentum. Unobservable: absolute direction. ‘Rotation’ means here turning of the physical system on an angel. It does not mean that the system is rotating. (c) Translation-symmetry of time (i.e. delay in time makes no difference). This leads to the principle of conservation of energy. Unobservable: absolute (point of) time. (d) Velocity-symmetry: Invariance with respect to transformations between systems of inertia, thereby assuming the existence of a universal time. Unobservable: absolute velocity. (a) – (d) plus the explicit assumption of Euclidean Space provide the full Galilean invariance (symmetry) which includes relativity with respect to inertial frames (d). This symmetry group (a) – (d) underlies Newton's Theory. Observe that condition (d) was already discovered by Gelileo although fully understood only by Newton. This is the important principle of relativity entering the Theory of Special Relativity when the existence of a universal time is dropped. There are two further symmetries, one according to the principle of Special Relativity, the other according to the principle of General Relativity (see below ch. 4 and 5). 4

Weinberg (1987, TFL), p. 73.

134

Paul Weingartner

2.2 Permutational Symmetry Permutational Symmetry means that “different individual” particles of the “same kind” are treated identical. Thus the laws and the respective physical reality described by these laws remain the same if we interchange any two electrons. The same holds for protons, neutrons, neutrinos and µ-mesons (according to the Fermi-Dirac statistics) and also for photons, π-mesons, Kmesons and gravitons (according to the Bose-Einstein statistics). By this symmetry the laws of nature unify all those elementary particles which belong to the same species of particles, i.e. all electrons into one group, all protons into another etc. 2.3 Discrete Symmetries (a) Charge-conjugation-symmetry or particle-antiparticle-symmetry. Unobservable: absolute sign of electric charge. This symmetry is slightly violated in decay phenomena. (b) Parity or right-left-symmetry or mirror-image-symmetry. Conservation: Parity. Unobservable: absolute right or left, viz. absolute orientation in space. This symmetry is satisfied for electro-magnetic effects, but not completely fulfilled for radioactive decay phenomena. (c) Time-reversal-symmetry or past-future-symmetry. Unobservable: Absolute direction of time. No conservation principle follows. All fundamental laws of physics in the domains of Classical Mechanics, of Quantum Mechanics and of the Theory of Relativity seemed (so far) to be invariant with respect to time reversal. According to Prigogine this is a sign that the laws of physics are still incomplete since many processes are irreversible in time.5 But time-reversal-symmetry seems not to hold on the micro level.6 (d) CPT-symmetry. This symmetry seems not to be violated by any processes known so far. It is a threefold symmetry combining particle-

5

Cf. Prigogine (1993, TDC) and Prigogine/Stengers (1993, PZt). There is indirect evidence via CP-violation and CPT-symmetry. There is direct evidence shown by new experiments. Cf. Schwarzschild (1999, TEO) and Hagiwara et al. (2002, PDG), Vaccaro (2011, TVU). 6

On the Unifying Role of Laws of Nature

135

antiparticle, right-left and past-future-symmetry (where ‘C’ stands for charge, ‘P’ for parity and ‘T’ for time). 2.4 Gauge Symmetries There are different Gauge symmetries. A consequence of them is the conservation of electric charge, of baryon number and of lepton number. With respect to all these four groups of symmetries a law of nature can describe, explain and predict non-symmetrical phenomena. The mathematical reason for that is that a differential equation can have non-symmetrical solutions. Every initial state of the universe or of a planetary system (i.e. every concrete constellation of the planets or stars) is an asymmetry or a symmetry breaking relative to the symmetry expressed by the law. The same holds for a concrete microconstellation (microstate) of the molecules of a gas (think of a snapshot). Such non-symmetrical phenomena do not imply non-symmetrical laws, but can be caused by non-symmetrical initial conditions. In general every symmetric law connects asymmetric phenomena (concrete states) and in this sense unifies them. However, it should be noted that invariance and symmetry must not be exaggerated. No law is invariant or symmetric w.r.t. every parameter or magnitude. Especially, and very important, laws of nature are not symmetric or invariant w.r.t. constants of nature. Therefore a change of a constant of nature would change that law which contains this constant. The most basic laws of Quantum Mechanics and the Theory of Relativity contain – and are therefore dependent on – the following constants of nature: c (velocity of light in vacuum), h (Planck's Constant), e (elementary charge), em (electron mass), G (gravitational constant).7 From this consideration the question arises: Where is the border between symmetry and symmetry breaking? This border is in fact relative to some extent and depending on the following aspects: (a) on the kinds of phenomena. For example parity is satisfied in all electromagnetic phenomena, but slightly violated in radiation phenomena; further: Lorentz’s Invariance is satisfied by Maxwell's Equations, but not by Newton’s Theory. 7

For details see Mittelstaedt/Weingartner (2005, LNt), ch. 8.2.

136

Paul Weingartner

(b) on the level (macro or micro). For example a litre of gas is highly symmetric at room temperature (machrostate), but a snapshot which shows the microstate (at that time) is a symmetry breaking. (c) on the level of universality: Symmetries might be hidden. Examples: the heart on the left side or the right hand screwed houses of snails might be a symmetry breaking which might have been built up through evolution, but which is based on a symmetric law which allows both sides or both screws. 3. Unification by Causal Relations Represented by Laws of Nature All laws of nature represent some causal relation provided that the causal relation is different for dynamical and statistical laws (3.1) and the causal relations satisfy both, the temporal order and the chronology condition (3.2). 3.1 The Causal Relation is Different for Dynamical and Statistical Laws The paradigm cases for dynamical laws are the laws of Classical Mechanics (or of Newton’s Theory). This type of law is described by Laplace in his Essay on Probabilities thus: “We ought to regard the present state of the universe as the effect of its anterior state and as the cause of the one which is to follow. Given for one instant an intelligence which could comprehend all the forces by which nature is animated and the respective situation of the beings that compose it – an intelligence sufficiently vast to submit these data to analysis – it would embrace in the same formula the movements of the greatest bodies of the universe and those of the lightest atom; for it, nothing would be uncertain and the future, as the past, would present to its eyes. ”8

This famous quote can be illustrated by the following thought experiment: Assume a film is made of the world, i.e. of the events happening in the whole universe. After the film is developed, we cut it into pieces corresponding to single film pictures. Now, we put the single pictures successively in time (in the order of time) into a long card index box, like the cards of a library catalogue. The one special state of the universe at a certain time t corresponds to one such card (film picture) of the catalogue. One can follow one trajectory across the (perpendicular to the) catalogue-cards. 8

Laplace (1814, EPr), ch. 2.

On the Unifying Role of Laws of Nature

137

Interpreted with the help of this illustration, Laplace’s idea expressed in the quotation means that it suffices to know the law(s) of nature and one single catalogue card (film picture) corresponding to one state (of the universe) at a certain time t in order do construct all other cards of the catalogue, i.e. to predict and to retrodict all the other states of the universe. Put more formally, the gist of this idea is expressed in the following condition: D1:

The state S1 at t1 (of the physical system, or any classical part of the system) plus the laws of nature determine sufficiently any other (any later or earlier) state S2 at t2.

The respective causal principle fitting the dynamical laws is the following: CP1: The same initial state leads – under the same laws – to the same series of successor states. That means that in all cases where dynamical laws are applicable for calculating predictions, the causal relation represented by the law unifies all anterior and successor states under one principle. It should also be noted that the description of Laplace does not give a definite answer to all questions about the universe (even if it would be deterministic). Already Thomas Aquinas pointed out that such a picture of the universe does not give an answer to the question whether the universe had a beginning, i.e. has a finite age or not. His reason was very simple: universal laws do not tell us whether there was a singularity (a certain point of time with which all began) because they abstract from “hic” and “nunc” (from here, place, and now, specific point of time). Thus in the above illustration the question whether there is a first or last catalogue card is not answered. According to Thomas Aquinas such an answer like “God created the world a finite time ago” can be given by faith only9. Today we may add that because of the discovery of the Cosmic Background Radiation (by Penzias and Wilson in 1965) the Standard Big Bang Theory which claims a finite age of the universe is more probable than the mathematical consistent (but physically not confirmed) theory (by St. Hawking) of a “nonboundary-universe” infinite in time. 9

Cf. Thomas Aquinas (STh) I, 46,2.

138

Paul Weingartner

There are further conditions for the application of dynamical laws, especially that the system has a certain type of stability. D2

Very small changes in the initial states lead to proportionally small changes in the final state.

This condition was not understood well scientifically until Hadamard and Poincaré investigated it theoretically (around 1900) and Lorenz, Chririkov, Miles and others investigated it experimentally (since 1963). But already Aristotle and Maxwell were aware of the problem: “Do similar causes produce always similar effects?”, asks Maxwell and says that although this holds in many cases, it does not hold always: “There are other cases in which a small initial variation may produce a very great change in the final state of the system.”10 And Aristotle saw this problem from the epistemological point of view in connection with increasing error: “The least initial deviation from the truth is multiplied later a thousandfold.”11 The violation of condition D2 as expressed in Maxwell's quotation is a necessary condition for chaotic motion in the sense of dynamical chaos.12 The underlying laws of chaotic motion in the sense of dynamical chaos13 are still dynamical laws. But there is no predictability (except for an extremely short time). Therefore we cannot have the causal relation CP1 as in the case of Laplace's universe of Classical Mechanics. But there is still causality involved as unifying the chaotic states. This may be expressed by the following weaker principle of causality: CP2 Two similar states lead – under a positive Ljapunov exponent – to two separated systems where the distance and the degree of disorder and the loss of information (between two related points) is measured by a positive Ljapunov exponent (Kolmogorov entropy).

10

Maxwell (1991, MaM), p. 13. Aristotle (Heav) 271b8. 12 For details see Weingartner/Schurz (1996, LPL). 13 There is also quantum chaos and cosmological chaos which has different properties from dynamical chaos. For eight conditions of dynamical chaos see Weingartner (1996, UWT). 11

On the Unifying Role of Laws of Nature

139

Since the 19th century it was known that phenomena of heat, friction, radiation, diffusion, electric transport and others cannot be described adequately by dynamical laws. There was a general doubt now concerning the mechanistic world view which tried to explain all complex systems like gases, swarms of mosquitoes and clouds as ultimately consisting of small particles interacting in a mechanistic way; this view may be put into the slogan used by Popper: “All clouds are clocks.”14 Such doubts lead first to the other extreme: Could it not be that all complex systems of the world are in fact, in their inmost structure, on the atomic level, like gases or swarms of mosquitoes or clouds? In this case the above slogan should be reversed: “All clocks are clouds.” A consequence would be that all laws of nature are in fact statistical laws and the deterministic outlook is only on the surface of macroscopic phenomena. In fact both extreme views are not correct. It can be convincingly shown that neither statistical laws are reducible to dynamical laws nor dynamical laws are reducible to statistical laws.15 A first proof of that is the fact that statistical laws do not obey the condition D1 for dynamical laws: The state Si of the physical system at ti is not a definite function of an earlier state Si-1 at ti-1. The same initial state may lead to different successor states (branching). A respective principle of causality for statistical laws can be formulated in the following way: CP3 The same initial state may lead to different series successor states. But those successor states which belong to the same initial state, obey the same statistics. That means that the successor states need not be directly causally related, but those are related and unified (by the causal relation) which obey the same statistics. 3.2 Temporal Order and the Chronology Condition The condition of Temporal Order means that in every causal relation the cause must be earlier than the effect. At Newton's time it was assumed that one could allow the cause to be simultaneous with the effect, too, such that the cause was understood to be either earlier or simultaneous with the ef14

Popper (1965, CaC), p. 210. This has been shown in ch. 7.2 of Mittelstaedt/Weingartner (2005, LNt), p. 145165. 15

140

Paul Weingartner

fect. But the Theory of Special Relativity together with its confirming experimental tests has shown unambiguously that every causal propagation or influence needs time and can proceed at most with the velocity of light. In other words: For two events x, y  M (where M is the Minkowski spacetime structure) the causality relation C(x, y) holds when some signal can propagate from x to y. Therefore we have to say that every cause must be earlier than its effect. The Chronology Condition says that the time coordinate is not closed (like a loop), or in other words that there are no closed time-like curves. If the Chronology Condition is violated, there are drastic consequences like “time travel” and “backward causation” for the causal relation which however do not seem to be realistic.16 This is expressed by the following quotation from Hawking/Ellis: “This shows that in physically realistic solutions the causality and chronology conditions are equivalent.”17 On the other hand we assume – in accordance with the General Theory of Relativity – that space is closed, i.e. the space-like curves are closed. In other words: the universe w.r.t. space is finite and closed. In the antiquity and later in the early Middle Ages there were theories of the universe which did not obey the Chronology Condition, but assumed a “cyclic universe”. Some of the proponents of a cyclic universe used this theory as an attack against a creator of the universe. Against such a view Basilius (…) pointed out that – although we should not slit open the circle at one place and putting there a first cause – this is not a refutation of creation at all, since we can always ask “Who invented (created) the Circle?”. 3.3 Teleological Cause? According to Aristotle, every cause is an arché, i.e. something which is the first from which either being or coming into being or knowledge starts.18 He distinguishes four types of cause: (1) Material cause (Question: out of what?). Interpretation (in the light of our knowledge today): The chemical elements, the material constants like the elastic modulus of a metal etc. 16 17 18

Cf. the investigations of Thorne (1994, BHW) and others. Hawking/Ellis (1973, LSS), p. 192. Cf. Aristotle (Met) V,2.

On the Unifying Role of Laws of Nature

141

(2) Formal cause (Question: what is it?). Interpretation: The atomic or the crystal structure (form). The DNA of a species, etc. (3) Efficient cause (Question: which is the starting point, who is the producer or mover?). Interpretation: The parents, the car driver, the initial state of a dynamical system, etc. (4) Final cause (Question: why?, for what?, what for?). Interpretation (w.r.t. why): The premises or the law w.r.t. the conclusion or to the explained phenomena. According to Aristotle, every scientific explanation is an answer to a why-question.19 Interpretation (w.r.t. “for what?” and “what for?”): The goal, purpose, aim, motive, intention, etc. This second interpretation is the usual one, but the answer to the “why?”-question is equally important for Aristotle. Now concerning teleological causes in modern science, we can distinguish two questions: (a) Are there teleological arguments which use teleological causes in their premises and which are logically valid? (b) Are there certain states of physical or biological systems which can be interpreted as teleological causes? Question (a) can be answered with “Yes”20 independently of a positive answer to question (b). An answer to question (b) cannot be given by the philosopher of science alone. He needs the scientist (physicist or biologist) to at least offer a suitable possibility for a teleological interpretation. But as modern biology shows, there are indeed some interesting candidates. One is the so-called “evolutionarity stabilised strategy” (ESS) investigated by Maynard-Smith, Szathmáry, Wickler and others.21 The new discovery is that ESS, which is a special kind of equilibrium in a population depending on different factors like the kind of consanguinity or the strength in competition, the time lost for recovering after being injured etc., will be ultimately realised.

19 20 21

Cf. Weingartner (1991, NAT). Cf. Weingartner (1984, ITA). Maynard-Smith (1984, ETG). Wickler/Seibt (1987, PEN).

142

Paul Weingartner

4. Special Relativity – Unifying Inertial Systems 4.1 The Unification of Newton’s Theory Already Newton’s laws unify all inertial physical reference systems via Galilei Invariance. That means that these laws do not care about -

where you locate your laboratory (physical reference system) how you orient you laboratory how you set your clocks (in the laboratory) how fast your laboratory is moving (on a straight line) without acceleration.

There are however hidden assumptions underlying Galilean Invariance, which are the following: -

the time scale is the same in all inertial reference systems simultaneity is the same in all inertial reference systems the spatial distance of two simultaneous events is the same in all inertial reference systems.

It was shown by Einstein’s Special Theory of Relativity that these hidden assumptions are not generally true. 4.2 How to Unify Newton (CM) and Maxwell (ME)? In his Special Theory of Relativity Einstein's problem was a question of the invariance of physical laws in the domain of Classical mechanics (CM) and Electrodynamics (Maxwell’s Equations, ME); or in other words a question of unification of CM with ME: Is it possible to formulate both, the laws of CM and ME, in such a way that both laws are invariant under transformations of inertial systems? That this is in fact possible is claimed in a postulate by Einstein on the first page of his essay on the Special Theory of Relativity of 1905 “Zur Elektrodynamik bewegter Körper”: “The same laws of electrodynamics and optics are valid for all frames of reference for which the equations of mechanics hold good… We will raise this conjecture (the purport of which will thereafter be called the ‘Principle of Relativity’) to the status of a postulate…”22 22

Einstein (1905, EBK), Engl. Transl. p. 37.

On the Unifying Role of Laws of Nature

143

To solve this problem Einstein had three options: (1) Assume an ether as inertial frame for ME where CM is unified by Galilean Invariance. (2) Adapt ME to CM such that both are Galileo invariant. (3) Invent a new unifying principle of invariance which is obeyed by ME and adapt and correct CM. Einstein has chosen (3) even before any experiment had confirmed his choice. The new unifying principle is the principle of Special Relativity (SR): The laws of nature are the same in all inertial reference frames without the three hidden assumptions. The new assumption is: v ≤ c. That means that no physical body or particle can travel faster than the velocity of light c (in vacuum). It means further that no causal propagation or influence can be faster than c. The new unifying invariance is Lorentz-Invariance (in contradistinction to Galilei-Invariance). Lorentz-Invariance corrects the following magnitudes of CM: mass, length, time and simultaneity: In case of velocity, close to c mass increases. length is contracted (in the direction of movement) and time is stretched or delayed; and simultaneity is no more a universal concept. 5. General Relativity – Unifying any Moved Systems To formulate a theory which unifies any moved systems with the help of most general laws was the original intention of Einstein when constructing his General Theory of Relativity (GR). Although this aim – to find equations for any moved system – has not been achieved, GR is today understood as the most general theory of Gravitation which includes SR (and with it an unification of CM and ME). To proceed from SR to GR, Lorentz-Invariance had to be transcended for a higher unification in essentially three steps: (1) by dropping the restriction to inertial reference frames, i.e. by allowing acceleration; (2) by dropping the restriction to straight Galilean Coordinates, i.e. by allowing curved space-time; (3) by incorporating gravitation. The underlying unifying principles used were the Principle of Equivalence and the principle of General Invariance and General Covariance.

144

Paul Weingartner

5.1 The Principle of Equivalence Physical laws are invariant under changes of “free falling" local inertial reference frames: -

reference frames (r.f.) at rest in gravitation field  r.f. moving with acceleration free falling r.f.  r.f. with inertial movement

The deeper reason for the principle of equivalence is the equivalence of inertial mass and gravitational mass. But there is no satisfactory explanation for it so far. 5.2 General Invariance and General Covariance E1 “The laws of nature must be of such a nature that they are valid with respect to systems of reference in arbitrary motion.”23 E2 “The general laws of nature are to be expressed by equations which hold in all systems of coordinates (i.e. which are generally covariant).”24 E1 focuses on physical reference systems under arbitrary motion (aspect: A). E2 focuses on mathematical coordinate systems under arbitrary transformation (aspect: B). The guess (and claim) that A and B are equivalent, led to great confusion. In fact not every transformation of a coordinate system does correspond (or is representable) by a change of the reference system. This can be seen from the subsequent examples: Galilean Invariance: No equivalence between A and B if the hidden assumptions (universal temporal metric, arbitrary velocity) are not restricted. SR: Equivalence between A and B for inertial systems. GR: B (the general covariant formulation of laws) does not imply a generalised relativity principle about reference systems under arbitrary motion (A).25 23

Einstein (1916, GAR), p. 772. Engl. Transl. p. 113. Ibid. p. 775f. Engl. Transl. p. 117. 25 For a detailed discussion of aspects A and B concerning E1 and E2 see Mittelstaedt/Weingartner (2005, LNt), ch. 6.7.2, p. 127ff. 24

On the Unifying Role of Laws of Nature

145

6. Unification of Forces. Unified Theory The four fundamental interactions The four fundamental forces (a) Weak Interaction (force) It is responsible for -decay; it violates parity and CP-(charge parity)symmetry; it is concerned with leptons (spin 1/2). (b) Strong Interaction (force) It is responsible for  and -decay; it increases with the distance (within certain limits); it concerns mesons (spin 1,0) and baryons (spin 1/2). (c) Electromagnetic Interaction (force) It is responsible for the atomic structure and the binding of atoms to molecules; it has an unlimited range; in its macroscopic domain it is described by Maxwell’s Theory; in its microscopic domain it is described by QED (Quantum Electrodynamics) and concerns leptons; it is used in scattering experiments with electrons to investigate hadrons. (d) Gravitation Gravitational force has an unlimited range; on the microlevel the electromagnetic force is much stronger, but on the cosmological level its role is dominant. Weak interaction and electromagnetic interaction is unified by the Glashow Weinberg Salam Model. This theory unifies QED and QFT (Quantum Field Theory). The unification begins at 10-16 cm (not above). Weak, Strong and Electromagnetic Interaction is unified (not above 10-29 cm) by the so-called Grand Unified Theory (GUT). All four interactions (forces) are supposed to be unified by a Theory of Everything (TOE)26.

26

“Theory of Everything” is however a very misleading name. What is supposed to be unified are forces applicable to physical bodies; but it is not clear whether they are sufficient to explain all biological processes, much less mental phenomena.

146

Paul Weingartner

References Aristotle (Heav) “On the Heavens”, in: Barnes, J. (ed.). The Complete Works of Aristotle. The revised Oxford Translation, Vol 1. Princeton, Princeton University Press 1985. -

(Met) “Metaphysics”, in: Barnes, J. (ed.). The Complete Works of Aristotle. The Revised Oxford Translation, Vol. 2. Princeton, 1985.

Einstein, A. (1905, EBK) “Zur Elektrodynamik bewegter Körper”, in: Annalen der Physik 17 (1905), p. 891-921. Engl. Transl. in: Lorentz, H.A. et al. (1923, PRT), p. 37-65. -

(1916, GAR) “Die Grundlagen der allgemeinen Relativitätstheorie”, in: Annalen der Physik 49 (1916)., p. 769-822. Engl. Transl. in: Lorentz, H.A. et al. (1923, PRT), p. 109-164.

Hagiwara, K. et al. (2002, PDG) “Particle Data Group”, in: Physical Review D66, 010001. Hawking, S.W. and Ellis, G.F.R. (1973, LSS) The Large Scale Structure of SpaceTime. Cambridge, Cambridge University Press. Laplace, P. (1814, EPr) Essai philosophique sur les probabilities. Paris, Courcier. Engl. Transl.: A Philosophical Essay on Probablities. New York, Dover 1951. Lorentz, H.A. et al. (1923, PRT) The Principle of Relativity. London, Dover. Maxwell, J.C. (1991, MaM) Matter and Motion. New York, Dover. (Based on the edition by J. Larmor of 1920) Maynard-Smith J. (1984, ETG) Evolution and the Theory of Games. Cambridge, Cambridge University Press. Mittelstaedt, P. and Weingartner, P. (2005, LNt) Laws of Nature. Berlin, Heidelberg, Springer. Popper, K.R. (1965, CaC) “Of Clouds and Clocks”, in: Popper, K.R. Objective Knowledge. Oxford, Oxford University Press 1972, p. 206-255. Prigogine, I. and Stengers, I. (1993, PZt) Das Paradox der Zeit. Zeit, Chaos und Quanten. München, Piper. Prigogine, I. (1993, TDC) “Time, Dynamics and Chaos: Integrating Poincaré's NonIntegrable Systems”, in: Holte, J. Chaos: The New Science. London. Nobel Conference XXVI. University Press of America. Schwarzschild, B. (1999, TEO) “Two Experiments Observe Explicit Violation of Time-Reversal Symmetry”, in: Physics Today, Feb. 1999, p. 19-20. Thomas Aquinas (STh) Summa Theologica. Transl. by Fathers of the English Dominican Province (Christian Classics). Westminster, Maryland, 1948, 1981.

On the Unifying Role of Laws of Nature

147

Thorne, K. (1994, BHW) Black Holes and Time Warps.New York, W.W. Norton & Co. Inc. Vaccaro, J.A. (2011, TVU) “T Violation and the Unidirectionality of Time”, in: Foundations of Physics 41, p. 1569-1596. Weinberg, S. (1987, TFL) “Towards the Final Laws of Physics”, in: Elementary Particles and the Laws of Physics. The 1986 Dirac Memorial Lectures. Cambridge, Cambridge University Press, p. 61-110. Weingartner, P. (1984, ITA) “On the Introduction of Teleological Arguments into Scientific Discourse”, in: La Science Face aux Attentes de l'Homme Contemporain. Bruxelles, Archives de l'Institut International des Sciences Théoriques 26. Bruxelles, p. 196-218. -

(1991, NAT) “A Note on Aristotle's Theory of Definition and Scientific Explanation”, in: Spohn, W. (ed.). Existence and Explanation. Dordrecht, Kluwer, p. 207-217.

-

(1996, UWT) “Under what Transformations are Laws Invariant?”, in: Weingartner, P. and Schurz, G. (1996, LPL), p. 47-88.

Weingartner, P. and Schurz, G. (eds.) (1996, LPL) Law and Prediction in the Light of Chaos. Research Lecture Notes in Physics 473. Berlin, Springer. Wickler, W. and Seibt, U. (1987, PEN) Das Prinzip Eigennutz. Hamburg, Fischer. Wigner, E.P. (1967, SRf) Symmetries and Reflections. Scientific Essays of Eugene P. Wigner. Moore, W.J. and Scriven, M. (eds.). Bloomington, Indiana University Press.

Finitist Objects SERGIO GALVAN, Milan

The aim of the present essay is to enquire the nature of finitist evidence by examining two fundamental problems: (1) what are the objects of finitist evidence, i.e. what are finitist objects? (2) what formal system characterizes finitist evidence? The first problem will be investigated by taking into consideration its historical roots up to Hilbert’s work and recent interpretations of Hilbert’s finitism by Parsons and Tait. Addresssing the second problem will consist of an examination of I0 and PRA, understood as systems that formalise the idea of finitist evidence, as they have been put forward respectively by Parsons and Tait. While Parsons defends the thesis that finitist objects are quasi-concrete objects and Tait, on the contrary, supports the abstract nature of these objects, in this essay the claim is made that finitist objects are quasi-abstract objects. 1. What are the objects of finitist evidence, i.e. finitist objects? There is no non-controversial interpretation of Hilbert’s thought on the precise nature of finitist objects. But there is no doubt that for Hilbert finitist objects are “extralogical concrete objects that are intuitively present as immediate experience prior to all thought”.1 As objects which are present to our immediate experience, there exist only a finite number of such objects. Thus, according to Hilbert’s dictum, finiteness and concreteness are the components of finitist evidence. Now, it seems that the concept of finite does not suffer any particular problem of interpretation. The concept of finiteness is a primitive concept we fully comprehend. That a quantity of objects is finite is something conceptually crystal-clear, even though it can be practically difficult to establish the finite quantity of very big sets. The relevant aspect of Hilbert’s theory is rather that we can have intuition only of finite objects. Infinite objects, instead, cannot be made accessible to intuition. This epistemically problematic aspect of the notion of infinite 1

See Hilbert (1926), p. 376.

150

Sergio Galvan

reminds us of the other characteristic of finitist objects: their concreteness. This raises a problem: why can we have intuition only of concrete objects? Examination of this problem is particularly difficult, since there are many notions of concreteness. Two of them are of particular importance in the context of this essay. 1. Concrete means particular. Precisely, concrete object = individual object = non universal object = complete object = object determined with respect to all properties. On this way of understanding it, concrete is the opposite of universal. 2. Concrete means this, here and now. On this way of understanding it concrete means located in space-time. According to this meaning, concrete = spatial-temporally determined object, from which it also follows that concrete = subject of perceptual intuition. Something would be nonconcrete on this way of thinking about concreteness if it is a-temporal and a-spatial. Consequently, deciding which sense of concreteness is implicit in the finitist notion of intuition is difficult unless one clears up the problem of what, in the philosophical tradition of the West, is meant by ‘intuition’. Obviously, this can only be done here in broad strokes. The idea of intuition, which relates to mathematical knowledge and should be sharply distinguished from the generic idea of ‘evidence’, arose with new scholasticism. The notion introduced by Scotus (and then developed by Ockham) was that of the ‘intuition of an object’, not ‘intuition that’. The novelty is that intuitive knowledge (notitia intuitiva) in Scotus and Ockham concerns only existing objects. It is of these that we can have intuition, while abstractive knowledge (notitia abstractiva) can concern properties and relations among possibles: For intuitive cognition of a thing is a cognition such that by virtue of it it can be known whether the thing exists or not, in such a way that if the thing does exist, the 2 intellect at once judges it to exist and evidently knows it to exist … Abstractive cognition, however, is that by virtue of which it cannot be evidently known of the thing whether it exists or does not exist. And in this way abstractive cognition, as opposed to intuitive cognition, “abstracts” from existence and non-

2

Ockham, Prologue to the Ordinatio, q. 1, a. 1, Opera Theologica, 10 Voll., St. Bonaventure, New York 1967-1984, I Vol., p. 31; english translation by P.V. Spade; see also Ockham, Reportatio, q. 15, E, ed. Lyon 1494-1496.

Finitist Objects

151

existence, because by it neither can it be evidently known of an existing thing that it exists, nor of a non-existent one that it does not exist.3

It is evident from the two quotations that what characterizes intuitive knowledge is the existence of the objects intuited. But, according to Ockam, what exists is particular. We can therefore intuit only particular objects. In this sense, the idea of intuition differs greatly from the idea of scientific evidence, which according to the Aristotelian tradition exclusively pertains to the universal, so that there can be science only about the universal. By contrast, intuition pertains to the singular, of which it is equally a true vehicle of knowledge. It is clear, however, why the objects of intuition are concrete. They are concrete because only concrete objects can exist and only existing objects can be intuited. The theory of intuitive knowledge established itself in the modern age, although until Kant the type of intuition involved remained indeterminate. The crucial change came with Kant, who declared that intuition puts us into an immediate relation with objects grasped in space and time. This was for the simple reason that, according to Kant, we can establish the existence only of empirical objects. In his opinion, in fact, the existence of an object is nothing other than its position within the sphere of experience. Therefore, empirical immediateness means presence to the mind of the particular object existing in space-time, and intuition cannot but be perceptive. The transformation of intuition into perceptive (sensible) intuition was thus accomplished. Hilbert inherited from Kant the idea of perceptive intuition and the change of meaning of the term ‘concrete’ from simple particular to space-time-determined particular. Of course, Kant introduced the fundamental distinction between empirical intuition and a priori intuition, but, in any case, this always concerned intuition in space and time. It is for this reason that, in conclusion, according to some interpretations of Hilbert’s thought, intuitive objects are concrete (anschaulich, i.e. space-timedetermined) objects: they are signs and configurations of signs located in space and time. As Gödel writes in footnote b of Gödel (1972): 3

Ockham, Prologue to the Ordinatio, q. 1, a. 1, Opera Theologica, 10 Voll., St. Bonaventure, New York 1967-1984, I Vol., p. 32; english translation by P.V. Spade. See also: Ockham, Prologue to the Ordinatio, q. 1, a. 1, Opera Theologica, 10 Voll., St. Bonaventure, New York 1967-1984, I Vol., p. 36; Ockam, Quodlibet I, q. 14, Opera Theologica, 10 Voll., St. Bonaventure, New York 1967-1984, IX Vol., p. 79.

152

Sergio Galvan

‘Concrete intuition’, ‘concretely intuitive’ are used as translations of ‘Anschaung’, ‘anschaulich’. The simple terms ‘concrete’ or ‘intuitive’ are also used in this sense in the present paper. What Hilbert means by ‘Anschauung’ is substantially Kant’s space-time intuition confined, however, to configurations of a finite number of discrete objects. Note that it is Hilbert’s insistence on concrete knowledge that makes finitary mathematics so surprisingly weak and excludes many things that are just as incontrovertibly evident to everybody as finitary number theory. … There is nothing in the term ‘finitary’ which would suggest a restriction to concrete knowledge. Only Hilbert’s special interpretation of it introduces this restriction.

However, there now arises a problem: Are finitist objects tokens or types? 4 The problem was acknowledged by Hilbert himself, when he affirmed concerning numerals that their “shape can be generally and certainly recognized by us independently of space and time, of the special conditions of the production of the sign, and of the insignificant differences in the finished product.”5 The solution to this problem is implicit, I believe, in the theory of the numerals as formal objects developed by Hilbert and Bernays in 1939 and anticipated by Bernays in 1930. Nevertheless C. Parsons has provided, in my opinion, the most systematic and explicit answer to the problem.6 Parsons’ answer has then formed the basis for his perceptual interpretation of Hilbert’s finitism; and it is the essential feature that sets his perceptual finitism in opposition to Tait’s abstract finitism.7 It is accordingly worth examining Parsons’ theses, beginning with his distinction among pure, quasi-concrete, and concrete objects: - Pure objects are totally abstract and hence cannot be intuited. For Parsons, natural numbers (one, two,...) are of such a type: they are purely abstract, because they result from the abstraction of quasi-concrete objects. - Quasi-concrete objects are abstract objects capable of having concrete instances. For example, numeral-types are quasi-concrete objects. Quasiconcrete objects can be intuited, because they are concretely instantiable. - Concrete objects are instances of quasi-concrete objects. For example, numeral-tokens are concrete objects, because they are exemplifications of numeral-types. Of course, they can be directly intuited and, therefore, space-time determined. 4 5 6 7

The problem was already raised by P. Bernays in Bernays (1923). See Hilbert (1922), p. 202. See Parsons (1980), Parsons (1983) and Parsons (1998). See Tait (2005).

Finitist Objects

153

Omega-sequences are sets of quasi-concrete objects. A paradigmatic example is the omega-sequence of stroke-numerals. The omega-sequence of stroke-numerals L is constituted by: (a) Elements of L: Base-element = | (one stroke). Finite sequences of strokes (numerals), as |||||||. They are the result of adding a stroke (succession function) to sequences previously obtained. (b) Structure of L: arrangement of the numerals in order of succession, i.e. L = |, ||, |||, ||||, … The first thing to be noted is that, for Parsons, there is no particular omega-sequence that can be identified with the structure of natural numbers. The latter is the structure of all structures isomorphic with omegasequences. Natural numbers are pure objects whose only essential properties are those that they share with the objects which form an omegasequence. From a structuralist point of view, this amounts to saying that the meaning of a number is the role of the position corresponding to it in any omega-structure. In other words, what is essential to a number is the set of relations that define the position corresponding to it in an omega-structure. Natural numbers, therefore, are not subject to intuition and consequently they are not the objects of finitary arithmetic. Instead, the objects of finitary mathematics are, for Parsons, the quasiconcrete objects of an omega-sequence whose concrete space-time determined instances (tokens) are susceptible to perceptual intuition. Parsons’s thesis contrasts with Tait’s thesis on finitism8. For Tait, in fact, finitist evidence does not rest ultimately on perceptual intuition. It instead derives from the intuition of the concept of finite sequence (an object constructed by iteration) of any objects, of which the numeral sequences are only physical representations and not in principle necessary.9 The independence of the concept of number from its intuitive representation is forcefully reiterated, when Tait discusses the Kantian thesis that numbers like 1010, although they cannot be the object of intuition, can be decomposed into intuitively perceptible parts. He writes: But to have this idea, itself not found in intuition, is to have the idea of number in10 dependent of any sort of representation in intuition. 8 9 10

See Tait (1981) and Tait (2005). See, in particular, Tait (1981), pp. 529-530. Ibidem, p. 539.

154

Sergio Galvan

So, for Tait, the objects of finitary mathematics are the sequences of any objects constructed inductively (of which stroke-numerals are only an exemplification). Of course, on this interpretation, finitist objects become abstract objects (in Parsons’ sense) and not quasi-concrete ones. And in this sense it cannot be said that they are the objects of perceptual intuition. Tait’s conception also seems to be closer to certain recent interpretations of Hilbert’s theory of finitist signs, according to which they are concrete not because they are spatio-temporally determined but because they are generally determined.11 However it may be, theories like Tait’s which are grounded upon an abstract conception of finitist objects, have to face the problem of their intuitability. Can it be maintained that the properties of any object inductively constructed are accessible to intuition just as are those of Parson’s quasiconcrete objects? We should not forget that Hilbert’s requirement of concreteness holds also if concrete objects are not concrete in the spatialtemporal sense. It holds because concreteness is the reason for existence. If we cannot intuit an exemplification of a concept, we cannot exclude that it is an empty concept. Referral to finistist objects is then fundamental, since, being concrete, they are in principle accessible to intuition and, as a consequence, they can be said to exist. Now, how can we affirm that Tait’s finitist but abstract objects are intuitable? We’ll tackle this problem in next paragraph, after a detailed analysis of the systems that Parsons and Tait propose in order to characterize the finitist evidence. 2. What formal system characterizes finitist evidence? 2.1 Parsons’s System Parsons proposes12 the following I0 (or Q2 = PFA = Polynomial functional arithmetic) system as a system characterizing finitary evidence – i. e. evidence relative to the finitary properties (intuitive properties) of finitary objects (concrete instances of the numeral-types):

11 12

See Zach (2006), pp. 420-424. See, in particular, Parsons (1998).

Finitist Objects

155

SYSTEM Q Q1. – x(0 ≠ x') Q2. – xy(x' = y'  x = y) Q3. – y(¬(y = 0)  x(y = x')) Q4. – x(x+0 = x) Q5. – xy(x+y' = (x+y)') Q6. – xy(x·0 = 0) Q7. – xy(x·y' = x·y + x) SYSTEM I∆0 I∆0 = Q + IP∆0 where IP∆0 = Induction Principle for ∆0-formulas, i. e. (0) x((x)(x'))x(x) (for = -formula) In Parsons’ view, the reason for choosing I0 as system characterising finitist evidence is that the axioms of this system are finitarily intuitable, as they are interpretable on the omega-sequence L of the stroke-numerals. Let us see, therefore, the interpretation of this system on the omega-sequence L of the stroke-numerals. Where x varies on the set of stroke-numerals, one has:13 Q1: x(x| ≠ |) Q2: xy(x≠y  x| ≠ y|) Q3: y(y≠|  x(y = x|)) Q4: x(x+| = x|) Q5: xy(x+y| = (x+y)|) Q6: xy(x·| = x) Q7: xy(x·y| = x·y + x) IP∆0 = A(|)  x(A(x) A(x|))  xA(x) 13

The interpretation of Q on the omega-sequence L here assumed is characterized by replacing zero with the stroke and modifying the axioms Q4 and Q6. Another possibility consists in maintaining the language and axioms, but in interpreting zero on the absence of strokes. This is characteristic of the language of a Turing machine, where there are two types of sign: the stroke | and the blank cell (as absence of strokes). This interpretation forces to extend the range of variables to the set of strokes-numerals + the blank cell.

156

Sergio Galvan

Now: ad Q1: The justification of Q1 is intuitive. In fact, on taking any string of strokes (stroke-numeral) resulting from the addition of a stroke to a previous string of strokes (stroke-numeral), we can intuitively verify that it is different from one stroke. Even if the previous string is not perceptually present in detail, because it is very numerous, at least its confused representation (a vague pattern of strokes) is present. Thus the whole is intuitively grasped as different from one stroke. ad Q2: Q2 is intuitively immediate for stroke-numerals of limited size. But size is irrelevant given a sufficiently abstract definition of strokenumeral. Therefore, by generalization it holds that, given any two different stroke-numerals, the successive stroke-numerals will also be different. ad Q3: Q3 can be justified in a similar way by generalization. If y is an intuitable string constituted by several (more than one) strokes, then one can intuit the string that precedes y – that is, the one which results from cancelling the last stroke in y. But with respect to the existence of the previous string, the quantity of the strokes constituting y is irrelevant. Therefore one can say in general that if y is a string constituted by several (more than one) strokes, then the string that precedes y exists. Note that both justifications of Q2 and Q3 represent abandonment of the pure terrain of intuition. Idealization presupposes the grasping of certain structural aspects in the notion of stroke-numeral which arise from generalization by intellective abstraction of the contents of the intuition. But the role of rational evidence involved is limited. It only consists in the logical regimentation of the concept of stroke-numeral. The single stroke-numeral that results therefrom is merely the sensible representative of the corresponding number.14 ad Q4 and Q5: Q4 and Q5 are the equations defining the sum. Their intuitiveness is indirect. They perform an operation that sums the strings of stroke-numerals which is intuitive. The intuitiveness resides in the fact that the result of the operation must be a string containing exactly the strokes which constitute the strings to sum. And this is guaranteed for the intuitable strings. By generalization, the result holds for all the strings. ad Q6 and Q7: Q6 and Q7 are the equations defining the product. The previous considerations apply to these as well. Note that the intuitiveness 14

See for this Page (1993).

Finitist Objects

157

of the product operation derives from the fact that the product of string x by string y (y times string x) is equivalent to the result of substituting string x for every stroke in string y. A result that is intuitable. ad IP∆0: the intuitiveness is the result of a somewhat complex reasoning. First, the inductive formulas are bounded formulas. This means that the use of induction up to the ∆0-formulas does not imply acceptance of the existence-in-act of the infinite domain of natural numbers. Moreover, the induction is justified by the inductive nature of the strings of strokenumerals. It can be shown inductively that induction up to the ∆0-formulas does not imply acceptance of the existence-in-act of the infinite domain of natural numbers. Basis: Let A(x) be an open-formula. Hence its inductive use does not imply the actual infinite. Firstly, it does not imply it because the meaning of the formula, being open, does not require it. Secondly, obtaining xA(x) by induction has the same meaning as obtaining A(x), because the generalization is the immediate consequence of having available a procedure able to establish for any number x that A(x) holds for it. Therefore, to give sense to the expression, it is not necessary to presuppose that the domain exists in act; but simply that whenever one constructs a new number x, it can be demonstrated (that is, one has the procedure to generate the result) that it possesses the property denoted by A(x). Step: Let A(x) be a ∆0-formula. By inductive hypothesis its inductive use does not imply any assumption of the actual infinite. Hence, also x(x≤z (x)) has the same property. The reason is that the procedure does not state that there exists simply an x such that A(x), but an x smaller than or equal to a certain z such that A(x). Consequently, having selected a particular z, the x of which the existence is declared can be exhibited. By inspecting the numbers from 0 to z, the x in question, if it exists, will be found. Naturally the same conclusion holds for the universal quantifier.15

15

Note, however, that in the case of the universal quantifier its bounding is essential only if it follows an existential quantifier (i.e. is not external), because in this latter case the formula can be taken as a formula open with respect to the variable bound by that quantifier.

158

Sergio Galvan

2.2 Tait’s System Tait proposes as a system characterizing finitary evidence – i. e. evidence relative to the finitary properties (intuitive properties) of finitary objects (concrete instances of the numeral-types) – the system PRA of Primitive Recursive Arithmetic, that can be formulated as Q = Q + the equations for every primitive recursive function + IPR, i. e. the Induction Principle for PR-formulas. PRA includes I0. I0, corresponds to the fragment which contains only addition and multiplication and is closed to bounded primitive recursion, while PRA contains all the primitive recursive functions. The reasons why Parsons excludes PRA as the system of finitary evidence are the same that highlight Tait’s specific standpoint about the abstract nature of finitist objects. So, why does Parsons doubt that intuitive arithmetical knowledge goes beyond the fragment I0 of PRA? Why, in other words, does Parsons believe that finitism does not extend to the system of primitive recursive arithmetic? Parsons’ thesis is grounded on two fundamental reasons. The first pertains to the fact that, according to Parsons and other authors like E. Nelson16, one cannot maintain that the recursion of rapidly increasing functions like exponentiation is susceptible to intuitive insight. In other words, it is not possible intuitively to take for granted that, given any argument, the value of the relative power exists, or that the exponentiation operation is total. Of course, this (the totality of all the primitive recursive functions) can be demonstrated, but only by using induction up to the Σ1-formulas, therefore passing from I∆0 to IΣ1 (i. e. the system Q + IPΣ1, where IPΣ1 is the Induction Principle for Σ1-formulas). And here arises the second reason. IΣ1 is not a finitary acceptable system. The Σ1-formulas correspond, in fact, to properties that are not acceptable finitistically because they involve the existence of the totality of natural numbers. But if IΣ1 is unacceptable, also PRA is unacceptable because IΣ1 is reducible to PRA. 2.3 Critical remarks on Parsons’ theory In Tait’s construal it is possible to argue against both these reasons. I will first discuss the question of the provable totality in PRA of the primitive

16

See Nelson (1986).

Finitist Objects

159

recursive functions. I will then address the problem of the connections between PRA and IΣ1. As regards the first issue, it is correct to dispute the intuitiveness of the exponentiation operation. Nevertheless, we have seen that also the intuitiveness of the other axioms of Q (except for Q1) is only partial. In the strict sense, intuition is guaranteed only where stroke-numerals of small size are concerned. The generalization is based on an idealization of the objects involved which pivots on the irrelevance of the number of the stroke-numerals. Therefore the truth of such axioms is ensured by intuition plus a portion of rational (intellective) evidence, and not only by perceptual intuition. Why cannot the same form of intuition guarantee the existence of value for functions that increase more quickly than the sum and the product but which are nevertheless still recursive primitives? The crucial point concerns, not the perceptual intuitiveness of such functions but their structural construction. It is structurally evident that a function defined by recursion from functions defined for all arguments (total) is in its turn such (total). And this evidence does not rely on the presupposition that there exists the whole-in-act of natural numbers. This finds formal expression in the fact that in PRA (therefore without Σ1-induction) it is possible to prove that a function defined by recursion, under the constructive conditions just stated, is provably total with respect to all numerals. In other words, the following theorem holds. Theorem: let f be introduced by means of the recursion scheme: f(0) = g f(y’) = h(f(y)) Let it be by inductive hypothesis that h is a function by hypothesis already defined. It is therefore total with respect to the numerals (that is, for all n there exists an m such that – h(n ¯) = m ¯ )). Hence f is total for every numeral; that is, for every n there exists a k such that – f(n ¯) = ¯k. Proof: – – – – – –

f(y’) = h(f(y)) f(0’) = h(f(0)) f(0) = g f(0’) = h(g) h(g) = m ¯ f(0’) = m ¯

2° eq. def. Substitution (y/0) 1° eq. def. Substitution (f(0)/g) inductive hypothesis Transitivity

160

Sergio Galvan

– – – –

f(0’’) = h(f(0’)) f(0’’) = h(m ¯) h(m ¯ ) = ¯n f(0’’) = ¯n

from 2° eq. def. Substitution (f(0’)/ m ¯) inductive hypothesis Transitivity

……..…….. Generalization: for every n there exists a k such that – f(n¯) = ¯k.17 I’ll now move to the second theme. The relationships between PRA and IΣ1 can be summed up as the interpretability of PRA in IΣ1, the reducibility of IΣ1 to PRA, and the conservativeness of IΣ1 over PRA with respect to the Π2-formulas. (i) PRA is interpretable in IΣ1 in the sense that all the recursive primitive functions characteristic of PRA (and present in PRA as primitive functions ruled by the corresponding defining equations) can be definitionally introduced into IΣ1 (through proof of the respective conditions of existence and uniqueness). In other words, all the primitive recursive functions are provably total in IΣ1.18 (ii) IΣ1 is reducible to PRA in the sense that all the provably total functions in Σ1 are recursive primitive.19 (iii) IΣ1 is a conservative extension of PRA with respect to the Π2formulas.20 These three properties may give the impression that there is a substantial equivalence between PRA and IΣ1. But this is not the case. The fact that PRA is interpretable in IΣ1 and IΣ1 is reducible to PRA and conservative, 17

It is obviously possible to show in PRA also the totality not restricted to numerals of the primitive recursive functions. The provable totality not restricted to the numerals of the primitive recursive functions trivially follows in PRA from the linguistic primitivity of these functions. This, however, is a circular proof, because it depends on the use of the substitution rule (Substitution(x/t), and applying the substitution rule to f(x) has the same meaning as assuming – xz(f(x) = z)), which is precisely what was to be proved. In effect, to obtain the totality of the function without substitution, and therefore in non-circular manner, one must use a particular instance of 1-induction. 18 See Buss (1998), p. 97. 19 See Buss (1998), pp. 123-125. 20 See Avigad (2003), p. 269.

Finitist Objects

161

does not allow us to identify the role of PRA with that of IΣ1 in characterizing finitist evidence. As we have seen, there exists in PRA a metatheoretical (inductive) demonstration that for every primitive recursive function there exists the numeral value for every numeral argument selected. The difference between PRA and IΣ1 depends on the way in which the primitive recursive functions are introduced into the system. Whereas in IΣ1 they are introduced through derivation via induction on Σ1 (therefore based on unbounded existential statements) of the existence and uniqueness conditions, in PRA they are introduced as (undefined) primitive functions simultaneously with the introduction of the relative equations. This has as a consequence that the theory can be conceived as the result of the linguistic extension obtainable through adding, one after the other, all signs of the primitive recursive functions and the respective definitional equations, starting from the smallest subsystem. This way of seeing things allows us to prove that PRA is a reflexive theory, that is to say a theory in which the consistency of any of its finite subtheories is provable. However, this implies that PRA cannot be finitarily axiomatised, because in this case it could prove its own consistency, so breaching G2. Instead, IΣ1 is finitarily axiomatisable.21 This is a not negligible difference, since it implies an important consequence as to the model of both theories. While the standard model satisfying IΣ1 is a model constituted by a unique block, since it is not possible to distinguish between the IΣ1 system and its finite subsystems, the model for PRA, on the contrary, can be conceived, by analogy with the theory’s structure, as the result of the model’s extension to any of its finite subtheories. Thus, while the standard model captured by PRA can be interpreted constructively, i.e. in a manner that only involves a potential notion of infinity, IΣ1, by contrast, captures a standard model based on an actualist notion of infinity. Ergo, PRA is finitarily acceptable and not IΣ1. Thus, criticism of the first kind of reasons, such as the non intuitive character of some primitive recursive functions, goes along well with criticism of the second kind of reasons, such as equivalence of PRA with IΣ1.

21

See Hájek and Pudlák (1998), p. 78.

162

Sergio Galvan

2.3 Remarks about Tait The criticisms moved against Parsons’ theses certainly support Tait’s fundamental argument, according to which the organ of finitist evidence is PRA. However, information deriving from examination of the differences among PRA, I0 and IΣ1 are not sufficient to settle the question about the nature of finitist evidence. It is necessary to assess this information in the light of the previous reflections upon the nature of finitist objects. In the comments on Parsons’ position on finitist objects as quasi-concrete objects, it has been said that the peculiarity of Parsons’ thought consists in the fact that although such objects are types and therefore abstract, they are types which only admit of spatio-temporally determined exemplifications. By examining the I0 system, however, it was also noted that sometimes such exemplifications are only possible in the context of strongly idealized space-time. This reflection then led to criticism of the non finitist, because not space-temporally accessible to intuition, character of objects like the function of exponentiation and, as a consequence, to acceptation of PRA instead of I0. Let us now ask ourselves: Is all this sufficient for accepting all aspects of Tait’s conception about finitist objects? Tait’s idea we intend to critically examine is that finitist objects are types of concrete objects whatsoever, that is to say types of concrete objects inasmuch as they are simply determined (even if not spatio-temporally). Here lies the whole difference with respect to Parsons. Parsons’ quasi-concrete objects (for example numeral 2¯) only admit of spatio-temporally determined instances (sequence of two stroke-numerals ||), while Tait’s objects do not necessarily require exemplifications of this kind. On the other hand, only spatio-temporally determined objects are accessible to intuition. How is it then possible to establish their existence, if in principle they are not always accessible to intuition? I think there are two ways of answering this question. According to the first one concreteness (as determinateness) is sufficient in order for the object to be intuited; in this case, however, intuitability is to be understood as an intuition of the way in which such objects can be constructed, which refers to the question of the abstract criterion of construction. Moreover, intuition would no longer have an empirical character, which is a necessary condition in order to be an intuition of the existent; it would instead be an intuition of an abstract, intellective kind.

Finitist Objects

163

The second way consists in emphasizing the idealization condition of the spatio-temporal component of the exemplifications (numeral instances). Objects are finitist to the extent in which they also have a spatiotemporally determined exemplification. Intuitability would so derive from the necessary presence of the spatio-temporal instance. The immediate consequence of the just given definition of finitist object is the finitness condition, that depends on the condition of determinateness. I believe that this second solution is preferable. It consists in conceiving finitist objects as quasi-abstract objects. They are not purely abstract, since among their exemplifications are included concrete signs and their finite configurations. One can agree with Tait that the conceptual space opened by the system PRA coincides with the space of the sequences of any objects constructed inductively. However, it must be added that among them also concrete exemplifications of stroke-numerals exist and of this some account must be given. This necessary referral to spatio-temporally determined forms of exemplification makes up the concrete component of the finitist object. It is however weaker than that required by Parsons, since it is present together with others which are excluded by the quasi-concrete conception of the latter. A further difference distinguishes the quasi-abstract notion of finitist object from Tait’s abstract one. The difference lies in the fact that finitness remains a fundamental requirement of finitist objects. Therefore, it is not a legitimate move, in my opinion, to deny outright, as Tait does in the following paragraph, the gap between finite and infinite: With the rejection of numbers as representable in intuition, I think that we are also led to reject Hilbert’s distinction between the finite sign as a real object and transfinite objects as ideal and ideas of pure reason. 1010 is no more representable in intuition than is the sequence 0, 1, 2, … of all the numbers. This too is obtained by ite22 rating the successor operation – only infinitely often.

The idea of an infinite iteration is finitistically meaningless. As Bernays says: It is by no means compatible, however, with Hilbert’s basic thoughts to introduce the numbers as ideal objects "with quite different determinations from those of sensible objects", "which exist entirely independent of us". By this we would go beyond the domain of the immediately certain. In particular, this would be evident 22

Ibidem, p. 540.

164

Sergio Galvan

in the fact that we would consequently have to assume the numbers as all existing simultaneously. But this would mean to assume at the outset that which Hilbert 23 considers to be problematic.

3. Conclusions The conclusions that we may draw are the following: (i) The finitist objects are all and only the quasi-abstracts objects (types) constructable in a primitive recursive manner. (ii) They are not directly intuitable: instead what is intuitable – at least in principle – are the relative concrete instances (tokens), that is, the numeral configurations (stroke-numerals) and their finite combinations. However the finitist objects are indirectly intuitable in (or by means of) their concrete instances, which they, as quasi-abstract objects and not purely abstract, necessarily – at least in principle – admit. We call this form of intuition finitist. (iii) The logical and arithmetical inference within PRA is based on finitist intuition of the basic primitive recursive functions, and on finitist intuition of the formal mechanism of constructing the functions and deriving their properties. (iv) Therefore, everything that can be said about any object within PRA exactly corresponds to what can be said within PRA about finitist objects and, then, about stroke-numerals as their concrete instances. It is not possible within PRA to attribute properties to numbers which are not representable without gaps also as spatio-temporal properties of stroke-numerals. For example, a finitist meaning cannot be given to the property "not to be derivable in PRA" by the numerical code of Gödel’s proposition. For these reasons finitist evidence is characterizable by the PRA system. Bibliographical References Avigad J. (2003), “Number Theory and Elementary Arithmetic”, Philosophia Mathematica, Vol. 11 (2003), pp. 257-284. Bernays, P. (1998), “Reply to the Note by Mr. Aloys Müller, ‘On Numbers as Signs’”, in Mancosu (1998), pp. 223-26, translation by P. Mancosu from “Erwiderung

23

See Bernays (1923), pp. 225-26.

Finitist Objects

165

auf die Note von Herrn Aloys Müller: Über Zahlen als Zeichen”, Mathematische Annalen 90, 1923, pp. 159-63. Buss, S.R. (1998), “First-Order Proof theory of Arithmetic”, in Buss, S. R. (ed.), Handbook of Proof Theory, Elsevier Science B.V. Butterfield, J. (1986) (ed.), Language, Mind and Logic, Cambridge University Press, Cambridge 1986. Gödel, K. (1958), “On a hitherto unutilized extension of the finitary standpoint (1958)”, translation from “Über eine bisher noch nicht benützte Erweiterung des finiten Standpunktes (1958)”, in CW 1990, pp. 240-251. -

(1972), “On an extension of finitary mathematics which has not yet been used (1972)”, translation from “Über eine bisher noch nicht benützte Erweiterung des finiten Standpunktes”, in CW 1990, pp. 271-280.

-

(CW 1986), Collected Works; vol. I, Publications 1929-1936 (eds. S. Feferman, J.W. Dawson Jr., S. C. Kleene, G. H. Moore, R. M. Solovay, J. van Heijenoort), Oxford University Press, New York – Oxford.

-

(CW 1990), Collected Works; Vol. II, Publications 1938-1974 (eds. S. Feferman, J. W. Dawson Jr., S. C. Kleene, G. H. Moore, R. M. Solovay, J. van Heijenoort), Oxford University Press, New York – Oxford.

-

(CW 1995), Collected Works; vol. III, Unpublished Essays and Lectures (eds. S. Feferman, J.W. Dawson Jr., W. Goldfarb, C. Parsons, R. M. Solovay), Oxford University Press, New York – Oxford.

Hájek, P. and Pudlák P. (1998), Metamathematics of First Order Arithmetic, Springer, Berlin – Heidelberg. Hart, W.D. (ed.) (1996), The Philosophy of Mathematics, Oxford University Press, Oxford. Hilbert, D. (1922), “Neubegründung der Mathematik: Erste Mitteilung”, Abhandlungen aus dem Seminar der Hamburgischen Universität, 1, pp. 157-177, english translation in Mancosu (1998), pp. 198-214. -

(1926), “Über das Unendliche”, Mathematische Annalen 95 (1926), pp. 161190, english translation, “On the infinite”, in van Heijenoort (1967), pp. 367392).

Hilbert, D. and Bernays, P. (19682), Grundlagen der Mathematik I, Springer-Verlag, Berlin – Heidelberg 1934, second edition 1968. -

(19702), Grundlagen der Mathematik II, Springer-Verlag, Berlin – Heidelberg 1939, second edition 1970.

Ioannis Duns Scoti, Commentaria Oxoniensa, Ad IV Libros Magistri Sententiarum, Ad Claras Aquas (Quaracchi), prope Florentiam, 1912. Mancosu, P. (1998), From Brouwer To Hilbert. The Debate on the Foundations of Mathematics in the 1920s, Oxford University Press, New York – Oxford.

166

Sergio Galvan

Nelson, E. (1986), Predicative Arithmetic, Mathematical Notes 32, Princeton University Press, Princeton, New Jersey. Page, J. (1993), “Parsons on Mathematical Intuition”, Mind, vol. 102, 1993, pp. 223231. Parsons, C. (1983), Mathematics in Philosophy. Selected Essays, Cornell University Press, Ithaca – New York. -

(1983), “Ontology and Mathematics”, in Parsons (1983), pp. 37-62.

-

(1986), “Intuition in constructive mathematics”, in Butterfield (1986), pp. 211-229.

-

(1993), “On Some Difficulties Concerning Intuition and Intuitive Knowledge”, Mind, vol. 102, 1993, pp. 233-245.

-

(1995), “Platonism and Mathematical Intuition in Kurt Gödel’s Thought”, The Bulletin of Symbolic Logic, 1, pp. 44-74.

-

(1996), “The structuralist view of mathematical objects”, in Hart (1996), pp. 272-309.

-

(1996), “Mathematical Intuition”, in Hart (1996), pp. 95-113.

-

(1998), “Finitism and Intuitive Knowledge”, in Schirn (1998), pp. 249-270.

-

(2000), “Reason and Intuition”, Synthese, 125, pp. 299-315.

Schirn, M. (ed.) (1998), The Philosophy of Mathematics Today, Clarendon Press, Oxford 1998. Schirn, M. and Niebergall, K.G. (2005), “Finitism = PRA? On a thesis of W. W. Tait”, Reports on Mathematical Logic, 39 (2005), pp. 3-26. Stamm, M. (ed.) (1998), Philosophie in synthetischer Absicht. Synthesis in Mind, Klett-Cotta, Stuttgart. Tait, W. (2005), The provenance of pure reason: essays in the philosophy of mathematics and its history, Oxford University Press, Oxford 2005. -

(1981), “Finitism”, The Journal of Philosophy, 78 (1981), pp. 524-556.

Van Heijenoort, J. (ed.) (1967), From Frege to Gödel. A Source Book in Mathematical Logic, 1879 – 1931, Harvard University Press, Cambridge, MA – London 1967. Zach, R. (2006), “Hilbert’s Program Then and Now”, in Jacquette, D. (ed.), Philosophy of Logic. Handbook of the Philosophy of Science, vol. 5, Elsevier, Amsterdam 2006, pp. 411-447.

Ontology and Ordinary Objects LYNNE RUDDER BAKER, Amherst

Edmund Runggaldier is well-known throughout the world for his work in Christian philosophy. To say that Edmund Runggaldier works in Christian philosophy, however, understates the breadth of his philosophical interests. In English alone, he has published papers on sortal continuity, on intentionality in Aristotle and Brentano, and on Carnap and other topics in analytic philosophy. He has edited books that cut a broad swath in philosophy generally. I take my theme from the broad swath in which there are many convergences of Professor Runggaldier’s views and my own. In this well-deserved Festschrift for Professor Runggaldier, I want to pursue some of the interests that he and I share. Most concern the reality of everyday objects. Although it may seem obvious that everyday objects – you, me, automobiles, computers, roses – are real, many philosophers deem them not to be objects at all, but just collections of particles. There is a venerable tradition in philosophy of clever arguments for unbelievable theses being countered by quick-and-dirty refutations. For example, when Zeno argued that there was no motion, Diogenes was said to walk around the room, thus disproving Zeno. When Berkeley argued that there was no material substance, but only ideas, Dr. Johnson famously kicked a stone and said, “I hereby refute you.” Today, clever philosophers are still advancing arguments with incredible conclusions. For example, Peter van Inwagen says, “My position vis-à-vis tables and other inanimate objects is simply that there are none. Tables are not defective objects or second-class citizens of the world; they are just not there at all.” (Van Inwagen 1990, 99-100) Now, I would like to pull a Dr. Johnson and say, “Then, what is that thing in the dining room supporting your knife and fork?” But van Inwagen has a ready response: Strictly speaking, nothing; there are just some simples (or particles) arranged tablewise (and knife-and-forkwise). Of course, in ordinary life, we say that there’s a table there. But we must sharply separate what we need to say to get along in the world and what we should say in the philosophy room when we do ‘serious metaphysics.’

168

Lynne Rudder Baker

This is precisely the view of philosophy that I reject. So, I believe, does Edmund Runggaldier. We think that serious metaphysics should help make sense of what we all know and need to know in order to get along in the world. The territory that interests me here contains things of familiar kinds – cats, keys, and credit cards, etc. I believe that they are irreducibly real. These things exist and cannot be reduced to things of other kinds, like sums of atoms. They belong in the ontology, the inventory of what exists. Any complete account of what there is must mention them as such – as cats, keys and credit cards. Ordinary objects are a diverse lot. They include any objects that can get lost or stolen, any objects that you can encounter or interact with. Almost everybody has to contend with dirty dishes and driver’s licenses. Hence, I’ve developed a view that allows these ordinary things, as such, a place at the ontological table. I use the term ‘ontological’ to signal that I am talking about reality, genuine reality with metaphysical heft; I’m not just talking about concepts or sentences that we accept as true. When I say that ordinary objects have ontological significance, I imply that if an inventory of what exists mentions electrons but not elephants and elevators, it is incomplete. Let me come clean about my own convictions. I believe that the things that we encounter and interact with – the parts of reality that include persons and their inventions – are no less ontologically significant than the microphysical parts of reality. Although philosophy can be technical and abstract, there should always be a thread that can be followed back to something that somebody might care about outside the “philosophy room.” That’s the pragmatic side of my approach. The ontological side stems from a Platonic conviction that the things that are valuable are real and not reducible to unfamiliar, microphysical entities. So, I want to present an ontological account of ordinary objects. My plan is this: After characterizing my ontological view of ordinary objects, I want to consider some objections that have been raised against it. Consideration of these objections will lead us to ask how we ought to evaluate a philosophical theory.

Ontology and Ordinary Objects

169

1. The Constitution View What is the relation between, say, the chair you are sitting on and the particles that make it up now? My answer is: constitution.1 Constitution is a relation of unity that holds between spatially coincident objects of different kinds: E.g., between pieces of paper and dollar bills, between the sum of cells in your body and your body, between a sum of threads and a tapestry – things of basically different kinds that are spatially coincident.2 Constitution, unlike identity, is time-bound. (But I’ll usually omit the time-index.) If a few threads come loose from the tapestry and are replaced, the tapestry will continue to exist but be constituted by a different sum of threads. Integral to my notion of constitution is the notion of primary kinds. Every material object is of some primary kind essentially: An object could not continue to exist without having its primary-kind property.3 Teacher is not a primary kind; you could cease to be a teacher and still exist. But person is a primary kind: You could not cease to be an person and continue to exist. God could not have created you without creating a person. Like person, barn and boat are primary kinds. Suppose that a barn is dismantled and the planks are used to build a boat. The object that was a barn no longer exists; a new object – a boat – exists and is constituted by the same sum of planks that previously constituted the barn. A thing’s primary kind determines what sorts of changes it can undergo and still exist and what sorts of changes will result in the thing’s ceasing to exist altogether – in

1

A number of prominent philosophers in recent years have endorsed some form of constitution-without-identity. The following are just a sample: Doepke 1982; Lowe 1983, Thomson 1998, Koslicki 2004; Yablo 1987; Rea 1998; Johnston 1992; Oderberg 1996; Sosa 1997; Burke 1994; Simons 1987; Wiggins 1968. My construal of constitution differs from those of all the above. 2 For a discussion of whether or not spatial coincidence, when joined with the causal efficacy of ordinary things, leads to intolerable causal overdetermination, see Chapter Five of Baker 2007. 3 To borrow some paraphrases about essential properties from Chisholm, if x has the property of being a horse essentially, then ‘x is such that, if it were not a horse, it would not exist’; or ‘God couldn’t have created x without making it such that it is a horse’; or ‘x is such that in every possible world in which it exists it is a horse.’ (Chisholm 1976, 25-6).

170

Lynne Rudder Baker

philosophical jargon, a thing’s primary kind determines its de re persistence conditions.4 The basic idea of constitution, then, is this: when things of certain primary kinds are in certain circumstances, a thing of another primary kind – a new thing, with different persistence conditions and different causal powers – comes to exist.5 When a yellow triangular piece of metal with the letters Y-I-E-L-D painted on it is in circumstances governed by laws and conventions of certain sorts, a new thing – a traffic sign – comes into existence. A traffic sign is a different kind of thing from, say, the triangular piece of scrap metal in your garage, even if your child had painted Y-I-E-LD on it for a school play. Although a traffic sign may be constituted by a piece of metal, a traffic sign is a different kind of thing, with different causal powers, from a piece of metal that doesn’t constitute anything. Primary kinds are not just “given” from the beginning of time. Some kinds of things, like human beings, evolve over time; and some, like MRI machines, are invented by human beings. But MRI machines and other human artifacts are fundamentally different kinds of things from natural objects (like turtles and trees), and the difference is that MRI machines have an intended function – a kind of essence – that turtles and trees and 4

My view thus follows Aristotle both in its essentialism and in its appeal to kinds. So far, I believe that Professor Runggaldier would agree with me. At least he would in 1998 when he published “Sortal Continuity of Material Things.” (Runggaldier 1998) However, my view of constitution differs from his in that I take constitution to be a relation between things of different kinds (e.g., pieces of metal and traffic signs), but he takes constitution to be a relation between a “objects proper” and some “compositional stuff...an object in a much looser sense.” I agree with Runggaldier that mereological sums are not (identical to) proper objects, but I think that they constitute proper objects. Nevertheless, I agree that where there is a statue made of copper, there are not two objects there. (Baker 2007, 169-172). 5 For detailed discussion, see Chapter Eight of Baker 2007. Consider a quantity of orange juice and a quantity of vodka. Call the orange juice and the vodka, considered together, the sum of the orange juice and the vodka. Now pour one into the other, and voilà, a new thing – a party drink, a screwdriver – comes into existence. It is not as if two things became one, but a sum of two things came to constitute one thing, a screwdriver. Now if you distil the screwdriver, you can separate the vodka and the orange juice back out; and the screwdriver goes out of existence. The sum of the vodka and orange juice existed before, during and after the party drink existed. Therefore, since identity is necessary, the sum of the vodka and orange juice are not identical to the party drink.

Ontology and Ordinary Objects

171

other kinds of natural objects do not have. Admittedly, something’s being an MRI machine depends on our practices and conventions. But people are part of nature and contribute to reality, just as molecules do.6 Ontologically – not just causally – some parts of reality depend on our activities and purposes. I call such objects ‘intention-dependent’ objects or ‘ID’ objects. (I am not at all sure the Professor Runggaldier would agree with me about ID objects.) An ID object could not exist in a world without other beings with beliefs, desires and intentions. Artifacts and artworks are paradigmatically ID objects, and they too bear constitution relations, just as natural objects do. The circumstances required for the existence of ID objects (like voting machines or statues) presuppose that there are people with intentions. The circumstances required for the existence of nonID objects (like turtles or trees) do not presuppose that there are people with intentions. But all objects – both those that presuppose intentionality and those that do not – all objects that we know of are constituted. Constitution is an engine of novelty. It brings into being new objects of new primary kinds. For example, a world with the same kinds of atoms that make up organisms (e.g., hydrogen, carbon, nitrogen, oxygen and sulfur)7 but with different environments may lack organisms, and a world without organisms is ontologically different from a world with organisms. A similar point can be made about inventions: the world before MRI machines or nuclear weapons was ontologically different from the world now. So, constitution makes an ontological difference.8 Although constitution is not identity, it is a relation of unity. This lectern is not identical to any sum of atoms; but it is constituted by a sum of atoms. Scrape off a few atoms and the lectern still exists. The constitution relation makes for a unity that is not identity. The constituted object (the lecturn) and the constituting object (a sum of atoms) share many exemplifications of properties by what I call ‘having properties derivatively’. The intuitive idea is simple: If x constitutes y at t, then some of x’s properties at 6

It is a profound error to take a distinction between what is mind-independent and what is mind-dependent as foundational for metaphysics. See Baker 2007, Chapter One. 7 Atoms of these kinds make up amino acid molecules. See http://www.biology. arizona.edu/biochemistry/problem_sets/aa/aa.html (accessed March 1, 2007).1 8 For greater detail, see Baker 2000 and Baker 2002.

172

Lynne Rudder Baker

t have their source in y, and some of y’s properties at t have their source in x. For example, my driver’s license is constituted by a piece of plastic: My driver’s license has the property of being rectangular only because it is constituted by something that could have been rectangular even if it had constituted nothing. And the piece of plastic has the property of easing the way through airport security only because it constitutes something that would have eased the way through airport security (a valid drivers’ license) no matter what constituted it. So, we have a drivers’ license constituted by a piece of plastic. The “source” of its property of easing the way through airport security is in the constituted license. The “source” of its property of being rectangular is in the constituting piece of plastic. I’ll say that the driver’s license has the property of easing the way through airport security nonderivatively, and the property of being rectangular derivatively; the piece of plastic that constitutes my driver’s license has the property of being rectangular nonderivatively, and of easing the way through airport security derivatively. I have defined these terms rigorously on the elsewhere.9 Not all properties may be had derivatively, but I’ll skip the details here.10 Enough properties can be had derivatively to see how property exemplifications may be shared and how constitution can be a relation of unity. Even primary-kind properties may be had derivatively. Something may have a primary-kind property without having that property as its primarykind property. There are two ways to fall under a primary-kind sortal: to be essentially of that kind or to be contingently related by constitution to something that is essentially of that kind. (Wasserman 2004). Since chair is 9

See Baker 2007, Chapter Six. Earlier versions appear in Baker 2000 and in Baker 2002. 10 Several classes of properties are excluded from being had derivatively: They are not shared. The excluded classes contain: (1) Properties expressed in English by locutions using ‘essentially’, ‘necessarily’, ‘possibly’, ‘primary kind’ and the the like (e.g., being a pencil essentially). (2) Properties of being constituted by x, or being identical to x (e.g., constituting a pencil). (3) Properties rooted outside the times at which they are had (e.g., having been quarried in 1500). (4) Certain combinations of properties (e.g., being a granite monument). I later define parthood – ordinary parthood – in terms of constitution as well as of mereology. So, the property of having part P at t is excluded since it is defined in terms of constitution. The property of having P as a part at t may not be had derivatively.

Ontology and Ordinary Objects

173

a primary kind, your chair is a chair essentially, but the sum of particles that make it up is only a chair contingently, in virtue of constituting the chair now. The sum of particles is not a chair unless there is something that is a chair essentially that it constitutes. If the chair is smashed and no longer exists, the sum of particles that once constituted the chair still exists, but no longer constitutes a chair. So, something (e.g., a sum of particles) may have a primary-kind property contingently when suitably related to something (e.g., a chair) that has it essentially.11 The chair is a chair nonderivatively, but the sum of particles is a chair derivatively – in virtue of constituting something that is a chair nonderivatively. The idea of having properties derivatively solves the problem of ‘too many Fs’ – too many persons, too many minds. If I am a person nonderivatively, and my body is a person derivatively (in virtue of constituting me), then there is just one person where I am – I nonderivatively, my body derivatively. I am a person nonderivatively; my body is a person derivatively, in virtue of constituting me. Since this body and I share a single exemplification of the property of being a person, my body and I are the same person. When this body becomes a corpse, it will still exist but no longer will it constitute a person. This body then will no longer be a person derivatively. This idea of unity can be sharpened by extending John Perry’s definition of ‘the same F’ to the constitution-relation. Constitution is like identity in some ways, and like separate existence in other ways. With respect to being the same F at t, constitution is similar to identity. Where I am, there is only one person – even though I am not identical to my body – because my body is a person derivatively.12 My body’s being a person is wholly parasitic on my being a person. 11

Many properties (unrelated to this discussion) may be had essentially by some things and nonessentially by other things. A planet has the property of having a closed orbit essentially; a comet that has a closed orbit has that property nonessentially. (This assumes that planets are planets essentially; otherwise it is only a de dicto necessity that planets have closed orbits.) 12 For an account of counting, based on Aristotle’s notion of accidental sameness, congenial to constitution-without-identity, see Brower and Rea 2005. Brower and Rea’s construal of constitution is significantly different from mine. They take constitution to be a mereological notion; I do not. I take sameness of parts at a time to follow from constitution, not to be constitutive of the idea of constitution itself.

174

Lynne Rudder Baker

Roughly, (omitting reference to times), x has F nonderivatively iff x’s having F does not depend on x’s constitution relations, and x has F derivatively iff x’s having F depends on x’s having constitution relations with something that has F nonderivatively. Now call the distinction between having a property nonderivatively and having it derivatively the ‘Key Distinction.’ The Key Distinction is coherent and clearly defined. (Baker 2007, 167-68) It shows how things in constitution relations share so many properties even though they are not identical, and hence it gives substance to the idea of unity-without-identity. 13 To sum up the Constitution View: There are three distinct ways that material objects, x and y, may be related at any time t: (i) x and y are identical, or (ii) x constitutes, or is constituted by, y at t, or (iii) x and y exist separately at t (i.e., they have different spatial locations at t). Although identity is necessary, nonidentity divides into two contingent relations – constitution and separate existence. Making room in metaphysics for two ways to be nonidentical is like making room in logic for two ways for a proposition to be untrue. Surprising, maybe; but not incoherent. 2. Objections and the Key Distinction The Key Distinction between having a property nonderivatively and having a property derivatively is a technical distinction that critics frequently 13

The Constitution View allows us to be nonreductionists and to take seriously the existence of all manner of things. To take seriously the existence of something, I believe, is to be able to distinguish between the thing’s changing a property and its going out of existence altogether. One reviewer of my new book misunderstood this point and complained about “Baker’s tendency to assume that, unless one believes Fs are essentially Fs, one does not take the existence of Fs seriously.” He pointed out that children are not essentially children, yet we take their existence seriously. Of course. Indeed, I made the analogous point about puppies. We can distinguish between the puppy’s losing a property (growing up) and going out of existence altogether (being run over by a truck); puppies have ontological significance in virtue of being dogs. In general, things have ontological significance in virtue of their primary kinds, of which they are essentially. We take puppies seriously in virtue of their being dogs essentially; we take children seriously in virtue of their being persons essentially. My view does not imply that we believe that puppies are essentially puppies or that children are essentially children in order to take puppies or children seriously. My point, which the reviewer missed, was that reductionists can’t distinguish between losing a property (like being a puppy) and going out of existence altogether.

Ontology and Ordinary Objects

175

either ignore or dismiss. One critic simply refuses to acknowledge the Key Distinction by taking ‘I am an animal derivatively’ to mean ‘I am not an animal.’ (Olson 2008) But this is like insisting that the sentence ‘this proposition is untrue’ means ‘this proposition is false’. In three-valued logic, a proposition can be untrue without being false; there is a third truthvalue – neither true nor false. Just as three-valued logic allows a proposition to fail to be false without failing to be untrue, so too on my view, an object may fail to have a property nonderivatively (i.e., independently of constitution relations) without failing to have the property altogether: the object it may have the property derivatively. There is a whole battery of objections – the many-minds objection, the counting objection, the too-many-thoughts objection (Olson, 2001; Zimmerman 2002), to name a few – that are all answerable in terms of the Key Distinction. Even though I believe that I have answered the objections, they keep popping up, Whac-a-Mole style. Here I want to try again. I’ll focus on a couple of examples. First: It is a consequence of my view that I am a person nonderivatively (indeed, essentially), and that this body is a person derivatively as long as it constitutes me. A recent critic has noted: “Whenever I think ‘I am not the body’, my body thinks ‘I am not the body’. But the body is mistaken. How do I know that I am not the one making this mistake?” The critic goes on to attribute to me – correctly – the view that “whenever my body says and thinks ‘I’, it does not refer to itself, but to me” (the person). (Johansson 2009, 368) But how, he goes on to ask, “can something be a person if it cannot perform the fairly elementary task of thinking and speaking about itself in the first person?” In fact, a human infant is a person who cannot perform the fairly elementary task of thinking and speaking about itself in the first person. But more to the point: This question simply ignores the Key Distinction between nonderivative and derivative persons. The referent of ‘I’ is never a body considered apart from a person. ‘I’ always refers unambiguously to a nonderivative person – a person constituted by, but not identical to, a body.14 As we saw in the extension of ‘the same F’ to the constitution14

If there is any ambiguity lurking, it is between the ‘is’ of identity and the ‘is’ of constitution. Since I am constituted by my body, but not identical to it, my utterance, “I am not my body,” is true if interpreted via the ‘is’ of identity, but false if interpreted via the ‘is’ of constitution.

176

Lynne Rudder Baker

relation, my body and I (though nonidentical) are the same person. So, the answer to the critic’s question, “How can something be a person if it cannot perform the fairly elementary task of thinking and speaking about itself in the first person?” is easy: I am a person nonderivatively (as long as I exist); my body is a person derivatively (as long as it constitutes me). There is no competition between my body and me. Since I am constituted by this body now, this body and I are the same person now. The word ‘I’ unambiguously refers to the entity who is a person nonderivatively – the person constituted by the body, but not to the body on its own. There is simply no mystery here. Second example: A different critic found it “absurd” for me to say that the body that constitutes me is a person derivatively. (Kearns 2009) This is supposed to be absurd because it leads to rejection of (what he calls) “a very plausible principle” – namely, that if I am a person and my body is a person and I am not identical to my body, then there are two persons. The critic’s “very plausible principle” generalizes to this: (VPP1) If x is a person and y is a person and x is not identical to y, then there are two persons. However, the use of the Key Distinction to extend the definition of ‘the same F’ to constitution-relations results in a modification of the critic’s “very plausible principle”: For x and y to be two persons, x and y must not be constitutionally related. As a result, the critic’s “very plausible principle” is modified to this: (VPP2) If x is a person and y is a person and x is not the same person as y, then there are two persons. (VPP2) is likewise very plausible. It is based not just on intuition, but also on a definition that does significant philosophical work in a theory. So, (VPP2) has a good claim to be preferable to (VPP1). Indeed, (VPP2) is an elaboration of Aristotle’s notion of numerical sameness without identity of being.15

15

For discussion of Aristotle’s notion of sameness without identity, see Matthews 1992. Matthews argues that Aristotle’s Theory of Kooky Objects should be interpreted as an ontological, not a linguistic, thesis.

Ontology and Ordinary Objects

177

So, both critics either ignored or dismissed the distinction between having a property nonderivatively and having a property derivatively, without raising a finger to argue against it. This leads me to ask: How should we evaluate a philosophical theory like mine? Should we base our verdicts on what strikes us as absurd or unattractive? 3. How To Evaluate a Philosophical Theory I don’t think so. Here, my pragmatic side comes to the fore. The proof of the pudding is in the eating. To evaluate a philosophical theory, a pragmatist like me would propose two questions: (i) Does the theory aim at a worthwhile goal? (ii) Does the theory accomplish what it set out to do as coherently and economically as possible? The goal at which my theory aims is an ontological account of ordinary objects that is comprehensive, unified and nonreductive. The theory is comprehensive in that it applies to every material object that we might encounter. It is unified in that it deploys the same two basic ideas – constitution and primary kinds – to understand all kinds of material objects. It is nonreductive in that things of one primary kind (a mallet, say) are not reducible to things of a different primary kind (such as two wooden cylinders of different dimensions). My goal of nonreduction may be challenged. Nonreduction is desirable because it allows for an ontologically robust account of things’ not just changing properties or being rearranged, but of going out of existence altogether. Suppose that you are a three-dimensionalist (like me), but that you have a reductive view according to which all medium-sized material objects were identical to (and hence reducible to) sums of particles. Now when the Twin Towers collapsed on 9/11, the sums of particles survived the collapse, but the towers did not. A three-dimensionalist reductionist cannot account for the difference between there being towers and there not being towers as an ontological difference. From a 3D-reductionist point of view, there is no ontological difference between particles-arrangedtowerwise and the same particles-arranged-rubblewise. Arrangements are not items in the ontology; they are not objects at all. Nor from a four-dimensionalist point of view, is there any ontological difference betweeen a 4D worm whose last temporal part has the towers

178

Lynne Rudder Baker

still standing, on the one hand, an a 4D worm whose last temporal part is just rubble, on the other hand.16 There is just a difference in the distribution of qualities. All 4D worms are ontologically on a par. We may pick out the worm that ends at the collapse of the tower, but that’s just a matter of our parochial interests. On a 4D view, there is no ontological content to a statement that a certain worm ends with the collapse of a tower. Appeal to composition does not help.17 Composition by particles does not distinguish between an ordinary object (like one of the towers) and something that is not an ordinary object (like the fusion of the bottom half of one of the Twin Towers and the top half of the other.)18 So, composition does not obviate the need for constitution. Unlike the reductive views, my view of constitution gives an ontological explanation of the towers’ going out of existence. Tower is a primary kind. The Twin Towers were towers essentially. So, when they collapsed and ceased to be towers, the objects that they were ceased to exist altogether. Since the constitution view can explain object’s beginning and ceasing to exist altogether, I believe that its goal of nonreduction is worthwhile. Does the Constitution View achieve this goal as economically as possible? Again, I would say yes. But for the Constitution View to achieve its goal, it needs the Key Distinction. To anyone who thinks that the Key Distinction is absurd, I would say this: The Key Distinction is coherent and earns its keep by contributing to a worthwhile goal that does not seem achievable without it. For a pragmatist, that’s good enough.

16

The difference between a 4D worm that is the tower-worm and a 4D worm that is the tower-and-rubble worm is no more an ontological difference than is the difference between the 4D worm that is the first-half-of-the tower’s existence and the 4D worm that is the second-half-of-the-tower’s existence. All four of these 4D worms are temporal parts of exactly the same sum of particles. 17 One critic suggested that we can skip constitution altogether and say something like this: “There were particles that composed the tower (and only the tower) before the attack on 9/11, and after the attack, those particles no longer composed the tower; they composed scattered objects or nothing at all.” (Kearns 2009, 534) But this is only an exercise in labeling; it makes no ontological distinction between worlds in which there are towers and worlds in which there is just rubble. 18 If we take the Twin Towers to be duplicates, and consider only qualitative properties of the particles and their relations to each other, the single tower and the two halftowers are compositionally indistinguishable.

Ontology and Ordinary Objects

179

4. Conclusion As I said at the beginning, I think that it is important for philosophy to keep in touch with things that we all care about. With that in mind, I have tried to develop an ontology of ordinary objects that we encounter and interact with – all day, every day. As the great American pragmatist Charles Sanders Peirce wisely urged, “Let us not pretend to doubt in philosophy what we do not doubt in our hearts.” (Peirce 1958, 40)19 I think that this is a sentiment with which Professor Runggaldier would agree. 5. References Baker, L.R. 2000 Persons and Bodies: A Constitution View, Cambridge: Cambridge University Press. -

2002 “On Making things Up: Constitution and its Critics”, Philosophical Topics: Identity 30, no. 1: 31-52.

-

2007 The Metaphysics of Everyday Life, Cambridge: Cambridge University Press.

Burke, M.B. 1994 “Preserving the Principle of One Object to a Place: A Novel Account of the Relations Among Objects, Sorts, Sortals and Persistence Conditions”, Philosophy and Phenomenological Research 54: 591-624. Chisholm, R.1976 Person and Object, LaSalle IL: Open Court Publishing Co. Doepke, F.C. 1982 “Spatially Coinciding Objects”, Ratio 24: 45- 60. Johansson, J. 2009 “Review of The Metaphysics of Everyday Life”, Philosophical Quarterly 59: 365-368. Johnston M. 1992 “Constitution is Not Identity”, Mind 101: 89-105. Kearns, S. 2009 “Review of The Metaphysics of Everyday Life”, Philosophical Review 118, no. 4: 533-535. Koslicki, K. 2004 “Constitution and Similarity”, Philosophical Studies 117: 327-364. Lowe, E.J. 1983 “Instantiation, Identity and Constitution”, Philosophical Studies 44: 45-59. Matthews, G.B. 1992 “Aristotle’s Theory of Kooky Objects”, unpublished manuscript. Olderberg, D. 1996 “Coincidence Under a Sortal”, Philosophical Review 105: 145171. 19

A predecessor of this paper was delivered at the Auburn Philosophy Conference on the Ontology of Ordinary Objects, February 27, 2010.

180

Lynne Rudder Baker

Olson, E.T. 2008 “Reply to Baker”, Abstracta Revista Special Issue I: 32-42. Peirce, C. S. 1958 “Some Consequences of the Four Incapacities”, in Philip P. Wiener, (ed.), Charles Sanders Peirce: Selected Writings, New York: Dover Publications: 39-72. Rea, M.C. 1998 “Sameness Without Identity: An Aristotelian Solution to the Problem of Material Constitution”, Ratio (new series) 11: 316-328. Runggaldier, E. 1998. “Sortal Continuity of Material Beings”, Erkenntnis 48: 359-369. Simons, P. 1987 Parts: A Study in Ontology, Oxford: Clarendon Press. Sosa, E. 1997 “Subjects Among Other Things” in M.C. Rea (ed.) Material Constitution, Lanham, MD: Rowman and Littlefield: 63-89. Thomson, J.J. 1998 “The Statue and the Clay”, Noûs 32: 149-173; Van Inwagen, P. 1990 Material Beings, Ithaca, NY: Cornell University Press. Wasserman, R. 2004 “The Constitution Question”, Noûs 38: 693-710. Wiggins, D. 1968 “On Being in the Same Place at the Same Time”, Philosophical Review 77: 90-95. Yablo, S. 1987 “Identity, Essence and Indiscernibility”, Philosophical Review 104: 293-314.

Can One Separate Me From My Life? A Runggaldier-sympathetic perspective on four- vs. three-dimensionalism CHRISTIAN TAPP, Bochum

1. The positions The debate between three- and four-dimensionalism is still alive in analytic ontology. It does not concern some special worries of some special metaphysicians, but a question of general importance, touching on the limits of scientific knowledge and the possibility of integrating it into one coherent world-view with our everyday ontology and the prerequisites necessary to explain agency. Four-dimensionalism comes in several variants. The two most important ones are the following:1 (Temporal Parts) Physical objects are four-dimensional parts of fourdimensional spacetime which have the form of spacetime ‘worms’ streched-out along the time axis. The most outstanding consequence of this thesis is that things are temporally extended and, thus, have temporal parts. Other consequences include that, at a given time t, objects are not wholly but only partially present. Their mode of persistence through time is perdurance, i.e., objects are identical to the series of stages of their existence. According to this understanding, four-dimensionalism is the denial of three-dimensionalism or the ‘no temporal parts doctrine’. (Reality of Non-Present Objects) There is no ontological difference between past and present objects. They both exist in the same sense. With respect to the status of future objects, one can distinguish eternalism, which claims that past, present, and future objects exist on a par, from the 1

Both meanings of 4-dimensionalism should be kept separately; see (Rea 2003, 246– 247), who understands 4-dimensionalism primarily as a thesis about the ontological status of non-present objects (as opposed to presentism) and proposes to use ‘perdurantism’ for the denial of three-dimensionalism. For the sake of this paper, I will stick to the widespread way of talking in calling perdurantism also (a variant of) ‘4dimensionalism’.

182

Christian Tapp

growing block theory which claims the existence of past and present objects only. According to eternalism, spacetime has no distinguished hyperplane orthogonal to the time axis which could be called ‘the present’, and hence there are no tensed facts in the strong sense of a McTaggart A-series of time. The opposite thesis to both variants, eternalism and growing block theory, is presentism, i.e., the claim that only presently existing objects really exist. In this paper, I will focus chiefly on the first variant of fourdimensionalism as opposed to three-dimensionalism. Four-dimensionalists hold that ordinary things have temporal parts, while three-dimensionalists deny that. In their eyes, things (not only a temporal part of them) are wholly present at each moment of time at which they exist. Things are extended in three spatial dimensions, but not in time. So, three-dimensionalism and four-dimensionalism disagree about the identity of temporally persisting objects. Following David Lewis, we can say that an object persists iff it exists at various times. Four-dimensionalism holds that a physical object persists through time by having different temporal parts at different times, ‘though no one part of it is wholly present at more than one time’ (Lewis 1986, 202). This mode of persistence through time is called ‘perdurance’. According to three-dimensionalim, in contrast, physical objects ‘endure’, that is, they are ‘wholly present at more than one time’. Edmund Runggaldier holds the attractive thesis that four- and threedimensionalism are in a way compatible (Runggaldier 2010, 28). According to him, four-dimensionalist models of reality can be successful in the sciences without forcing us to accept them as adequate pictures of reality in itself (Runggaldier 2007a, 2007b). I propose to extend this into the thesis that, in an integrative metaphysical perspective in the spirit of Muck (Löffler 2010), one cannot even separate the two theories. My argument for their intricate interrelation is grounded on the assumption that one cannot separate me from my life – although we are surely not identical. 2. The battle In this section, I want to discuss some (not all) of the prominent arguments on three- vs. four-dimensionalism. To begin with, three-dimensionalists find the idea of temporal parts of natural objects in our Lebenswelt strange. They resort to our natural ways of speaking according to which we say that physical objects exist at some

Can One Separate Me From My Life?

183

times t1 and t2, but not that temporal parts of them exist at those times. If I have changed between t1 and t2, then there must be differences between me at t1 and me at t2, although “I” must somehow refer to the same person about which one says that it had changed from t1 to t2. Hence there is a substantial question for identity conditions. The whole point of personal identity is to theoretically combine a change in properties or relations with the presupposition that it is nevertheless one and the same subject (me) who has undergone that change and remained identical to itself during it. To illustrate this, imagine me at the age of 2 and my son having now grown to this age. He resembles me as a 2-year old much more than I as grown up today could ever resemble myself as a 2-year old boy. But nevertheless, the point of personal identity is that I (today) am me (some 30 years ago), and not my son (today). Another point in favour of three-dimensionalism concerns the explanation of our actions and the articulation of the beliefs relevant to actions (Runggaldier 2007a, 2007b). This includes the phenomenon of indexicality. What a person refers to by using expressions like ‘here’, ‘now’, or ‘I’ depends on when and where she utters a statement containing such indexical expressions. Four-dimensionalism cannot, it is said, take this point into account. One may call it ‘subject-centeredness’ or ‘prevalence of a firstperson-perspective over a third-person-perspective’. A four-dimensionalist’s semantics has to replace all occurences of, say, ‘I’ and ‘today’ in an utterance by Edmund Runggaldier on 24 August 2011 by non-indexical expressions like ‘Edmund Runggaldier’ and ‘on 24 August 2011’. This admittedly eliminates some philosophical puzzles about indexicals. But it also obstructs the way to access important aspects of the first-person-perspective in the explanation of actions – like our reasons guiding them, the viewpoint ‘from somewhere’, the way it feels to act, the presupposition of being the same person who will enjoy (or suffer) the consequences of that action etc. The ascription of moral and legal responsibility points in the same direction. If I did something wrong yesterday, it is me who can be blamed for it today. Obviously, referring to the fact that it was some other part of me who lapsed, but not the same part one is actually talking to, is never a good excuse. However, this is not what a four-dimensionalist is committed to. He can reconstruct the relevant responsibility ascription via a little detour, starting from today’s part of me (=whom one is talking to) to the fourdimensional whole it is part of (=me), to yesterday’s part of this whole

184

Christian Tapp

(=who lapsed). So, the critique shrinks to charging four-dimensionalism such a detour in reconstructing a morally and legally basic mechanism. In contrast, some philosophers have argued that the natural sciences strongly suggest a four-dimensional ontology. One of the greatest advocates of four-dimensionalism was Willard Van Orman Quine (see for example Quine 1960, §36; 1981, ch. 2). The doctrine of ontological naturalism instructs us to consider science as the authoritative source for learning what there is and what the basic structures of reality are. In a naturalistic mindset, therefore, the success of the scientific four-dimensional picture of the natural world is pressing towards a general metaphysical claim that things are indeed four-dimensional. Sometimes an argument from Leibniz’ principle of indiscernibilitas identicorum to four-dimensionalism is brought forward. According to Leibniz’ principle, if a and b are identical, they must share each and every property. In a non-temporal version, this principle seems to contradict the combination of change and identity through time. It does not allow that a has P and non-P (at different times). As opposed to enduring objects, perduring objects do not suffer the same problem, for what is usually predicated of one object is now predicated of two stages or phases of an object. This argument, however, is not cogent, since instead of integrating time into the objects one could also read Leibniz’ principle in a temporal way, for example: If a and b are identical, then for every time t, a and b must share each and every property they have at t. This temporal version well allows for a having P and non-P at different times.2 Metaphysical four-dimensionalism, on the other hand, brings with it some problems when it comes to reconstructing our everyday ontology or the advanced ontology required for a philosophical analysis of agency. Some of the problems have been discussed above. As far as I can see, here lies the reason why so many philosophers strongly resist fourdimensionalism although they are interested in a scientific picture of the world and an integrative understanding of our lives including scientific 2

It may be interesting to note that Bernard Bolzano builds this point right into his definition of time. According to him, time is a determination which we have to add to the idea of a substance in order to find among each predicate P and its contradictory non-P one which can truthfully be ascribed to the substance while the other is denied of it (see Paradoxes of the Infinite, § 39).

Can One Separate Me From My Life?

185

elements. To be more precise: many of them resist the amalgamation of four-dimensionalism and ontological naturalism, not four-dimensionalism per se. If it were correct that four-dimensionalism was indispensable for natural sciences, the counter-position would necessarily be hostile to the scientific endeavour. Runggaldier, like many other proponents of three-dimensionalism, is by no means an ‘enemy of science’. Instead he sympathizes with an analysis of scientific reasoning by Muck, who holds that the success of scientific reasoning is founded on its distance to subjectivity and its restriction of perspective and method (Runggaldier 2007a, 2010). Scientific claims therefore can be true and successful, but insofar as they are won by sort of parenthesizing parts of reality, they cannot claim unrestricted universal validity. If one makes use of effective models one must not forget that they are, in fact, models that reproduce only part of reality. Hence, four-dimensionalism is all right for the regional ontology of the sciences, while this does not force us to accept its counterpart in fundamental or general ontology. Even if four-dimensionalism is an indispensable ingredient of the scientific quest of knowledge, we need not embrace it as a thesis about how things really are. This is due to the fact that the scientific quest makes up only part of our general quest for understanding ourselves and the world. Parenthesizing some parts of reality surely is a good starting point for the generation of scientific knowledge, but it is a bad starting point for an argument against the existence of the parenthesized parts. So, Runggaldier favours a three-dimensionalist theory which better allows for the explication of indexical speech acts, human agency, and the linguistic inconvenience of talking about temporal parts of me. Let me thicken his argument slightly: When by ‘I’ I refer to me, then in statements about what I did yesterday, do today, and will be doing tomorrow, I do not refer to other parts of me, but to myself in different circumstances. Temporal circumstances, however, resemble much more the places at which I can be than the places of my spatial parts. Paradoxically, just this resemblance between time and space seems to me to be one of the best arguments in favour of three-dimensionalism.

186

Christian Tapp

3. The pacification (attempt) In our everyday language we indexically refer to ourselves and explain the actions of persons in ways that seem to fit much better into a three- than a four-dimensionalist ontology. We say that you celebrating your 65th birthday are the same person as you when you were born – otherwise there would be no birthday to be celebrated (or uncountably many birthdays, which is likewise unsatisfactory). Hence, in our everyday metaphysics people are conceived of as enduring things rather than as spacetime worms. Yet, there is something corresponding quite well to the spacetime worms of four-dimensionalism, namely our lives (or our personal history). Whilst the three-dimensionalist has it that I am a three-dimensional object enduring through time, he will probably agree that my life is not such a threedimensional object enduring through time. My life, instead, is extended in time. It is more of a process than of an object. And it has temporal parts: my birth, my childhood, my studentship etc. Furthermore, it is quite plausible to conceive of my life as having also a spatial extension at a given time, namely my (living) body: there is no living of me outside of my body and no part of my body not belonging to the respective moment of my life.3 If this is true, then the difference between ontological threedimensionalism and four-dimensionalism seems to reduce to the question of what the personal pronoun ‘I’ refers to: me or my life? The threedimensionalist holds that ‘I’ refers to me, that I do not have temporal parts, while my life has such parts. The four-dimensionalist, in contrast, maintains that ‘I’ refers to that which the three-dimensionalist calls ‘my life’, 3

This thesis needs further discussion and, maybe, modification. Does my life extend to the tumor which has just been excised from my stomach? One is inclined to say no, in perfect harmony with saying that the tumor is not part of my body anymore. Though, does my life extend to the finger which I had lost in an motor saw accident and which was successfully sewed-on again? And what about the finger during the time of its detachment? If one wants to hold that my life extends also to biologically living parts which are temporally detached from the rest of my body, the thesis would require us to call detached (ex-)parts of my body parts of it. But this may lead into aporia in case a former part remains detached. So, by modus tollens, the thesis would require us to say that although a detached part of my body is biologically living and the detachment is only for a limited period of time, my life is not extending to that part. If this consequence turns out too uncomfortable, one had to modify the thesis that my body is the spatial extension of my life.

Can One Separate Me From My Life?

187

and that what the three-dimensionalist calls ‘me’ is just a temporal part of what the four-dimensionalist calls ‘me’, namely, a slice through the worm orthogonal to the time axis. Storrs McCall and Jonathan Lowe (2006) have proposed that even the notion of a temporal part can be reconstructed in three-dimensionalist terms. As physics tells us, continuously existing physical objects are made up of changing sets of particles (atoms or subatomic particles – whatever one likes here). For any continuously existing physical object O we can, therefore, consider as the ontological sum of the particles which constitute O at time t. so defined is readily available for a threedimensionalist for it consists of the particles which endure through time. however corresponds precisely to that which the four-dimensionalist calls a ‘temporal part’. Hence, there is a translation of statements (in a standard version) of three-dimensionalism into statements (in a standard version) of four-dimensionalim. McCall and Lowe conclude that the threevs. four-dimensionalist controversy is a ‘storm in a teacup’ (McCall&Lowe 2006, 577). What, then, is the actual difference between (the standard accounts of) three- and four-dimensionalism? It seems to lie in the difference between me and my life. But can they be separated? Am I something different from my life? Is my life something I could be removed from? Of course, nonremovability is not sufficient for identity. Anyway, I would suggest that just as Aquinas once said ‘anima mea non est ego’ (Super I Cor. c.15 l.2), we can also say ‘vita mea non est ego’. Bluntly identifying me with my life would indeed be a category mistake (Lowe 2009, 51). Do I have to blame the four-dimensionalist for a blunt category mistake then, as Lowe suggests? In my opinion, a reaction in Muckian style is much more attractive. As we have seen above, Muck holds that science has to parenthesize part of reality and that this is what makes science so successful in describing nature, finding laws, and explaining natural phenomena (Runggaldier 2007a, 89–90). For a four-dimensional spacetime perspective, I am not available in ways other than (synchronically) a spatial object with certain material constitution and (diachronically) an object with certain duration of life during which my synchronic constitution undergoes change. The physicalist perspective results in an abstraction, a limited picture of reality, like a photograph which falls short of reality not only in ommitting two dimensions, but also sound, smell, touchability, temperature, air pressure, color fastness etc. Hence, the four-dimensionalist is per-

188

Christian Tapp

haps just speaking a little loosely when saying that I am the spatiotemporal worm of my life. He should better say that I am not available to the physical spacetime model of reality in other ways. Here we touch the large-scale discussion about reductionism and the critique on the ‘nothingelse-buttery’ (Julian Huxley, as reported by Lorenz 1983, 196f.). Even if I am not identical to my life, we cannot be separated either. We ontologically depend on each other in a very strong sense. This was also Aquinas’ position. According to him, ‘being alive’ is a substantial or essential predicate for living things (STh p.I q.18 a.2). It is by no means accidental. If a living thing is no longer alive, it is no longer identical to itself, it would not exist anymore. ‘Vivere nihil aliud est quam esse in tali natura’ – in this nature (the nature of living things, of that what has the inbuilt capacity to move itself) to live is nothing else than to be. Hence, according to Aquinas, the life of the living is its being. My life and my being are identical. In contrast to God, I am not my life or my being, but the fact of my existence is the same fact as that of me living. Put in a nutshell: In the four-dimensionalist spacetime perspective, I show up only in form of my life. My life is only part of what makes me up. Hence, the four-dimensionalist account of ‘me’ need not necessarily be condemned as a category mistake. Instead it can be taken as the result of a restricted perspective. I take this to show that four-dimensionalism in the sciences and three-dimensionalism in ontology are in fact more than compatible. They are strongly connected to each other. 4. Concluding remarks Let me conclude with some methodological observations. Obviously, I was not able to hide my sympathy to many of the standpoints of Edmund Runggaldier with respect to three- vs. four-dimensionalism. We shall, however, give room for a very general critical question here: How do we know about all that? How do we know that, even though the four-dimensional model in the sciences is so successful, things truly are different? What those questions ask may be called ‘sources of metaphysical knowledge’. Which sources do we have? In the works of Edmund Runggaldier, one can find some hints at sources of metaphysical knowledge (explicit or implicit): everyday ontology as built up during practical life in the Lebenswelt; explanatory power of agency in describing and understanding intentions (first-person perspective);

Can One Separate Me From My Life?

189

power to integrate different regional subtheories; rational deliberation; and, last but not least, scientific success. One may think about amending this list by adding basic truths like Moorean truths (part of Lebenswelt) and transcendental conditions of experience (Kant). And perhaps one could spend a thought about other possible sources like intuition, introspection, and revelation. All these sources considered (except, perhaps, the last block of three), do we then have enough evidence to judge how things really are? Runggaldier’s critique on the inference from scientific success to the truth of four-dimensionalist ontology teaches us a different lesson: Maybe not only science, but also metaphysics aims at models of reality. For they can come more or less close to reality, metaphysics should be freed from the overload of pretensions of certainty. I find the idea convincing that there is a methodological convergence between philosophy of science and metaphysics, not only with respect to integrative explanations (Löffler 2010), but also with respect to a model-based distance from excessive claims of truth and certainty. Science and metaphysics produce theories which are better or worse, depending on criteria such as coherence, extension, compatibility (to everyday metaphysics and the sciences) etc. And, if a last word of speculation is granted to me, critical realism may be a very promising metametaphysical doctrine. It would mean on the one hand that metaphysics is about reality so that theories can objectively be better or worse than others while this quality does not depend only on the purely internal properties of the theories. On the other hand, it would mean that metaphysics is aimed at models which are continuously in need of improvement, be it by critique or, rarely, by revision. Bibliography Haslanger, S. 2003: Persistence through Time. In: The Oxford Handbook of Metaphysics. Ed. by Michael J. Loux and Dean W. Zimmerman. Oxford: Oxford University Press, 315–354. Hempel, C.G. 1974: Grundzüge der Begriffsbildung in der empirischen Wissenschaft. Düsseldorf: Bertelsmann Universitätsverlag. Löffler, W. 2010: Integrative Erklärungen: Konvergierende Tendenzen zwischen Metaphysik und Wissenschaftstheorie? In: Metaphysische Integration. Essays zur Philosophie von Otto Muck. Ed. by Winfried Löffler, Heusenstamm: Ontos, 91–111.

190

Christian Tapp

Lorenz, K. 1983: Der Abbau des Menschlichen. München: Piper. Lowe, E.J. 2009: A Survey of Metaphysics. Oxford: Oxford University Press 2002, reprint 2009. McCall, S. and Lowe, E.J. 2006: The 3D/4D Controversy: A Storm in a Teacup. In: Noûs 40/3, 570–578. Quine, W.V.O. 1960: Word and Object. Cambridge: MIT Press. -

1981: Theories and Things. Cambridge: Belknap Press.

Rea, M.C. 2003: Four Dimensionalism. In: The Oxford Handbook of Metaphysics. Ed. by Michael J. Loux and Dean W. Zimmerman. Oxford: Oxford University Press, 246–280. Runggaldier, E. 2007a: Das menschliche Lebewesen als Einheit in der Zeit (Kontinuant), in: Naturalismus als Paradigma. Ed. by L. Honnefelder, M. Schmidt. Berlin: University Press, 87–104. -

2007b: Die Raum-Zeit: Eine naturalistisch verengte Sichtweise, in: Philosophisches Jahrbuch 114, 399–408.

-

2010: Eingeschränkte Sichtweisen: Mucks Antwort auf die naturalistische Herausforderung. In: Metaphysische Integration. Essays zur Philosophie von Otto Muck. Ed. by Winfried Löffler, Heusenstamm: Ontos, 13–29.

Dimensions of the Mental FRANZ VON KUTSCHERA, Regensburg

Today’s philosophy of mind is less concerned with exploring the mind itself, the varieties of mental phenomena and their interrelations than with demonstrating that the mind is nothing else but the product of neural processes. The optimism inspiring the work on this reductionistic program, however, derives mostly from a fundamental lack of knowledge of the mental. A serious philosophy of mind should first investigate mental phenomena before it enters into a discussion of their reducibility to something else. In my paper I want to justify this claim by showing that simple antimaterialist arguments can be based on fundamental structures of the mental. The first section is about intentionality and reflection. The second shows how reflections gives rise to the hierarchies of abstract objects. Since these hierarchies demonstrate the dimensions of the human mind only from a conceptualistic standpoint which sees them as mental constructs, I shall discuss the problem of the ontological status of abstract objects in the third section. In the last section I formulate two anti-materialistic arguments relying on these hierarchies 1. Intentionality and Reflection Our consciousness is mostly intentional: a consciousness of an objective reality. For our purposes we can think of the objective reality as a state of affairs. I see (know, doubt, wish etc.) e. g., that the apple tree in my garden begins to blossom. The words “objective” and “object” have to be taken in a very wide sense here so that it applies to everything our attention is directed at. The state of affairs we are conscious of may belong to the external world as in our example or to the mental life of another person, but it may also concern my own mental life. I may be conscious, for instance, of the fact that I see, that the apple tree is blooming. Normally our attention is focussed on the objective reality while we ourselves and our acts and attitudes remain in the background. We see the world as in the light of a torch that leaves the person who holds it in the dark. We are also implicitly

192

Franz von Kutschera

aware of ourselves and what we do, but if we want to be clear about that, we have to reflect on ourselves or our attitudes or acts. “To reflect on something” means in general to think about it. More specifically, to reflect is to turn the attention from objects of the outer world to the subject or its activities and states, to its perceiving, knowing, doubting or wishing. Still more specifically, reflection consists in making precedent mental acts and states the topic of new ones. Therefore a reflection can be described as leading from one act or state to another intentional act that has the first one as its object. I listen to a symphony, for instance, and then realize that my thoughts have wandered and I have taken in only very little of the music. I make a claim and then consider whether it was justified. I feel unhappy and reflect on the reasons. In all such cases to reflect means to disengage from what we are doing, feeling or wanting, to take a step back, bring it into focus and think about it. The capacity for reflection is a fundamental trait of the human intellect. Its theoretical and practical importance is hard to overestimate. If we do something, want something or feel something we are more or less aware of it. To get a clear idea of it, however, we have to reflect on it and concentrate our attention on ourselves. The capacity of reflection, then, is a precondition for forming a clear and distinct notion of ourselves and for a critical evaluation of our own actions. A scientist constantly has to reconsider the justification of his assertions and his theories. In everyday life we have to examine our own actions as to their expediency or to their moral permissibility. In order to control it we have to reflect on what we say, feel and want. Our ability for reflection depends largely on language. Language is not only a means for the expression and communication of thoughts but thinking and speaking are inseparable and are developed in close correspondence. This has been emphasized, first of all, by Wilhelm von Humboldt.1 He pointed out that linguistic articulation is essential for the distinctness of thought. There are two reasons for that: First, expressing a thought linguistically means giving it an objective form in which it can be perceived optically or, in spoken language, acoustically. Now, our cognitive competences are best developed for comprehending the outer world. Only in sensual perception are we confronted with well-defined objects, only by it can we 1

Humboldt 1960, vol. VII, p. 53.

Dimensions of the Mental

193

make fine and accurate distinctions. In the field of mental phenomena they are much more difficult. Linguistic expression, then, allows additional distinctions for whatever is expressed by correlation with sensible differences. It also makes reflection on them much easier. You cannot distinguish pure emotions or impressions by putting one besides the other and comparing them for, roughly speaking, the mental realm admits of no simultaneity but only of succession . If you describe them linguistically, however, you can compare the descriptions and analyze them even if the emotions and impressions themselves have passed away. Secondly, the linguistic expression of mental states and activities makes it possible to talk about them. We live in a common world, and therefore agreement with others is a criterion for reality. The expression of our opinions opens the way to additional, intersubjective criteria for their correctness and saves us from having to follow just our own impressions. In one sense the development of a common language presupposes common experiences, of course, but in another sense common experience presupposes a common language since it is only by it that we can determine what contents of experience we share. 2. Abstract objects It is by reflexion that abstract objects are formed. Besides the realm of the physical and of the psychological phenomena there is also the realm of ideal objects, “der Bereich des objektiven Nichtwirklichen”, as Gottlob Frege called it, i. e. the realm of causally ineffective objective being. That is also Karl Popper’s “third world”. This world contains theories, norms and ideals and the like, but here we shall be concerned with abstract objects only, with concepts, propositions and sets, especially. A concept, first of all, is a way to conceive an object, to determine what it is or distinguish it from others. It is only by reflecting on this determination that we come to concepts as abstract objects and topics for further determinations or distinctions. Likewise for propositions: We first determine an object conceptually and then we reflect on this act. But the determination of an object presented as an object is a proposition. Reflection, then, is the way to the infinities of the world of abstract objects.

194

Franz von Kutschera

a. Concepts: Let us take number theoretic concepts like ‘is a prime number’ or ‘is smaller than’ as an example. They are defined on the set U0 of natural numbers. Since natural numbers themselves are not concepts such number theoretic concepts are termed “first-order concepts”. If we now consider such first-order concepts as new objects and add them to U0 , thereby going from U0 to a more inclusive set U1, we can also define second-order concepts on this new set like ‘is a transitive relation’ or ‘is an instance of the property ...’. We can then repeat this step and count also second-order concepts among our objects, arriving at a still larger set U2. On it we may then also define third-order concepts like ‘applying to all second-order concepts’, and so on. In this way we arrive at a hierarchy of concepts and their domains as it is envisaged for instance in (cumulative) type theory. This hierarchy of concepts remains open since we can always enlarge our last universe of discourse by adding the concepts we have used to characterize its objects. The set of all concepts does not exist, therefore. b. Propositions: There is also an infinity of propositions. First-order propositions are the ones we can form without presupposing other propositions. An example is the proposition that Cairo has approximately 18 million inhabitants. Second-order propositions may presuppose only such first-order ones, as for instance the proposition that the proposition, that Cairo has approximately 18 million inhabitants, is true or is about Cairo. In this way we again arrive at an infinite and open hierarchy of propositions. c. Sets: My third example for a mental infinity are sets. Since set theory is the fundamental mathematical discipline it is here that the infinities have been studied most extensively. If we start from the classical concept a set is the extension of a property. The set of frogs, e.g., is the extension of the property of being a frog and contains as elements exactly those things that are instances of this property, i.e. frogs. This conception of set has also been that of the founders of set theory, Gottlob Frege and Georg Cantor. Their as hindsight has it - „naive“ set theory rests on just two simple principles that derive directly from this concept. The first of them is the axiom of comprehension: ‘Every property has an extension’, i.e. for every property

Dimensions of the Mental

195

there exists a set of all its instances. The second principle is the axiom of extensionality: ‘Sets that have exactly the same elements are identical.’ Together they imply that every property has exactly one extension, that it determines exactly one set of its instances. If we start from objects that are no sets – mostly called individuals – we can form sets of them, let us call them first-order sets. Of them plus the individuals we may then also form second-order sets, i.e. sets containing only individuals and first-order sets, and so on. In this way we again get an open hierarchy of sets. In the hierarchies of abstract objects the infinity of the mental become especially clear. They represent our ability to create new objects, concepts and propositions. The essential incompleteness of these hierarchies mirrors the creativity of the human mind. Therefore one should not ask, what set theory as a mathematical theory has to do with the philosophy of mind: Set theory is the most developed theory of mental constructions and therefore deserves much more attention from the philosophy of mind than it has received so far. 3. Conceptualism In presenting the infinite hierarchies of concepts, propositions and sets I have tacitly presupposed a conceptualistic understanding of them for which abstract objects are constructions of the human mind. The opposite stand is that of Platonism or realism with respect to universals. For it abstract objects are not mental constructs but exist independently of human thinking; they are, as Frege said, not created, but only grasped by us. The strongest argument for conceptualism is that its rival, Platonism, leads to contradictions. Take concepts first again. If they exist ready-made out there, there must be the set of all concepts. But then there is also the set of all concepts or properties not applying to themselves. Such a property is being a frog, e.g., which being a concept is not a frog. But if we ask whether the property of not being self-applying is self-applying or not we get a contradiction because it is evidently self-applying if and only if it is not self-applying. In a conceptualistic understanding, however, concepts are something we form, and as we form concepts with respect to given objects every concept presupposes its domain of definition and is therefore never defined for itself. There is then no set of all properties and no over-all property of being not self-applying. There are only the properties of being a not self-applying

196

Franz von Kutschera

property of an order smaller than n, for all n, and these properties are of order n. The same holds for propositions. A Platonistic understanding of them has to face the paradox of self-applying propositions that Alfred Tarski has formulated for sentences:2 A universal proposition of the form ‘All propositions have property E’ - for instance: ‘All propositions asserted by Max are wrong’ - is called self-applying if and only if it has the property E itself - in our example: if the proposition is either not asserted by Max or is false. Now, is the proposition that all propositions are not self-applying selfapplying or not? Again we find: it is self-applying if and only if it is not self-applying. This construction corresponds closely to that in the paradox of the concept of not self-applying concepts. In the case of propositions we need only the Platonistic assumption that propositions are not something we form but something that exists independently of human thought. Then there is the set of all universal propositions, there are self-applying and not self-applying ones among them and the proposition that all universal propositions are not self-applying makes sense and must be contained in the propositional universe. For conceptualism, however, the construction falls flat: If we have a set P1 of propositions we can make assertions about the propositions in P1 and form the corresponding propositions. Since the formation of such propositions presupposes those in P1 they cannot belong to P1 themselves. If we add these propositions to P1 and thereby arrive at a larger set P2 the sense of the word „all“ is changed if applied to the elements of P2 instead of to those of P1. The proposition „All propositions in P1 are not self-applying“ is not in P1 therefore. It refers not to itself, and thus the contradiction evaporates. The problems of Platonism have been discussed most extensively for sets, for it is there that they can be represented in the most simple way. The two principles of set theory I have named above, although they are immediate consequences of the concept of a set as the extension of a property, give rise to a host of contradictions, the paradoxes of naive set theory. The most simple one is that of Russell: There are sets that are not elements of themselves as our set of frogs testifies. The Russell set - let us call it R - is the set of all such sets, i.e. the set of all sets that are not elements of themselves. Now, is R an element of itself or not? It contains itself as an element 2

Tarski 1949, p. 80, fn. 11.

Dimensions of the Mental

197

if and only if the defining property of not containing itself applies to it, i.e. if it does not contain itself, and that again is a contradiction. In fact Russell’s paradox follows from the principle of comprehension alone. It has baffled generations of mathematicians where the mistake in that simple and self-evident principle could be. To see that one has to realize that the construction of the paradoxes presupposes the existence of the set of all sets, the distinctive principle of Platonism. The founders of set theory, Cantor and Frege, were both staunch Platonists, as is the majority of contemporary mathematicians. To them the metaphysic reality and eternity of abstract objects seemed overwhelmingly plausible and therefore also the existence of a closed universe of sets. Platonism as to sets, however, is really not very convincing, if you come to think of it: According to the classical notion a set is an extension of a concept. Now concepts are forms of conceiving which we form, which come into use and pass out of it. They are therefore naturally thought of as mental constructs. But then this conceptualism which is so natural for concepts should also be extended to sets as their extensions. And the same holds for propositions: Propositions are formed by applying concepts to objects, and if the concepts are something mental so are the propositions. These brief remarks cannot, of course, serve as a refutation of Platonism. The main argument is simply the fact that there are no intuitively convincing and consistent Platonistic theories but intuitively well-founded and consistent conceptualistic alternatives.3 In a conceptualistic theory concepts and sets form an open hierarchy the elements of which are all grounded. In the case of sets this means that there are no infinite elementchains. Therefore no set can be an element of itself. Russell’s set does not exist, and neither the universal set which would have to be an element of itself. 4. Anti-materialistic Arguments a. There is no set of all mental propositions: The infinity of mental constructions is the basis of the most effective antimaterialistic arguments. Materialism maintains that the mental is reducible 3

Cf. Boolos 1971, and for another version of a conceptualistic set theory Kutschera 2001.

198

Franz von Kutschera

to the physical but is divided as to the sense in which such a reducibility should hold. The different concepts of reducibility can best be compared if they are expressed by supervenience relations. The weakest of the relevant supervenience relations is that of global supervenience. A set of propositions S is globally supervenient on a set of propositions T if any two worlds differing with respect to the S-propositions holding in them also differ with respect to the T-propositions holding in them. If S is the set of all mental propositions and T that of all physical propositions we get the claim of a global supervenience of the mental on the physical. The first argument against materialism can then simply be stated thus: There is no set of all mental propositions, as we have seen. We can, of course, define the supervenience relation without mentioning sets of propositions and say simply that if two worlds differ with respect to some mental propositions, they also differ with respect to some physical propositions. This reformulation doesn’t eliminate the problem, however: The words “all” and “some” are meaningful only with respect to a well-defined domain, and there is no well-defined domain of all mental propositions. b. The argument from dimension: Let T be a set of states of affairs, complete in the sense that together with a state also its negation is in T, and together with a sequence of states also their conjunction.4 Consider now the possible doxastic attitudes of a person P to the states of T, i.e. her beliefs that exactly such and such states of T hold. Such an attitude may be represented by the subset of T containing exactly those states of T P believes to obtain. The set A of possible doxastic attitudes of P to the T-states is then the power set of T, i.e. the set of all the subsets of T, and the set of possible beliefs of P concerning states of T may be represented by the power set of A.5 That P believes that a proposition p 4

S, then, is to be a complete Boolean algebra. In doxastic logic which studies the principles of belief - cf. Lenzen 1980, e.g. - we distinguish between a descriptive and a rational concept of belief. While for the latter there are logical principles saying, e.g., that you cannot at the same time believe a state and its negation, and that together with a state you also believe what follows logically from it, there are no such principles for the first one, to which I refer in the text. Because of the principles for rational belief doxastic attitudes may be represented by a single state instead of by a set of states. In the case of descriptive believes there is no such possibility.

5

Dimensions of the Mental

199

from T holds means, after all, that P has some doxastic attitude represented by a subset of T containing p. Now Georg Cantor has shown that the power set of a set always contains more elements than that set itself. Therefore there are more doxastic attitudes towards the states of a set than there are states in that set. Applied to physical states: There are more possible beliefs about physical states than physical states. Thus not all mental states can possibly be physical states. This is an argument against an analytical global supervenience of the mental on the physical. But there is still a weaker thesis, that of a nomological global supervenience. It is defined in the same way as the analytical relation but only with respect to nomologically possible worlds, i.e. worlds in which the same physical laws hold as in the real world. Now it is doubtful whether such a claim can still be classified as materialist since it can also be made by dualists. Psycho-physical parallelism as represented, e.g., by Arnold Geulincx, Nicole Malebranche and, most prominently, Gottfried Wilhelm Leibniz, makes it too. To be sure, parallelists maintain in addition that there is no causal interplay between the two realms, a claim that a materialist would not accept, of course, but global supervenience is neutral with respect to causal relations. Nomological global supervenience just says something about the coexistence of mental and physical states while causal relations always connect earlier with later events. A distinctive materialistic content, therefore, is only to be found in claims about an analytic global supervenience of the mental on the physical. But there is also an argument against a nomological global supervenience: It would very severely restrict our nomologically possible doxastic attitudes towards the physical world. It should be possible, for instance, for every physical fact that person P recognizes this fact. Otherwise there would be physical facts unknowable to P. This would be contrary to the intentions of materialism since it would be absurd to maintain a nomological identity of mental states of fact with physical ones which we cannot possibly know. But this harmless additional assumption, together with nomological supervenience, implies that P must be omniscient.6 6

Cf. Kutschera 1994 and 2009, 4.4. – If you feel that the possibilities of belief are very restricted anyhow you can reformulate the argument with another propositional attitude, for instance that of thinking about a state of affairs. It should surely be nomologically possible to think about every particular physical state.

200

Franz von Kutschera

The result, then, is this: Even the weakest claim of a supervenience of the mental on the physical is untenable. If no supervenience relation holds that does not mean that there are mental processes which are not accompanied by physical ones, especially by brain processes, of course. It only means that we cannot always distinguish neurologically between psychologically different events, and that neurological classifications must sometimes remain coarser than psychological ones. The infinity of mental constructions, then, is a decisive obstacle to materialism. The reason, in a nutshell, is this: Each person at every time can take divers propositional attitudes towards physical states of affairs like believing, wanting, fearing etc. These attitudes give rise to a much larger set of states of affairs, and they may also be iterated. Physics does not know any attributes of propositions, and even if you count probabilities and possibilities among physical properties of states of affairs and causal relations as physical relations between them, they cannot be iterated and therefore do not give rise to an infinite hierarchy comparable to those we find in the mental realm. Therefore the mental is not reducible to the physical simply in view of its dimensions, no matter what correlations the brain specialists will turn up in the future. 5. References Boolos, G. 1971 “The iterative concept of set”, The Journal of Philosophy, 68: 215-32. Humboldt, W. v. 1960 Werke in fünf Bänden, ed. A. Flitner and K. Giel, Darmstadt 1960-81: Wissenschaftliche Buchgesellschaft. Kutschera, F. v. 1994 „Global supervenience and belief”, Journal of Philosophical Logic, 23: 103-10. Kutschera, F. v. 2001 “Concepts of a set”, in A. Newen, U. Nortmann, R. StuhlmannLaeisz (eds.): Building on Frege, Stanford: CSLI, 319-27. Kutschera, F. v. 2009 Philosophie des Geistes, Paderborn: Mentis. Lenzen, W. 1980 Glauben, Wissen und Wahrscheinlichkeit, Wien: Springer. Tarski, A. 1949: “The semantic conception of truth and the foundations of semantics”, in H. Feigl and W. Sellars (eds.): Readings in Philosophical Analysis, New York: Appleton-Century-Crofts, 52-84.

3. Ontology in Application

Philosophy and Neuroscience ANTONELLA CORRADINI, Milan

1. Introduction For most part of its long history, philosophy has been considered as the queen of the sciences, as an intellectual enterprise occupying an overarching position over the sciences. Philosophy’s privileged position has often been related to its epistemological status as an a priori discipline, and as such has been conceived of as fundamentally opposed to the empirical sciences, whose status is a posteriori. However, in mid-twentieth century, the authoritative philosopher of science W.V.O. Quine launched a powerful attack against the traditional thesis that a clear-cut line can be traced between a priori and a posteriori sentences (Quine 1951). Since justification of a scientific theory is holistic, all its components are subject to empirical testing. Even the most internal part of the theory, its logical core, is in principle empirically testable. Later on, Quine developed the idea of a naturalized epistemology, that is to say of a discipline that, divested of its traditional a priori methods, was searching in the empirical domain for new instruments for the analysis of science (Quine 1969). As had already occurred in the 19th century, naturalism in the Quinean fashion took the form of psychologism. This explains why the blueprint for philosophy had to be behaviourism, that is to say the most influential psychological theory of the period in which Quine was outlining his point of view. As behaviourism’s success began to decline, psychologism drew its nourishment from the psychological current directly deriving from behaviourism, that is to say cognitive science. In fact, in the last decades a trend was developing, mostly in the fields of epistemology and philosophy of science, but not alone in these domains, according to which the logical, a priori component of the philosophical theories should be replaced by the computational dimension coming from cognitive science (Bechtel 2008, 2009; Nersessian 2008; Thagard 1988, 2009). Nowadays, the computational approach has been smoothly extended to cognitive neuroscience, while the function of philosophy in this framework has not been subject to any remarkable change. If we leave aside the harmless philosophy of neuroscience, understood as the traditional philosophy of science as applied

204

Antonella Corradini

to neuroscience, the contemporary philosophical and scientific scene has been dominated so far by neurophilosophy. In search of a definition for it, we can find characterizations of different philosophical impact. In some places of her 2002 book, P.S. Churchland seems to confine neurophilosophy to the mere acknowledgment of neuroscience’s relevance to philosophical issues (Preface, vii). This is a quite modest requirement, and most “traditional” philosophers wouldn’t have any objections against it. But, some pages after, Churchland tells us that philosophical questions can be addressed through experimental techniques and that this trend “is putting blood back to philosophy” (Preface, x). I don’t know if philosophy is in need of medical assistance, but what is certain is that neurophilosophy is a development of the movement initiated by philosopher Quine in last century’s fifties. As cognitive science represents an updating of Quine’s psychologism based on behaviourism, neurophilosophy extends the range of the methods of which a naturalistically oriented philosophy should avail itself including, among others, neuroimaging techniques. The picture of philosophy we obtain from this short excursus is that of a kind of “ancilla scientiarum”: philosophy and the sciences are a long continuum, philosophy adopts scientific methods and procedures too, its nature is a posteriori as that of the sciences. In this essay I intend to examine whether these results necessarily derive from the development of neuroscience, or if there are other plausible ways of conceiving the relationship between philosophy and neuroscience. Although psychologism on a neurobiological basis is a mainstream position in the current philosophical landscape, in the last few years rival standpoints have been developed, with the aim to restore philosophy’s birthright. I shall first start with an analysis of one of such proposal (Bennett and Hacker 2003), then move on to make some criticisms of it and to expound my own point of view. 2. Bennett and Hacker1 on Philosophy and Neuroscience 2.1 Exposition of the theory Neuroscientists are used to complain about philosophers, arguing for ex1

In the following section I will mention mainly Hacker, who represents the philosophical side of the Bennett-Hacker couple.

Philosophy and Neuroscience

205

ample that they are unable to tackle in a rigorous and effective way problems of common interest, such as the mind-body relation or the nature of consciousness. But in Hacker’s view, this criticism depends on considering philosophy as a kind of neuroscience manqué, which does not succeed in achieving the results that progress in neuroscience increasingly show. Hence the claim either that philosophy should step aside in favour of neuroscientific experimental evidence or that it should give up its autonomy and instead be grounded in the firm terrain of neurobiology (Bennett and Hacker, 396 ff). But, according to Hacker, philosophy does not have a duty to bring about the results required by neuroscientists, because its nature and function are very different from that of neuroscience. As an analytic philosopher in the Wittgensteinian tradition, Hacker maintains that philosophy is above all a conceptual inquiry. Since words are combined into sentences in rule-governed ways, one of the main tasks of philosophy consists in discriminating between the forms of correct and incorrect combination (400). Of particular importance in philosophy is connective analysis, “which traces, as far as is necessary for the purposes of clarification and for the solution or dissolution of problems and puzzles at hand, the ramifying logico-grammatical web of connections between the problematic concept and adjacent ones” (400). Throughout Bennett and Hacker’s volume, connective analysis is applied to some fundamental psychological concepts – such as memory, perception, volition, emotions … –, in order to examine whether the use that cognitive neuroscience makes of them is or is not correct. In what follows, I shall enumerate some of the many differences that Hacker detects between philosophy on the one hand and neuroscience conceived as one of the empirical sciences on the other hand. Philosophy deals with matters of meaning, not with matters of fact. Philosophy is not interested in empirical truth and falsehood, but in the sense and non-sense of sentences. Philosophy is concerned with logical possibilities, not with empirical actualities. Philosophical results can’t be approximations in the way scientific truths are, since approximation to sense is one or another form of non-sense (401). In the same manner, philosophy does not have hypotheses to put forward as science does, because there is nothing hypothetical about describing the bounds of sense (403). What the outlined differences between philosophy and neuroscience have in common, is that they can all be traced back to the distinction between a priori and empirical. A mistake

206

Antonella Corradini

neuroscientists often make is to neglect this fundamental difference, thus requiring from philosophy services it can’t and mustn’t render. So much said, it is evident that from this perspective philosophy can’t be grounded in neuroscience. Let’s take the example of epistemology. This is an a priori inquiry into the web of epistemic concepts. Connexions among these are logical, thus normative, in nature, just as is the procedure of justification of knowledge claims (406). But Hacker does not limit himself to claiming the autonomy of philosophy; beyond this, he affirms that philosophy is presupposed by science, since it would be impossible to build up an empirical science without a prior clarification of the conceptual frame that establishes the relations of sense among the involved concepts. So philosophy is not just autonomous from science but it also displays a kind of primacy with respect to science, being the necessary condition of any empirical enterprise (2003, 407; 2007, 4-5). This primacy finds expression in Hacker’s conviction that a philosopher is not required to abandon his theories if empirical evidence is against them. Sure, philosophers commit mistakes too, but these mistakes are a priori: that is to say, not arising from failures in experiment (2003, 404). At the root of many cases in which Bennett and Hacker complain about the lack of sense in the neuroscientific use of psychological concepts, one special source of non-sense should be picked out. The authors maintain that most neuroscientists commit a fundamental logical mistake, the socalled mereological fallacy in neuroscience. Let us examine what such fallacy looks like. Many contemporary neuroscientists are used to ascribe psychological attributes such as believing, categorizing, reasoning and understanding to the brain, whereas these capacities are commonly attributed to the whole human being. In the authors’ construal, this use is due neither to new discoveries about the brain nor to any linguistic innovation, but, instead, to conceptual confusions, to nonsensical combinations of words. Bennett and Hacker support their thesis by claiming that these sorts of psychological ascriptions conflict with a logical principle (2003, 81), or axiom, as we could perhaps define it, that reads as follows: “Only of a human being and what resembles (behaves like) a living human being can one say: it has sensations; it sees, is blind; hears, is deaf; is conscious or unconscious”

Philosophy and Neuroscience

207

(Wittgenstein 1953, § 281).2 Hence, ascription of psychological attributes to the brain or its parts is proved to be incoherent (2007, 6). Every time neuroscientists ascribe to a human being’s brain attributes which logically belong to the whole human being, they are committing a mereological fallacy. Of course, the same can be said about the mind. An earlier generation of neuroscientists (Sherrington, Eccles and Penfield) at the beginning of 20th century championed Cartesian substance dualism, which ascribes all psychological predicates to the mind, conceived as a substance separate but causally connected to the brain. But, according to the above mentioned logical principle, it is incoherent to be a substance dualist too. In fact, “it is not the mind that feels pain, perceives, thinks and desires, makes decisions and forms intentions, but the person, who is a psychophysical unity, not a duality of two conjoined substances, a mind and a body” (2003, 106). Contemporary neuroscientists pay their tribute to a mutant form of Cartesianism, which ascribes to the brain, mainly conceived as identical with the mind, the same psychological attributes previously ascribed to the mind (2007, 20-21). But, if the human being is a psychophysical unity, it does not make any sense to try to explain human powers and faculties by recourse to the brain and it parts. Ordinary psychological explanation based on reasons cannot be replaced by neurological ones (2007, 7). 2.2 Critical remarks In this paragraph I intend to critically examine the theses put forward by Bennett and Hacker on the relationship between philosophy and neuroscience. We have seen that the major difference between philosophy and neuroscience is that the first is an a priori discipline, while the second is an empirical enterprise. From this major distinction all the above mentioned distinctions follow. However, what does it mean to say that philosophy is an a priori discipline? The authors tell us that, unlike the empirical sciences, philosophy is concerned with sense and non-sense, not with empirical truth or falsehood. But, does philosophy dispense just with empirical truth or with truth completely? In reply to Dennett’s objections, Bennett and Hacker declare that philosophy only dispenses with empirical truth. “As is patent, it is our view that philosophy is concerned with conceptual 2

Henceforth “mereological principle”.

208

Antonella Corradini

truths and that conceptual truths determine what does and does not make sense” (2007, 207). Yet, how are such conceptual truths to be understood? Do they go beyond the pure description of the rules which govern the web of words? Do they overcome the mere syntactical level to signify reality? I think the answer is in the negative. Actually, other characterizations that Hacker attributes to philosophy are plausible only if the semantic dimension of truth is fully ignored. For example, it can be hardly maintained that philosophy has no hypotheses to put forward, even though they are different from the empirical hypotheses of the sciences. The whole history of philosophy contradicts this claim! It is equally difficult to agree with Hacker that philosophical results can’t be approximations to truth. In fact, it is a long time since philosophy has dispensed with the pretension to possession of absolute, indefeasible truths. Still more unreasonable is to affirm that a philosopher should never abandon his theories if they do not fit with empirical evidence. This claim is valid in its generality only under the questionable assumption that philosophers’ business is “with the forms of thought, not with its content” (2003, 404). Anyone who thinks that philosophers’ business is also with thought content can neither rule out a priori the possibility that whole philosophical theories or parts of them will conflict with empirical evidence, nor foresee how the conflict will eventually be settled. I believe that, without casting into doubt the a priori nature of philosophy in the slightest or its primacy over empirical knowledge, another interpretation of the a priori is possible. I shall try to show what it looks like, by dealing first with the parallelism between philosophy and mathematics and secondly with the logical status of the mereological principle. Neuroscientists often balk at the a priori character of philosophy but they shouldn’t do so, because it is comparable to physicists balking at the a priori character of mathematics. As Bennett and Hacker put it, “conceptual questions are no more amenable to empirical methods of investigation than problems in pure mathematics are solvable by the methods of physics” (2007, 5). Thus, philosophy and mathematics have aprioricity in common. “Mathematics is concept formation by means of a priori proof – for a mathematical proof forms a novel conceptual connection articulated by the theorem proved. Philosophy is concept elucidation by means of the description of the rule-governed use of words” (2003, 402). Now, it is true that philosophy deals with the analysis of concepts. The concepts subjected to analysis are specific to philosophical inquiry, such as for example the

Philosophy and Neuroscience

209

concept of object, complete object (i.e. individual substance), incomplete object (that is to say abstract object, like mathematical entities), existent object, possible object, and so on. The method of inquiry that philosophy applies to the investigation of these concepts is fundamentally a priori and under this aspect philosophical method is tightly connected to mathematical method. In its turn, mathematical knowledge is based on analysis of essential properties of abstract objects such as numbers. They are conceptually defined through the axioms of the theories put forward to characterise them. For this reason these properties are the subjects of direct analysis in axiom formulation and are made indirectly explicit, owing to logical deduction, in the theorems. But it is mathematical knowledge itself, in its analogy with philosophical knowledge, that allows us to understand the essential role that the semantic dimension plays in conceptual analysis. In fact, a purely syntactic conceptual analysis of the fundamental notions of a mathematical theory has been made impossible by the discovery of formalization limits in mathematics. This implies that the analysis of the primitive concepts of a mathematical theory cannot consist in the individuation of merely syntactic rules which govern their use. Instead, such rules should be conceived as expressions of properties and relations among the entities that make up the semantic model of the theory itself. Thus, ontological laws are not mere rules only designed for elaborating the conceptual framework within which to systematize information coming from experience. Rather, they are descriptions, even though fallible, of the general structure of reality or, at least, of some aspects of it. In other words, something putatively existing in reality has to correspond to the semantic connections resulting from a non reductive analysis of philosophical concepts, in the same way as mathematical theories regard the contents of their respective models. Let us turn now to the mereological principle. As we have seen, it does not hold in virtue of some matter of fact (2003, 71); rather, it is a logical principle and this means, in Hacker’s view, that it is a convention,3 although not any convention whatsoever, but the one that determines what does and what does not make sense. It is a convention deeply rooted in the linguistic and conceptual context in which we live. In principle it would be possible to change it by stipulation,4 but only if we were ready to also 3 4

My italics. My italics.

210

Antonella Corradini

change the whole conceptual framework that we share (2003, 81). Since the mereological principle is a convention, it is simply given: no further justification of it is needed. However, in his criticism of Hacker, Searle traces the supposed plausibility of the mereological principle to the role that in Hacker’s theory behaviour plays in the ascription of psychological predicates (2007, 101). In fact, Hacker maintains that behaviour is the primary evidence for the ascription of psychological predicates to another. But this evidence is not inductive evidence; rather, behavioural evidence is conceptually tied to the meaning of the relevant predicate. “Pain-behaviour is a criterion - that is, logically good evidence for being in pain … That such-and-such kinds of behaviour are criteria for the ascription of suchand-such a psychological predicate is partly5 constitutive of the meaning of the predicate in question” (2003, 82). Things being so – Searle argues – it follows immediately that only a living human being can be the bearer of psychological predicates, since brains are unable to show the appropriate behaviour (2007, 101). Now, if Searle’s interpretation of the mereological principle is correct, Bennett and Hacker’s theory is at risk of facing a paradoxical situation. On the one hand, the mereological principle is invoked in order to show that ascription of psychological properties to the brain is non-sensical. On the other hand, however, this very principle is derived from the thesis that behaviour is criterial and this thesis is equivalent to the assertion of logical behaviourism: that is to say to an arch-reductionistic stance. But Bennett and Hacker are aware of the danger and explicitly declare that they are not defending a form of behaviourism. As they put it, “the conceptual nexus we are insisting upon is an a priori evidential, but not a reductive, one” (2003, 82, footnote 35). In fact, that the behavioural description does not exhaust the meaning of the predicate also emerges from the quotation above, where it is argued that behaviour is only partly constitutive of the meaning of the predicate. But what does this mean? I’ll try to elucidate this point. In the sentence quoted above Bennett and Hacker make no distinction between the logical and the conceptual level, but in other places they seem to presuppose this distinction. On page 82 of 2003, for example, they affirm that whereas the logical relation of entailment – to be a bachelor entails being a non-married man – is not defeasible, the conceptual nexus between psychological predicates and their be5

My italics.

Philosophy and Neuroscience

211

havioural description can sometimes be overridden by countervailing evidence. This example clearly shows that there is a difference between, on the one hand, the logical, purely syntactical relation between ‘bachelor’ and ‘non-married man’ and, on the other, the conceptual relationship between being-in-pain and pain-behaviour. That pain-behaviour is only partially constitutive of the meaning of ‘is in pain’ has the following consequence: that there are other aspects beyond behaviour that make up the meaning of ‘is in pain’– for example the subjective aspect of pain. What all these aspects are can be established through meaning postulates (holding between autonomous meanings), whose acceptance implies adopting a semantic stance that is incompatible with the logico-grammatical, formal or conventional standpoint that Bennett and Hacker defend in their work. To sum up: It is true that the authors avoid falling into contradiction thanks to their refusal to adhere to behaviourism. However, by choosing the second horn of the dilemma they are forced to accept a semantic dimension that confers a different status on the mereological principle to that originally intended. It is no longer possible to affirm that it only derives from the grammar of concepts like whole, part and their respective attributes and that it cannot conflict with the data of experience, since it is the result of a conventional stipulation regarding the rules of use of these concepts. Instead, it must be emphasized that it derives from the real meaning of these concepts and from the laws which depend on them. Only so understood can the mereological principle be claimed to say something about the world and to possibly conflict with the data of experience. How to solve this conflict will depend both on the degree of the principle’s a priori reliability and on that of the data of experience. 3. Against physicalistic ontology Once it is established that the mereological principle has an ontological and not a purely logical nature, it still remains to explain what role it plays in the current debate about the philosophical interpretation of neuroscience. From my own point of view, its relevance is quite limited. The mereological principle forbids ascription of psychological predicates to parts of human beings such as brains, and it does so in order to fulfil an anti-reductive function. I believe, however, that the mereological principle is neither necessary nor sufficient to ground an anti-reductive approach.

212

-

-

Antonella Corradini

The mereological principle is not a necessary condition of antireductivism because, even if anti-reductivism implies the principle, this latter can be subject to a reductive reading, according to which the whole person coincides with her brain. To exclude this circumstance, we are in need of further arguments, which, however, are independent from the principle itself.6 Nor is the mereological principle a sufficient condition of antireductivism. In fact, I submit, a holistic approach which considers the human being as a psychophysical unity can produce reductive outcomes, if it is based upon a physicalistic ontology.

Moreover, while on the one hand acceptance of the mereological principle does not ensure the validity of anti-reductivism, on the other hand it implies the rebuttal of substance dualism as its further logical consequence.7 But it would be too a hasty conclusion to affirm that any kind of substance dualism is a wrong philosophical standpoint. In the remains of this section I shall try to show what results can be drawn from a non-reductive ontology without mereological principle. The development of the philosophy of mind starting from the late fifties of the 20th century has shown a tendency to avoid extreme positions, that is to say to outline intermediate positions between the two extremes of materialism8 and of dualism. Strangely enough, however, the philosophical theories which have been elaborated within this framework have not 6

For a critical discussion of such conceptions see Lowe 2008. Bennett and Hacker have two distinct reasons to argue against substance dualism. The first one, as we have seen, is tied to adherence to the mereological principle. The second one goes along with Bennett and Hacker’s view of mind. In their construal, “mind” is a mere facon de parler, since it does not correspond to a real entity, but it is merely a way of describing the complex of human powers and character traits. Via this de-ontologization of the mind they reach the same result they attained through the mereological principle, that is to say that it is the whole person and not the mind that is the bearer of psychological attributes. On top of that, since the mind “is not a nothing, but it is not a something either” (2003, 105), according to the authors it does not make any sense to identify the mind with a something like the brain, as neuroscientists mostly do. The de-ontologization of the mental substance is likely to be a worse move than abiding by the mereological principle. For a fuller treatment of this point see the conclusion of this essay. 8 In this essay I will not sharply distinguish between “materialism” and “physicalism”. 7

Philosophy and Neuroscience

213

proved to be genuinely impartial between materialism and dualism, because they have many traits in common with materialism and only few or none with dualism. Let us first reflect on the passage from reductive physicalism to nonreductive physicalism, the first theory in the 20th century which aspired to guarantee a proper level of autonomy to the mental. By definition, reductive and nonreductive physicalism are both physicalistic theories, since their difference only consists in the reductive or nonreductive nature of physicalism. Physicalism is non reductive if it fulfils the requirement of multiple realizability, that is to say the principle according to which the same mental state can be realised by different physical states. In other words, if mental states and cerebral states are identical, such identities are contingent and not necessary ones. (Fodor 1974, 99-101). We could ask ourselves if nonreductive physicalism sufficiently safeguards the autonomy of the mental and its specific characteristics. To begin with, physicalism, either reductive or nonreductive, relies on the thesis that every mental entity, at least at the moment of its realization, is physical. But then, what about the phenomenal qualities of our subjective experience (the qualia of experience), which cannot be reduced to the physical dimension? Nonreductive physicalism’s answer won’t be more liberal in this regard than that of reductive physicalism, since both confine themselves to taking only objective, third-person cognitive processes into consideration. On top of that, physicalism, both in its reductive and in its nonreductive variant, is a thesis about the nature of causality, according to which every physical event is exclusively determined by physical causes. If this thesis of completeness or of causal closure of the physical world is valid, as physicalism maintains, how can mental causation be autonomous with respect to physical causation? Once again, as far as crucial questions for the credibility of the theory are concerned, nonreductive physicalism does not seem to perform better than reductive physicalism, that is to say, it does not seem to represent the intermediate position between materialism and dualism it aspired to. Discontent about nonreductive physicalism induced some philosophers, at the beginning of the nineties, to regard emergentism as a candidate for the golden mean between materialism and dualism. Emergentism seems to be a good candidate because it appears to be more able to address the difficulties raised by the phenomenal dimension than nonreductive physicalism, such as the previously mentioned qualia of experience and autonomous mental causation. As to the former, Broad, one of the most authorita-

214

Antonella Corradini

tive representatives of the British Emergentists, maintains that there is at least a kind of phenomena that cannot be explained reductively, that is to say the qualia of experience: they display an ultimate character, irreducible to physical data (1925, 52). Moreover, when Broad examines reductive materialism, which he identifies with behaviourism, he argues that it results in an identity thesis between certain kind of mental processes and certain kinds of observable behaviours (612). Instead, according to Broad, the sensation of a certain colour and a certain physical movement are two different characteristics and the former is not reducible to the latter. As to the issue of mental causation, emergentism conceives it within a wider framework: the world would be constituted by levels of reality of different complexity, whereby the more complex levels of reality are able to exert a causal action on the lower levels. Without disregarding the role of bottomup causation, the higher levels, including the mental, are supposed to be able to change the course of the events running at the lower level. The typical example of downward causation in the field of the mental is that of mental states which are at the origin of the planning and execution of an action. Has emergentism thus achieved the desired goal of preserving the fundamental traits of subjective experience and of mental causation without being forced to adhere to dualism? Broad’s further claims make us doubt that this is the case. Even though he rejects reductivism, in fact, he does not hesitate to define himself as a materialist, that is an emergentist materialist. As to downward causation, Broad declares that the mental properties of material events are “completely determined” by the material properties of these events (623). But is a kind of nonreductive materialism, for which bottom-up determination holds, really able to account for subjective experience and mental causation? Broad does not seem to believe that the irreducibility of secondary qualities must lead to any breach of materialism, since, from his point of view, secondary qualities do not exist in reality, but just seem to exist. In other words, the qualia of experience can be considered compatible with materialism only if they are interpreted as Broad proposes, that is to say in a purely phenomenal sense (204 ff). To Broad it can however be objected that, whatever the ontological status of qualia may be, they are characterized by the fact that they appear to a subject, who, as such, belongs to a dimension of reality that is different from the material one. It follows that it is the ineliminable existence of subjects of experience which represents a strong argument in favour of a non mate-

Philosophy and Neuroscience

215

rialistic conception of reality. As to downward causation, some contemporary philosophers have maintained that it is contradictory, because it violates the most elementary principle of the transitivity of causation to suppose that the properties caused by the micro-elements of a system have an autonomous causal influence on the very properties that have caused them.9 It can be counter-objected that bottom-up determination implies that the micro-properties are both necessary and sufficient conditions of higherorder properties. This point of view, while consistent with physicalism, raises the further question of what causal role is left to higher-order properties. My answer is: none, as all the causal work is done by the physical micro-components. This is why Kim affirms that downward causation does not imply any ontological commitment, as it is only a way of describing the world (1999, 33). If, on the contrary, we intend to defend the ontological character of downward causation, we have to give up the physicalistic thesis of micro-determination and to claim that micro-levels are at most a necessary but not a sufficient condition of higher levels. Thus, good initial intentions notwithstanding, even emergentism is not a position neutral between materialism and dualism. As nonreductive physicalism, it is a kind of materialism, which makes it ultimately impossible to really account for mental phenomena. A first conclusion is that, if we have a (truly) nonreductive stance about the relationship between mind and brain we should not only not trust the mereological principle, as we already stated above, but moreover make a crystal-clear choice against physicalism and in favour of a non-physicalistic ontology. A further question is what such an ontology should look like. The logical consequence to draw from my foregoing argument is that a nonreductive ontology ought to be a dualistic one, that is to say an ontology which is ready to encompass not only the physical or material but also the mental. However, there are and there have been many kinds of dualism in the history of philosophy, whose variety cannot be discussed here. Elsewhere (Corradini 2008 and 2010) I have argued in favour of an emergentist form of dualism, which I’ll now try to briefly sketch. Even though the original emergentistic theory was monistic, materialistic and empiricistic, there are at least two good reasons to champion a dualistic kind of emergentism. Firstly, fundamental concepts for emergentism such as the phenomenal qualities of subjective experience and downward 9

Searle 1994, Kim 1999 and 2006. For a reply see O’Connor-Wong 2005.

216

Antonella Corradini

causation loose their radicalism and pregnancy if put in a monistic framework. Pace the British Emergentists, it does not seem sensible to coin conceptual categories that, upon closer inspection, reveal themselves to be insignificant if included in an overall view of a materialistic sort. A second, more ambitious line of defence of emergent dualism consists in questioning whether it really contradicts the Emergentists’ original intentions. They were against both mechanicism (reductivism) and vitalism (a kind of dualism). The reason they put forward to reject the latter was that it introduced into biological evolution entities such as the élan vital and the entelechies, which escape empirical control. However, it must still be proved that emergent dualism has the same fault that the British Emergentists attributed to vitalism, that is to say to introduce into reality obscure entities belonging to a realm outside the natural course of events. Instead, it can be shown that emergent dualism meets the requirement to proceed along purely naturalistic lines of advancement. Let us turn back to what we have written above about downward causation, that it to say that it is impossible if the causal influence of lower-level properties is both a necessary and a sufficient condition of the mental properties’ causal action. If we formulate this statement of impossibility in the positive, we get at the conceptual core of emergent dualism: higher-level mental functions, although presupposing activation of the neurophysiological level (since there is no thought without brain), cannot be produced by their neurophysiological basis alone. In order for the mental to emerge from its neurophysiological basis, a non material dimension of reality is necessary, endowed with ontological independence and existent from the very beginning of any emergent process. Only if this is the case, higher-level mental functions can exert an influence upon the brain’s activity, that is to say they are a sufficient condition of downward causal action. Their neurophysiological basis, in its turn, while insufficient to generate higher-level mental functions, constitutes the necessary condition of these latter. This short description of the salient characteristics of emergent dualism lets us understand that it does not imply any insertion of strange entities from outside nature: the non material, mental dimension is present in reality from the very beginning of any emergent process, because it is, by full right, an integral part of nature. The notion of nature deriving from emergent dualism is significantly different from that we are used to, so that a gestaltic re-orienteering of our conceptual categories is surely indispensable. Nonetheless, why should it be implausible that psychology, understood as the science of the mental, may acknowledge in

Philosophy and Neuroscience

217

the mental a further natural dimension beyond the purely physical one? To support mind-brain dualism does not mean abandoning the naturalistic point of view outright, but, instead, conceiving it within a wider context, so as to also include the mental and its relationship to the neural. It is because the mental is a natural phenomenon that it is reasonable to affirm the evolutionary significance of the aspects which are typical of the mental but not compatible with physicalism.10 These are the product of the evolutionary history of the human being and display a high adaptive value, inasmuch as they confer on the living beings possessing them an evolutionary advantage over those which only possess physical properties.11 Moreover, a further peculiar notion of emergent dualism should be emphasized, that of the co-evolution of mind and brain. As we have just seen, emergent dualism postulates that a non material dimension of reality is at the origin of the developmental potentialities of the human being’s mental component. Such a component becomes actualised at the very moment in which the biological structure reaches the necessary degree of complexity. According to mind-brain co-evolution, once the realisation of non biological potentialities is induced by the development of the biological structure, this latter becomes the subject of the causal activity of the conscious mind. Finally, a last remark. Following a stale common-place, substance dualism is often considered as an old-fashioned conception, above all because it would be not compatible with contemporary science. From my point of view, however, it is possible to show that emergent substance dualism is even more plausible and coherent than property dualism (Corradini 2008, 203-6). But, however this may be, I want to stress here that, even though it supports substance dualism, emergentism is fully able to justify the existence of the many correlations between mind and brain that neuroscience testifies. Despite the fact that the mental substance is endowed with ontological independence, from the functional standpoint it is nonetheless dependent on the brain, since none of its typical functions such as perceiving, thinking or deciding could occur without the activation of the brain. This shows that even the allegedly most extreme form of dualism, if emergentis-

10

For example the existence of a self characterized by a first-person ontology. See Hasker 1999 and Meixner 2004 and 2008. To Meixner’s pioneering studies on dualism I owe much inspiration and insight.

11

218

Antonella Corradini

tically understood, can be wholly compatible with the most recent discoveries of neuroscience. References Bechtel, W. 2008 Mental Mechanisms. Philosophical Perspectives on Cognitive Neuroscience, New York – London: Routledge. Bechtel, W. 2009 “Constructing a Philosophy of Science of Cognitive Science”, Topics in Cognitive Science 1, 548-569. Bennett, M.R. and Hacker, P.M.S. 2003 Philosophical Foundations of Neuroscience, Malden, MA: Blackwell Publishing. Bennett, M.R et al. 2007 Neuroscience and Philosophy. Brain, Mind and Language, New York: Columbia University Press. Broad, C.D. 1925 The Mind and Its Place in Nature, London: Routledge and Kegan Paul. Churchland, P.S. 2002 Brain-Wise. Studies in Neurophilosophy, Cambridge, MA: the MIT Press. Corradini, A. 2008 “Emergent Dualism”, in A. Antonietti, A. Corradini and E.J. Lowe (eds.) Psycho-physical Dualism Today. An Interdisciplinary Approach, Lanham, MA: Lexington Books, 185-209. Corradini, A. 2010 “How Special are Special Sciences?”, in A. Corradini and T. O’Connor (eds.) Emergence in Science and Philosophy, New York – London: Routledge, 289-304. Fodor, J.A. 1974 “Special Sciences (or: The Disunity of Science as a Working Hypothesis)”, Synthese 28, 97-115. Hasker, W. 1999 The Emergent Self, Ithaca and London: Cornell University Press. Kim, J. 1999 “Making Sense of Emergence”, Philosophical Studies 95, 3-36. Kim, J. 2006 “Emergence: Core Ideas and Issues”, Synthese 151, 547-59. Meixner, U. 2004 The Two Sides of Being. A Reassessment of Psycho-Physical Dualism, Paderborn: mentis. Meixner, U. 2008 “The Reductio of Reductive and Non-reductive Materialism – and a New Start”, in A. Antonietti, A. Corradini and E.J. Lowe (eds.) Psychophysical Dualism Today. An Interdisciplinary Approach, Lanham, MA: Lexington Books, 143-166. Nersessian, N.J. 2008 Creating Scientific Concepts, Cambridge, MA: MIT Press. O’Connor, T. and Wong, H.Y. 2005 “The Metaphysics of Emergence”, Noûs 39, 65979. Quine, W.V.O 1951 “Two Dogmas of Empiricism”, Philosophical Review 60, 20-43.

Philosophy and Neuroscience

219

Quine, W.V.O 1969 “Epistemology Naturalized”, in W.V.O. Quine (ed.) Ontological Relativity and Other Essays, New York: Columbia University Press, 69-90. Searle, J. 1994 The Rediscovery of Mind, Cambridge MA: MIT Press. Thagard, P. 1988 Computational Philosophy of Science, Cambridge, MA: MIT Press. Thagard, P. 2009 “Why Cognitive Science Needs Philosophy and Vice Versa”, Topics in Cognitive Science 1, 237-254. Wittgenstein, L. 1953 Philosophical Investigations, Oxford: Basil Blackwell.

How Molinists Can Have Their Cake and Eat It Too* GODEHARD BRÜNTRUP AND RUBEN SCHNEIDER, Munich

Introduction Edmund Runggaldier has always pursued the goal of reconstructing positions in the classic tradition of Christian philosophy within the framework of contemporary analytic philosophy. As a Jesuit he will certainly be delighted to see that one of the greatest contributions to philosophy by a fellow Jesuit, Luis de Molina S.J. (1535–1600), finds so much interest in contemporary analytic philosophy. With Alvin Plantinga’s influential “The Nature of Necessity” (Plantinga 1974) and David Lewis’s possible-world analysis of counterfactuals (Lewis 1973) there were new conceptual tools available to tackle traditional problems in the philosophy of religion, like that of theodicy or that of reconciling the orthodox concept of God with that of libertarian human freedom. This latter topic had already been conceptualized within a rather similar logical apparatus in the late periods of the Schoolmen, the Baroque Scholastics, especially by Molina. It should come as no surprise that analytic philosophers took interest in the work of this great philosopher and theologian (cf. Flint 1998, 122, fn. 1). The problem of Divine foreknowledge and human freedom bundles many great topics of analytic metaphysics into a focal point: causality and freedom, time and modality, and the semantics and ontology of counterfactuals. In this paper we will try to reconstruct Molina’s position within the contemporary analytic debate and specify to what extent Molina’s account of freedom can still be counted as a libertarian view: The Molinist claims to be a libertarian and nevertheless grants God full knowledge of future *

Incidentally, this common English proverb has its earliest recording in the time of Molina. The author is the English playwright John Heywood who was imprisoned for his Catholic beliefs and later fled England to avoid religious persecution. In a collection of proverbs, published by him in 1546, one finds: “wolde you bothe eate your cake, and have your cake?”

222

Godehard Brüntrup and Ruben Schneider

contingents with regard to human freedom. With disbelief many philosophers reply to this double-claim with “you can’t have it both ways” or “you can’t have the best of both worlds.” We try to show that the Molinist can actually “have it both ways.” There is a price to be paid, however: the notion of libertarian freedom has to be defined in a way that does not require an open future. In this paper we suggest such a notion based on the mathematical concept of a “choice-function.”1 1. Overview of the debate The following system of propositions can serve as a prima facie representation of the logical structure of contemporary positions regarding the Divine attributes of omniscience and omnipotence and their relation to human freedom (cf. Rhoda 2007, 301-303). The following four propositions cannot collectively be true. Only certain selective combinations of these four propositions are logically coherent. The denial of at least one of the four propositions is required: (1) God has the attributes of absolute independence, omniscience and omnipotence. (2) There are future contingents, the future is causally open. (3) It is impossible that the future be causally open and epistemically determined for someone. (4) The future is epistemically open for God. The Open Theist assumes (2), (3), and (4). Since (4), however, contradicts (1) it has to be denied by the Open Theist. Several theories try to temporalize God’s knowledge and thus God Himself (non-eternalism).2 The denial of (4) is common to Theological Determinism, Analytic Thomism, and 1

We will have to assume a certain familiarity with this debate on the part of the reader. It is not possible to give a full theoretical explication of all the technical notions used in this paper. 2 A weaker version of Open Theism could replace (3) by this assumption: It is impossible that the future be open and all propositions about the future have definite truth values. The future would be unknown to God because the truth-values of propositions concerning the future have not yet been determined. Since knowledge implies truth, God knows all knowable propositions and is thus omniscient (cf. Rhoda 2007, 303304). This position runs into a conflict with the Law of Conditional Excluded Middle.

How Molinists Can Have Their Cake and Eat It Too*

223

Analytic Molinism because they reject the notion of a non-omniscient God (Deus ignorans). However, a further distinction has to be made. Proposition (4) is a logical consequence of (2) and (3). Denial of (4) implies thus the denial of either (2) or (3). The theological determinist denies (2) while holding on to (3), thus claiming strict causal determination of the future. Thomism and Molinism can be subsumed under the negation of (3) while still asserting (2). Thus, they both believe that even though the future is not causally determined, God can still have foreknowledge of it. In Thomism this is explained by an eternalist concept of time in combination with a specific understanding of causality (cf. Craig 1988, 166-118): future contingents cannot be known as future events; rather, God contemplates them sub specie aeternitatis as presently given to Him. He knows tenseless propositions with definite temporal indexicals. For the Thomists3 this does not entail that God is in any way determined by temporal events, His aseity is uncompromised: all events, including free decisions and actions, are premoved (praemotio physica) and pre-determined (praedeterminatio physica) by the prime cause.4 This is also the medium (medium quo) by which God has knowledge of future contingents. Molinism rejects the Thomistic explanation which is based on eternalism and praemotio physica. The latter, according to the Molinist, is a first step on a slippery slope towards determinism. Molinism assumes only a rather general entitative concurrence of God, one that remains indifferent with regard to free acts of the human will (concursus simultaneus et indifferens): “the ‘primary cause’ of everything is God. Yet a human action is free only if the creaturely agent gives God’s ‘general concurrence’ that makes that action possible a particular direction by causally contributing to it as well” (Jäger 2011, 89). In order to safeguard God’s sovereignty over 3

We refer here to the historically influential Bañezian brand of Thomism (Domingo Bañez, 1528-1604). 4 Can the free human will resist the pre-determined praemotio by God? The Thomists have a rather complex answer to this question which we cannot address here. In a nutshell: God causes these free actions contingently and as contingent events. God predetermines the free will in its free nature. In sensu diviso there is even in a premoved will the potentiality to resist, at least counterfactually. In sensu composito, however, the finite will is unable to make another choice (cf. Thomas Aquinas, De Ver, q. 6, a. 4, ad 8; q. 23, a. 5, ad 3).

224

Godehard Brüntrup and Ruben Schneider

his creatures, Molina postulates a Divine knowledge of certain counterfactuals: the theory of middle-knowledge. He distinguishes three specifications of Divine knowledge: first, the pre-volitional knowledge of all necessary truths and all possible worlds (scientia naturalis); second, the postvolitional knowledge of all contingent actual truths, including future ones, after creation (scientia libera); third, and in between the other two, the prevolitional knowledge of all contingent truth-values of all counterfactuals or pre-factual conditionals, especially those involving the counterfactuals of freedom (scientia media): “If creaturely essence P were instantiated in nondetermining complete circumstances C at time t, the instantiation of P would (freely) do A” (Flint 1998, 47). (To formalize this, we will be using the notation of David Lewis: “C □→ A.” “□→” denotes the “would”implication: “If it were the case that C, then it would be the case that A”. This has to be distinguished from the “might”-implication “C ◊→ A”: “If it were the case that C, it might be the case that A,” cf. Lewis 1973, 2.) Neither the counterfactuals of freedom nor the Divine act of creation cause (by efficient causation) free human actions. God’s sovereignty is established by the fact that He has access to this middle knowledge logically prior to His free choice in the act of creation. God can freely create circumstances (or worlds) and place free creatures in them in such a way that they freely choose under those circumstances those free actions that fulfill the Divine creative plan. The scheme of the four propositions enabled us to systematize the positions on God’s omniscience and the openness of the future. When it comes to the problem of freedom and determinism regarding human actions, another way of systematization is helpful. It is a scheme of three propositions forming a logical trilemma. Any two of the three propositions are logically consistent, but the third has to be denied (cf. Flint 1998, 22-23; Brüntrup 2000, 102-10): (a) Some human actions are free. (b) All humans actions are determined by circumstances and events which are outside of the agent and not under the agent’s control (remote past and laws of nature, Divine foreknowledge and predestination). (c) It is impossible that a free action be determined by circumstances and events which are outside of the agent and not under the agent’s control. Theological Determinism denies propositions (a) and affirms (b) and (c). This is the claim that there are no free human actions. Open Theism denies

How Molinists Can Have Their Cake and Eat It Too*

225

proposition (b) and asserts (a) and (c); it is clearly a libertarian position. Bañezian Thomism denies (c) and affirms (a) and (b) with respect to determination by God. It can thus be classified as compatibilism. In the literature Molinism is commonly labeled as a libertarian position (cf. Flint 1998, 24-34). This implies the denial of (b) and the affirmation of (a) and (c). However, classifying Molinism as a libertarian position gives rise to some controversial claims which will be systematically addressed in what follows. 2. The anti-molinist consequence argument It is our contention that if Molinism is to be a libertarian position, the Molinist had better consider the Consequence Argument against compatibilism a valid argument. The conclusion of the Consequence Argument is that nobody has a free choice about anything, assuming that we live in a deterministic world, and the past and the laws of nature cannot be changed. We will label this argument “CAComp” (cf. van Inwagen 1983, 96-98, 147). The problem of Divine foreknowledge and human freedom is logically analogous to the problem of freedom and determinism. It is of considerable significance in this context that some arguments against Molinism have the logical structure of the Consequence Argument against compatibilism. In that case the Molinist might be forced to use similar strategies to those put forth by the compatibilists who argue against the Consequence Argument. Let us call such an argument against Molinism “CAMol” (cf. Perszyk 2003). Let C be a set of circumstances and A be an action, then let “C □→ A” be any true counterfactual of freedom known by Divine middleknowledge. We also define a “Choice-operator”: Chxp := “p obtains and it was in x’s power to take an action A such that A occurs and therefore p is true” (cf. Rugel 2006, 41-44). We define the “No-Choice-Operator” in the standard way: Np := p ‫( ר‬x: Chxp) For any free action the following principle holds: (F) A is free : x: ChxA  NA We also assume the rules α and β, well-established in the literature:

226

Godehard Brüntrup and Ruben Schneider

(α) □p → Np (β) Np, N(p → q) ├ Nq We can then argue against Molinism in a way strictly parallel to the classical Consequence Argument (cf. Perszyk 2003, 148): (CAMol,1) (1) (2) (3) (4) (5) (6) (7)

□{[(C □→ A) ‫ ר‬C] → A} □[(C □→ A) → (C → A)] N[(C □→ A) → (C → A)] N(C □→ A) N(C → A) NC NA

Consequence of Molinism from 1) (2), α-rule Fixity of middle knowledge (3), (4), β-rule Axiom (5), (6), β-rule

But since principle (F) holds for any free action, CAMol,1 implies the incompatibility of Molinism and libertarian freedom. Due to difficulties with the original Van Inwagenian Consequence Argument, especially with the βrule, many improved versions have been advanced in the literature. For reason of simplicity, elegance and the relative lack of problem cases, we thus suggest the following definitions (cf. Brüntrup 2000, 112; Rugel 2006, 56): N*p := p ‫( ר‬x: Chx(◊p)) (β*) N*p, □(p → q) ├ N*q The Anti-Molinist Consequence Argument can now be rephrased: (CAMol,1*) (1) N*[(C □→ A) ‫ ר‬C]

Fixity of middle knowledge and antecedents (2) □{[(C □→ A) ‫ ר‬C] → A} Consequence of Molinism (3) N*A (1), (2), β*

In his “Molina on Foreknowledge and Transfer of Necessities” (2011), Christoph Jäger draws on an informal version of this argument. As a stronger and more coherent replacement for Molina’s notion of “accidental necessity” (of propositions known by middle knowledge), Jäger introduces

How Molinists Can Have Their Cake and Eat It Too*

227

the notion of “causal-impact necessity” or “CI-necessity” (cf. Jäger 2011, 84-88):5 (CI-necessity) A metaphysically contingent proposition or state of affairs p is CI-necessary for a given agent at t iff it is not within the agent’s power at t to contribute causally to some thing that constitutes, or grounds, a necessary or sufficient condition for p (cf. Jäger 2011, 88). The principle (F) can thus be read as: (F’)

For libertarian freedom with respect to an action A, it is necessary that that A is not CI- necessary for the agent.

The transfer principle given by Jäger is accordingly: (β’)

“If it is CI-necessary for a given subject at t that p, and p [strictly6] entails q, then it is CI-necessary for that subject at t that q” (Jäger 2011, 92).

Jäger’s overall argument now runs like this (slightly modified by us, cf. Jäger 2011, 91) (CAMol,2) (1) At any time in S’s life it is CI-necessary for S that God knows (prevolitionally and from extra-temporal eternity) via middle knowledge that S, when placed in circumstances C at t, would freely do A. (2) God’s knowing (prevolitionally and from extra-temporal eternity) via middle knowledge that S, when placed in circumstances C at t, would freely do A, entails [strictly] that S, if placed in C at t, freely does A. 5

“A proposition p is accidentally necessary at t iff p is (i) metaphysically contingent and (ii) true at t and every moment after t in every possible world that has the same history as our world at t” (Jäger 2011, 86), or, as defined by William of Sherwood: “That is accidentally necessary which neither now or in the future can be false, but once might have been false” (Sherwood, Introductiones in logicam, 11, 34, cf. Jäger 2011, 84). 6 That is, by metaphysical necessity (there is no possible world, such that God knows a counterfactual of freedom in it, and the appropriate event does not occur if the the relevant circumstances have been actualized). This clarifying and relevant addition can be found in Jäger 2010, 313.

228

Godehard Brüntrup and Ruben Schneider

(3) Hence, it is CI-necessary for S that, when placed in C at t, S freely does A. (4) S is not free in doing A freely when placed in C at t. The inference from (2) to (3) is based on the β’-rule, and the inference from (3) to (4) is based on the definition of libertarian freedom given in (F’) above. The CI-necessity of a proposition p implies, that the subject has no choice such that possibly p (i.e. Chx(◊p). Thus, the β’-rule operating in CAMol,2 is as strong as β* in CAMol,1*. These inferences can, just as the classical replies of the compatibilists to CAComp, be undermined by a refutation of the relevant β-rules. This is the route taken by Jäger, following an earlier argument by Freddoso (cf. Jäger 2011, 92-96; Freddoso 1988, 53-62). The appropriateness of this move seems to be vindicated by the fact that Molina himself seemingly rejects modal transfer principles in Concordia 4.52.34: For in such a case, even if (i) the conditional is necessary (because in the composed sense these two things cannot both obtain, namely, that God foreknows something to be future and that the thing does not turn out that way), and even if (ii) the antecedent is necessary in the sense in question (because it is past-tense and because no shadow of alteration can befall God), nonetheless the consequent can be purely contingent” (transl. Freddoso, Molina 1988, 189).7

A correct analysis of this passage is, however, more difficult than one might initially think. It is by no means clear whether Molina wants to reject a transfer principle of the form (Np, □(p → q) ├ Nq), or whether he rather intends to reject an inference of the form (Np, □(p → q) ├ □q) with respect to metaphysical necessity in the consequent. And also, in the same paragraph 34 Molina argues against this inference by counterfactual situations: “For if, as is possible, the opposite of the consequent were going to obtain, then the antecedent would never have obtained beforehand [...].” It is, however, not the topic of this paper to settle these exegetical questions. Jäger argues that Molina not only rejects the transfer principle as a matter of fact, he is also committed to this rejection to avoid inconsistencies. 7

“Tunc enim, esto condicionalis sit necessaria, quia in sensu composito cohaerere non possunt ista duo, quod Deus aliquid praesciat futurum et illud non eo modo eveniat, et esto antecedens illo modo sit necessarium, quia praeteritum et quia in Deum nulla possit cadere vicissitudinis obumbratio, nihilominus consequens potest esse mere contingens.”

How Molinists Can Have Their Cake and Eat It Too*

229

Jäger tries then to refute the transfer principle by making use of Ravizza’s overdetermination objection (cf. Ravizza 1994). A fact H is assumed to be caused by two independent events E1 and E2. Each of the two events is by itself sufficient for the truth of H. Now let E1 be a free choice of an agent S. With regard to the moral responsibility of S for the occurrence of H, the following inference is possible: (i) S is not responsible that E2; (ii) S is not responsible that (E2 → H); (iii) Thus: S is not responsible that H. However, (iii) is false. Since the inference from (i) and (ii) to (iii) relies on a βtransfer principle, and both premises are true, the transfer principle must be false. Transfer principles can be defended against this line of argument. Jäger himself offers an improved transfer principle for the case of moral responsibility which is immune to overdetermination objections (Jäger 2006, 14-19). The more general problem of these counterexamples to βprinciples, regardless of whether they are directed against moral responsibility or the no-choice operator, is this: If H follows deterministically and S is responsible for H, then compatibilism is already assumed. Assuming compatibilism in a refutation of a version of the Consequence Argument, an argument that is meant to defeat compatibilism, is obviously a petitio principii. If, however, H follows indeterministically, then this assumption of indeterminism in a line of reasoning that is – against a version of the Consequence Argument – ultimately meant to prove the compatibility of freedom and determinism is pointless (cf. Brüntrup 2000, 109). 3. Counterfactual power over middle knowledge A much more important problem in the general rejection of the β-rule seems to be that the Molinist would have to reject not only CAMol,2 or CAMol,1 but also CAComp along with it since in all those arguments the same structure of β-rules is at work (the objection made by Jäger stems from the debate about CAComp) – which is not admissible for the Molinist (as a libertarian), because the assumption of the validity of CAComp is a necessary (if not sufficient) condition for being a libertarian. If the Molinist, along with Jäger, rejects the β-rule’s validity it becomes difficult for him to maintain a libertarian position. So another way taken in the literature to refute the anti-Molinist Consequence Argument consists not in the rejection of the β-rule but in the repudiation of the premises, that is to deny the statement that we have no choice at all regarding the truth values of the counterfactuals of freedom

230

Godehard Brüntrup and Ruben Schneider

concerning us. Even if we are unable to exert any causal power on the knowledge of God and its contents (for this would deny the independence of God), we could still possess counterfactual “power” over it, since the scientia media’s counterfactual conditionals for their part have no causal influence on our actions and are only contingently true. Rather, they are explanatorily prior to our actions but logically and causally insufficient for them (cf. Morriston 2001; Perszyk 2003, 141): lacking causal power over a truth p is not the same as lacking counterfactual power over p. The Molinist has the option to allow this counterfactual power only over the counterfactuals of freedom: in CAComp he does not have to allow counterfactual power regarding the remote past P and the laws of nature L as compatibilists do, for the past and laws, contrary to the counterfactuals of freedom, are not subject-related. Indeed, Van Inwagen models his definition of determinism □[(P  L) → Future] explicitly in such a way that P is a state of the past before the existence of free subjects (cf. Schneider 2009, 23). This counterfactual power over counterfactual conditionals and, along with it, over the scientia media is explicitly conceded by Molina himself; he writes on middle knowledge: “[it] would be different in God if, as is possible, created free choice were by its innate freedom going to turn itself to the opposite part [...] ([...] aliter se haberet in Deo, si liberum arbitrium creatum pro sua innata libertate, ut potest, in oppositam partem foret inflectendum [...])” (Concordia 4.52.31, cf. 4.52.30, 32 and 34).8 This can be specified further (cf. Perszyk 2003, 140f.): if M0 is the state of God’s middle knowledge which contains the counterfactual C □→ A, and is counterfactually dependent on a decision on our behalf, then this means that the counterfactual power for the backtracking counterfactual A □→ M0 is true. But that also means that – if M0 holds – it is not possible for an agent to actually choose A (within the “physical state” of being created, the scientia media is virtually unavoidable for it precedes our temporal existence, cf. Ramelow 1997, 263) since this would entail the

8

Cf. also Concordia, 4.52.10: “For if created free choice were going to do the opposite, as indeed it can, then God would have known that very thing through this same type of knowledge, and not, what He in fact knows [Si namque liberum arbitrium creaturam acturum esset oppositum, ut revera potest, idipsum scivisset per eandem scientiam, non autem quod reipsa scit].”

How Molinists Can Have Their Cake and Eat It Too*

231

impossibility of counterfactual power over the counterfactual conditional, as can clearly be seen in the logic for counterfactuals by David Lewis: (A ‫ ר‬M0) → (A ◊→ M0) ↔ (A □→ M0) Within the logic of counterfactuals it is impossible for us to have counterfactual power over the scientia media (i. e. that the appropriate backtracker be true) and still, assuming M0 holds, to be able to act differently in the actual course of the world. The Molinist is therefore unable to claim both. To put it differently: if we have counterfactual power over the knowledge of God it is, under actualized circumstances C, logically if not actually possible to act in a different manner from the one which is contained “beforehand” in God’s middle knowledge. Our actual actions are accidentally but not metaphysically necessary. There are possible worlds containing actions that originated from free choices and the scientia naturalis contains (among other things) all possible worlds in which an individual P decides differently under the same circumstances C (reasons, dispositions etc.) than in the (assumedly) true counterfactual C □→ A, namely in choosing A. Therefore it is logically or metaphysically possible that there should be different courses of the world than those provided actually and contingently by the middle knowledge. 4. The individual essence as “notio completa” 4.1 Libertarianism without open future Now, does this mean that the Molinistic approach will slide into a form of compatibilism? For compatibilism also chooses a way out of CAComp using counterfactual power, counterfactual power over the laws of nature and the past in this case. But the crucial difference between a compatibilist and Molinist is that for the Molinist neither the past nor the laws of nature nor Divine knowledge influence the free choice causally in any way. For the Molinist the rational agent is the cause of his decisions completely within himself – he is an independent cause in the sense of agent causation. The Molinist therefore has to maintain CAComp along with the libertarian and at the same time has to reject CAMol. This means especially that the Molinist has to assume the β-rule to be valid and to set the crucial difference in the possibility of counterfactual power: while counterfactual power over the past and the laws of nature has to be rejected it has to be assumed regard-

232

Godehard Brüntrup and Ruben Schneider

ing Divine knowledge, which is of a completely different nature from the past and the laws of nature.9 Nevertheless there is a crucial difference between Molinism and the classical libertarian position: a possible world’s future regarding free actions may be causally open within the scientia media (sentence (2) of the initially introduced system of sentences) but still the future does not constitute a “garden of forking paths” (Kane 2005, 7) in a (possible or actual) creation presented by the scientia media and the scientia libera: The courses of action are decisively predefined. If God were to rewind time by some units to a point before the actual choice and would let it proceed again, the resulting decision would always be the same. So, if we analyse freedom in a libertarian sense we can discern two crucial aspects of freedom for the classical libertarian (cf. Morriston 2001, 30-31) of which the Molinist can accept only one: (I) Agent causation and autonomy: Nothing (else) caused the agent to act as he did. (II) Principle of alternate possibilities: Exactly as things were at the time of action, the agent could have refrained from doing the act. Molinism embraces Principle (I) as a crucial element of freedom against determinism (and the trilemma of freedom initially does not require more for a libertarian position). Principle (II), formulated as (F) or (F’) above, does not hold for the Molinist in the actual world. We are facing a libertarianism without an open future. This means that the definiteness of the course of decision is not synonymous with lawlike determination: The decisions originate in the individual himself and are not heteronomous but still follow an unambiguous path through time. Using “Frankfurt-style counterexamples,” Bergmann (2002) argues that the “Principle of alternate possibilities” is not necessary to ascribe created beings responsibility for their actions.10

9

Perzyk 2003 extensively discusses the problems of this distinction, but we cannot explore these in this paper. If Molinism is supposed to be a libertarian position, then it is inevitably is bound to this distinction. 10 Frankfurt-cases have been invoked explicitly against the tansfer-principles. As counterexamples they are only valid against β though, not against β* (cf. Brüntrup 2000, 110-111).

How Molinists Can Have Their Cake and Eat It Too*

233

4.2 Intentional preexistence and “notio completa” But how is this possible? Without venturing into the exegesis of the concept of freedom in Molina’s Concordia, it still is possible to trace a historical line of Molinist concept formation which renders some insight into our systematical question: For Molina within the counterfactual C □→ A there is no deterministic connection between C and A. From Molina’s point of view, the only connection between them is formed by the supercomprehension of God (cf. Concordia 4.49.11, 4.52.33): the indeterminateness of the freedom of choice is balanced by the superdeterminateness of Divine knowledge which infinitely transcends its object of knowing, and therefore “supercomprehends” it. The infinite power of representation of the Divine essence, by which God knows everything without requiring external information, and the infinite depth of His knowledge replace, as it were, the lacking determinateness of the human will (cf. Ramelow 1997, 56; Craig 1988, 178f.).11 From this concept of supercomprehensive knowledge we can draw a line to the Leibnizian term conceptus completus, or the individual idea of a free agent. This connection can be traced best by quoting the Molinist Hieronymus Fasolus S.J. (1568-1639) (cf. Ramelow 1997, 228): “According to Molina […] the free cause [...] cannot be known in a perfect manner, unless everything that is contained in this cause is also known, as well as everything what can possibly be caused by it, what has been caused by it, will be caused by it, and would be caused by it; for the effects too, and indeed all these effects, are something of this cause. So he who perfectly knows the cause, necessarily also knows these effects that depend on it in any respect whatsoever […]. But it is evident that this perfect knowledge must be infinite regarding future effects.” (In primum partem Summae D. Thomae Commentariorum, T.2, Lyon 1629, 269a, cf. 12 Knebel 1991, 3)

11

For the Molinist, this means a receptive, not a creative form of knowledge on this point. 12 “Mente Molinae [...] causa libera [...] non potest perfectissimo modo obiective cognosci, nisi simul cognoscantur et omnia, quae sunt in causa, et praeterea omnia quae ex causa vel esse possunt, vel erunt, vel sunt, vel fuerunt, vel essent; nam effectus etiam, atque adeo omnes isti effectus, sunt aliquid causae; ergo qui cognoscit perfectissimo modo causam, eius etiam effectus, quavis ratione ab ea pendentes, cognoscat necesse est [...]. Quod autem haec perfectissima cognitio respectu effectuum futurorum esse debeat infinita, patet.”

234

Godehard Brüntrup and Ruben Schneider

Even if he rejects the concept of a middle knowledge in God, Leibniz has a very similar intuition regarding the Divine foreknowledge: „God preserves our being and continually generates it, namely in such a way that we encounter thoughts spontaneously or freely in that order which is carried by the concept of our individual substance and in which it could be forseen from all eterni13 ty“ (Discours de métaphysique, §30).

Without having to follow Leibniz in his particular theory of individual ideas we can use the quotations by him and Fasolus as a backdrop for forming the following model: The individual essence14 of a free actor P pre-exists intentionally in the Divine mind logically prior to God’s act of creation (cf. Ramelow 1997, 222). For a formal exposition, let U be the set of possible conditions C, CU, and let A be a set of possible actions to be taken and M a set of subsets of A, {Ai : i  I}  M. Then, by His scientia naturalis, God knows which sets of actions are coherent with which conditions, formulated as a function from U to M: FP: U → M, FP(C) = {Ai : i  I} The actor P to be created should be free, therefore several choices among actions coherent with respect to conditions U are available for him. Formally, he can avail himself of several choice-functions ch: M → A. These choice-functions have sets as arguments and one element of the respective arguments as value: ch({Ai : i  I}) = AC, where AC  {Ai : i  I}. The actor to be created has a set of choice-functions CH at his disposal: in the strict logical sense there are several free choices consistent with the single circumstances and the actors individual essence. This also may be regarded as belonging to God’s scientia naturalis. By His scientia media, God knows which specific choice-function the actor to be created is going to select. So God knows not only the family of 13

For a precise reconstruction of the development of the Leibnizian concept of the individual idea or the “notio completa” within the Molinist debate from the perspective of the age cf. Ramelow 1997, 401-419. 14 Cf. Kvanvig 1986, 122-126.

How Molinists Can Have Their Cake and Eat It Too*

235

choice-functions CH available to the actor to be created but also which one the actor would select, He knows: ch*  CH, ch*: M → A, ch*(Ai : i  I) = AC*. In sum, by His middle knowledge, God knows the composed function SP* := ch*○FP : U → A, the first component being the choice of the actor to be created, and the second simply indicating which actions are coherent with which conditions (within a coherence framework – a possible world – W0. A world in which ch*○FP holds can thereby become actual, i.e. it is a “feasible” world, cf. Flint 1998, 51-54). All these functions, i.e. the conditional connections between C and the Ai, are pre-volitional for God. Not until God creates one (and for each point in time only one) C*U does He have any scientia libera. The scientia libera arises in evaluating the composed function ch*○FP at the created condition (the argument C*), ch*[FP(C*)]. So the free choices originate from the individual’s essence instead of receptively being the consequence of external causes. The composed functions SP* = ch*○FP present a “Logic of Freedom” inherent in possible beings. Their decisions intentionally anticipated by the Divine mind in the guise of the abstract and therefore trans-temporal essential functions SP* are the medium of knowledge (medium quo) by which God contemplates15 the counterfactuals of freedom from His eternal point of view. These functions also provide the truth-makers for counterfactuals demanded by the famous Grounding Objection against Molinism (cf. Flint 1998, 123-126; Hasker 1989; Adams 1977, 30). Since the functions are pre-volitional, the scientia media is not caused by an act of the Divine will, and since they are pre-existent, the actual decisions within the actual creation are also not the cause for the truth values of the counterfactuals (which would mean the aporetical external determination of the actus purus). 15

Speaking of distinct, finite functions within the Divine mind can of course only be understood modo nostro concipiendi. The Divine mind is not discursive but rather contemplates his “objects” of knowing.

236

Godehard Brüntrup and Ruben Schneider

4.3 Man as the image of God So creatures in actual reality cannot claim freedom of choice. Therefore, Molinism is able to radically construct man as the image of God, for in classical scholasticism God also does not have “alternative courses of action” within Himself in the strict sense: He does not change and is still free, even the source of all freedom. For Thomas Aquinas, God does not have freedom of choice regarding Himself since he recognises Himself as the absolute good and complete perfection, and therefore necessarily affirms Himself. He only has freedom of choice “ad extra” – what makes the Divine freedom of choice a relation of the perfect being toward imperfect beings, so it cannot be regarded as part of the actual, original freedom of God. Freedom of choice according to Aquinas can thus only be attributed to God in a secondary sense (cf. Schöndorf 2006, 654-655). For the classical Molinist too, God does not have freedom of choice in the actual meaning: the only freedom of choice that could occur within God Himself would be a freedom of the Divine hypostases toward each other by which they freely choose each other (cf. Schöndorf 2006, 658). But for the Molinist Luis Perez, for example, this is explicitly impossible since the Divine persons are logically prior than the choices of God. To prove his point, Perez explicitly refers to Molina himself (cf. Ramelow 1997, 252). The classical position also does not attribute freedom of choice to the angels and saints in heaven since they are facing the infinite good which is God and necessarily affirm it as their ultimate goal.16 True freedom is seen as the striving for the highest good, as well as its affirmation.17 The freedom of beings within the world also has its core in this affirmation of the good. But in a fallen world they are possessed by a receptiveness for the deviation from perfection resulting from original sin. Their morally wrong actions in the sinful maze of circumstances within this world are also contained within the functions of their essences as possibilities. Within the actual world freedom at its core thus consists in an independent self-realisation of the creature in an internal coherence with itself as 16

Tradition explains the fall of the angels by an initially obscured perception of the absolute perfection on their behalf which enabled them to sin by deprivation. 17 The rejection of the good in this case is synonymous with the loss of freedom, comparable to a drug addict losing his freedom by abandoning himself to the mechanism of evil - he becomes a “slave of sin.”

How Molinists Can Have Their Cake and Eat It Too*

237

well as in a comprehension of the inner logic of its being: “become what thou art.” We encounter freedom if the created being acts of its own accord in such a way that is coherent with itself. The created being independently and uncoercedly strives for realising its very own, complete form.18 5. Concluding remarks We have seen that the Molinist has to accept the consequence-argument CAComp against compatibilism regarding the determinateness of nature, if he wants to remain a libertarian, while he has to reject the consequencearguments CAMol against the “compatibility” of Gods foreknowledge and human freedom. But then he is unable to reject the validity of the principles of transfer that are at work in both consequence-arguments. The Molinist has to assume a counterfactual power over the (subject-related) contents of the middle knowledge, while rejecting such a counterfactual influence on the past and the laws of nature. Thus in the actual state of being created, it is only logically and not actually possible to act differently from what is known by the middle knowledge. So the Molinist, as we have seen, has to advocate a libertarianism without open future regarding the actual world: Libertarian freedom is conceived of in a weaker sense, as agent causation with an unambiguous course of action through time. However, this unambiguity is not synonymous with lawlike determination. The possibility of this is modeled best by an intuition shared by Molinists from the time between Molina and Leibniz: within the Divine mind (and only within His mind) there pre-exists an individual idea of a free creature P which contains all of its possible actions, as well as true counter- or prefactual conditionals in the form of the function SP*. The truth values of these conditionals cannot be ascribed with metaphysical necessity. This function is the counterfactuals’ truth-maker before the act of creation, thereby refuting the notorious grounding objection against Molinism. But this raises the question of a much deeper grounding-objection which does not relate to the question of truth-makers: Where do the functions SP* originate? Are they created by God? Then the scientia media is not prevolitional. Or is it something that confronts God in a relative form of autonomy from all eternity as the Possibilia in the thought of Suárez do (cf. 18

For a criticism of this point cf. again Schöndorf 2006.

238

Godehard Brüntrup and Ruben Schneider

Mahieu 1921, 229-232)? But then, who defines that they are exactly as they are? Are they brute facts? William Lane Craig suggests that a further grounding of the truth values of the counterfactuals of freedom would make freedom impossible (cf. Craig 2001; Davidson 2004, 368), and the same goes for the essential functions SP*; they would simply form the fundamental, irreducibly basic elements in the architectonics of Molinistic theory that are founded in the supercomprehension of God, in the infinite depth of His knowledge. This research was supported by a grant of the John Templeton Foundation, Philadelphia, USA: Analytic Theology Project, Cluster Initiative “Divine Omnipotence, Human Freedom and the Problem of Evil – Towards an Analytic Concept of Theodicy”. We wish to express our gratitude to Christina Schneider, LMU, Munich.

References Adams, Robert 1977 “Middle Knowledge and the Problem of Evil”, American Philosophical Quarterly 14, 109-117. Bergmann, Michael 2002 “Molinist Frankfurt-style Counterexamples and Free Will Defense”, Faith and Philosophy 19, 462-478. Brüntrup, Godehard 2000 „Der Metaphysische Begriff der Willensfreiheit und das Transferprinzip des Keine-Wahl-Habens“, in Peres, C. and Greimann, D. (eds.), Wahrheit – Sein – Struktur. Auseinandersetzungen mit Metaphysik, New York – Zürich: Olms, 102-119. Craig, William L. 1988 The Problem of Divine Knowledge and Future Contingents from Aristotle to Suárez, Leiden: Brill. -

2001 “Middle Knowledge, Truth-makers and the ‘Grounding Objection’”, Faith and Philosophy 18, 337-352.

Davidson, Scott A. 2004 “Craig on the Grounding Objection to Middle Knowledge”, Faith and Philosophy 21, 365-369. Flint, Thomas P. 1998 Divine Providence. The Molinist Account, Ithaca: Cornell University Press. Freddoso, Alfred J. 1988 “Introduction”, in Luis de Molina, On Divine Foreknowledge (Part IV of the Concordia), transl. by Alfred J. Freddoso, Ithaca: Cornell University Press. Hasker, William 1989 God, Time and Knowledge, Ithaca: Cornell University Press. Jäger, Christoph 2006 „Drei Konsequenzargumente für eine inkompatibilistische Theorie moralischer Verantwortung“, Zeitschrift für philosophische Forschung 60, 504-527.

How Molinists Can Have Their Cake and Eat It Too*

239

-

2010 “Göttlicher Plan und menschliche Freiheit. Vorhersehung und ‚Mittleres Wissen‘ bei Luis de Molina”, in Philosophisches Jahrbuch 117, 299-318.

-

2011 “Molina on Foreknowledge and Transfer of Necessities”, in Tapp, Chr. and Runggaldier, E. (eds.), God, Eternity, and Time, Farnham/Burlington: Ashgate, 81-98.

Kane, Robert 2005 A Contemporary Introduction to Free Will, New York – Oxford: Oxford University Press. Knebel, Sven 1991 “Necessitas moralis ad optimum. Zum historischen Hintergrund der Wahl der besten aller möglichen Welten”, Studia Leibnitiana 23, 3-24. Kvanvig, Jonathan 1986 The Possibility of an All-Knowing God, London: Macmillan Press. Lewis, David 1973 Counterfactuals, Oxford: Blackwell. Mahieu, Léon 1921 Francois Suarez, sa philosophie et les rapports qu’elle a avec sa théologie, Paris: Desclée de Brouwer. Molina, Luis de 1988 On Divine Foreknowledge (Part IV of the Concordia), transl. by Alfred J. Freddoso, Ithaca: Cornell University Press. Morriston, Wes 2001 “Explanatory Priority and the „Counterfactuals of Freedom”, Faith and Philosophy 18, 21-35. Perszyk, Kenneth 2003 “Molinism and the Consequence Argument – a challenge”, Faith and Philosophy 20, 131–151. Ramelow, Tilman 1997 Gott, Freiheit, Weltenwahl. Die Metaphysik der Willensfreiheit zwischen A. Perez S.J. (1599–1649) und G.W. Leibniz (1646– 1716), Leiden: Brill. Ravizza, Mark 1994 “Semi-Compatibilism and the Transfer of Nonresponsibility”, Philosophical Studies 75, 61-93. Rhoda, Alan 2007 “The Philosophical Case for Open Theism”, Philosophia 35, 301311. Rugel, Matthias 2006 Handlung und Freiheit. Analysen zur Akteursverursachung, München: Philosophia. Schneider, Christina 2009 Metaphysische Freiheit – Kohärenz und Theorie, München: Philosophia. Schöndorf, Harald 2006 „Ist Willensfreiheit immer Wahlfreiheit?“, in: Philosophiae et Musicae, Liber iubilaris Professori Stanislao Ziemianski SJ septuagesimum quintum annum celebranti dicatus, Darowski, R. (ed.), Cracow: Ignatianum, 647-658. Van Inwagen, Peter 1983, An Essay on Free Will, Oxford: Clarendon Press.

Scientia Media and Freedom to Do Otherwise CHRISTOPH JÄGER, Innsbruck

1. Introduction One of Edmund Runggaldier’s leading intellectual concerns is the philosophical reconstruction, and rational assessment, of classical Christian metaphysics. This enterprise stands in the great tradition of Jesuit philosophy, which to the present day continues to enrich philosophical and theological debates with insightful and important contributions. Among the key topics of Christian metaphysics is the free will problem. More precisely, one of the most intriguing problems that metaphysicians and philosophers of religion have wrestled with since antiquity is the relation between divine omniscience, causal determinism, and human free will. One of the most ingenious solutions to this problem ever developed in the history of philosophy has been proposed by one of Runggaldier’s fellow Jesuits, the 16th century philosopher and theologian Luis de Molina. In his main work, Liberi Arbitrii Cum Gratiae Donis, Divina Praescientia, Providentia, Praedestinatione et Reprobatione Concordia (henceforth for short: Concordia), Molina extensively discusses, and dismisses, theological determinism. This is the view that, since the world is governed by divine providence and omniscience, no human action or decision is free. Molina emphatically rejects this view and argues for a position I call theological compatibilism. This is the claim that divine providence and omniscience are compatible with human free will. Theological incompatibilism is the negation of theological compatibilism. However, while Molina rejects theological incompatibilism he endorses nomological incompatibilism. This is the view that if an action or decision is the outcome of laws of nature and prior states of the universe on which the agent has no causal influence, then that action or decision is not free. Nomological compatibilism is the negation of nomological incompatibilism. In this paper I examine whether it is coherent for Molina to endorse both theological compatibilism and nomological incompatibilism.

242

Christoph Jäger

In two recent papers I have defended Molina’s solution to the theological free will problem, arguing that it essentially rests on his rejection of a modal closure principle concerning “causal impact necessity” (CInecessity; cf. Jäger 2010 and 2011). I characterized this notion as follows: (CI-Necessity)A metaphysically contingent state of affairs (or proposition) p is causal-impact-necessary for a given subject X at a time t iff it is not within X’s power at t to contribute causally to something that constitutes, or grounds, a necessary or sufficient condition for p’s obtaining (or the truth of p).1 Talk about “contributing causally” to something is meant here to include “full causation”, in the sense that no other cause is involved. Talk about “constituting or grounding” is meant to preserve neutrality towards particular theories of causation. If causation is a relation between events, we may say that people causally contribute to something that grounds necessary or sufficient conditions for some fact or state of affairs. If causation is conceived as a relation between states of affairs, then we can say that people causally contribute to states of affairs that constitute necessary or sufficient conditions for other states of affairs. Intuitively, an entity is CI-necessary for a subject at a given time iff it is causally completely inaccessible to that subject at that time. CI-necessity is a time-relativized kind of necessity. For example, it is CI-necessary for you now that Caesar crossed the Rubicon in 49 B.C. What Caesar did in 49 B.C. might not have been CI-necessary for you had you lived in 49 B.C. In that case you might have had the chance to affect the course of events in such a way that Caesar would not have crossed the Rubicon. However, it is not only past states of affairs which are CInecessary (for a given subject at a given time). For example, among the things that are causally untouchable in this sense are also laws of nature. It is CI-necessary for every actual human being that nothing travels faster than light. Applied to such cases, the reference to times can be omitted.

1

For present purposes, I shall ignore (potential) metaphysical distinctions between propositions and states of affairs.

Scientia Media and Freedom to Do Otherwise

243

The closure or transfer principle concerning causal-impact necessity that Molina rejects, according to my argument2, says that: (P)

If it is CI-necessary for a given subject X at t that p, and p entails q, then it is CI-necessary for X at t that q; for short: NX,tp, □(pq) |– NX,tq3

Molina, I claim, must reject (P) in order to block arguments for theological incompatibilism. Consider God’s eternal omniscience before and after creation. According to the Molinist, after having actualized one of the many possible worlds He could have actualized, God has “postvolitional” eternal knowledge of every event in the universe. Assume with Molina (and other medieval Thomists) that God exists, and hence that His knowl2

The relevant passage is this: “Tunc enim, esto conditionalis sit necessaria, quia in sensu composito cohaerere non possunt ista duo, quod Deus aliquid praesciat futurum et illud non eo modo eveniat, et esto antecedens illo modo sit necessarium, quia praeteritum et quia in Deum nulla possit cadere vicissitudinis obumbratio, nihilominus consequens potest esse mere contingens.” (“For in such a case, even if (i) the conditional is necessary (because in the composed sense these two things cannot both obtain, namely, that God foreknows something to be future and that the thing does not turn out that way), and even if (ii) the antecedent is necessary in the sense in question (because it is past-tense and because no shadow of alteration can befall God), nonetheless the consequent can be purely contingent. (Molina, Concordia, 52, 4, 34, p. 189/p. 353) Page numbers refer, in this order, to Rabeneck’s Latin edition and to Freddoso’s translation (which I use here and elsewhere in this paper; see Freddoso, ed., 1988.) My interpretation of this passage in terms of CI-necessity departs from standard interpretations according to which Molina rejects the closure of what William of Sherwood and Ockham called “accidental necessity” (see William of Sherwood, Introductiones, 11; Freddoso 1988; Zagzebski 1991; Zagzebski 2002). Roughly, a contingent state of affairs (or proposition) is accidentally necessary at a given time if it cannot be made not to obtain (or cannot be rendered false) at that time because it (or the state of the world it describes) is already past. My chief reason for rejecting the accidentalnecessity interpretation of Molina’s reflections is that his account of the divine stands in the Boethius-Augustine-Aquinas tradition and maintains that God exists in extratemporal eternity. If so, Molina’s real opponent is not the critic who argues that God “foresaw” every human action at some time in the past, and who thinks that it somehow follows from this that no human action is free. To this charge the Molinist can respond that an extratemporal God does not have knowledge at times (see Jäger 2010 and 2011). 3 The principle is to be read in such a way that, if it is CI-necessary for S that p (q), then p (q) is the case.

244

Christoph Jäger

edge occurs, in extratemporal eternity. Let S be an arbitrary human subject, and A an action S performs at a given time T. Given (P), Molina’s opponent argues as follows: A. The objector’s argument in favour of theological determinism (1) (2) (3)

At any time in S’s life, it is CI-necessary for S that God know (postvolitionally, and from extratemporal eternity) that S does A at T (for S never has any causal access to God). Necessarily: If God knows (postvolitionally, and from extratemporal eternity) that S does A at T, then S does A at T. Hence, at any time in S’s life, it is CI-necessary for S that S do A at T.

(3) says that at no time in his life does S have any causal influence upon the fact that he does A at T. Is such an action free? Molinists are causal or nomological incompatibilists. That is, they deny that an action that is fully determined by causes outside the agent is a free action. In other words, they endorse the following requirement for freedom: (RF) An agent X performs an action freely at a time t only if the fact that it occurs is not CI-necessary for X at t (i.e., only if it is within X’s power at t to contribute causally to something that constitutes, or grounds, a necessary or sufficient condition for the fact that the action takes place). With (RF) it follows from (3) that: (4)

S does not do A freely at T.

So, this argument says that God’s eternal knowledge of our actions destroys our causal impact on those actions, and hence our freedom with respect to them. Where, if anywhere, can the Molinist find fault with this argument? He accepts (RF). If it is a necessary condition for a libertarian free action that the subject have at least some causal influence on the occurrence of that action, then (RF) is true. Second, it is an essential tenet of Molinist theology that no human being can have any causal impact upon anything that happens in the realm of the divine. Besides, how could spatio-temporal beings causally affect anything in the realm of the extratemporal? So prem-

Scientia Media and Freedom to Do Otherwise

245

ise (1) cannot reasonably be disputed either. Premise (2) is true as well. What it says is a direct consequence of the concept of knowledge. Nihil potest sciri nisi verum, no special theological doctrine needs to be invoked to argue for that claim. So what remains to be examined is the intermediate conclusion (3). How does (3) follow from (1) and (2)? In order to derive (3) we need (P). Hence rejecting (P) blocks argument A because without (P) we cannot derive (3). Precisely this, I claim, is an essential part of Molina’s defence of theological compatibilism: since (P) is false, he maintains, arguments such as Argument A do not get off the ground. The Molinist can present a similar argument in defence of his famous notion of scientia media (which I will discuss in the next section). Moreover, I think that he is right in rejecting (P). Although Molina himself does not present much of an argument against (P), there are counterexamples to this principle, two of which I presented in Jäger (2010) and (2011). They draw on cases of causal overdetermination. An event E on which S has no causal impact deterministically brings about some other event E*. Still, S does have causal impact on E* because S causally brings about a sufficient condition, or something that grounds a sufficient condition, for E*. If stories of this kind are coherent, they show that (P) is false. Below I will discuss in detail a counterexample with this structure. First however let us consider an alleged problem with this reasoning. In their paper “How Molinists can have their cake and eat it too” (2011, this volume) Godehard Brüntrup and Ruben Schneider raise two interesting objections to my argument. First, drawing on observations in Brüntrup (2000), they argue that my (alleged) counterexamples to (P) are either guilty of a petitio principii or pointless. Second, they maintain that my defence of (this aspect of) classical Molinism faces the following problem. Being a libertarian, Brüntrup and Schneider claim, Molina is committed to accepting a van Inwagen-style consequence argument for the incompatibility of free will with nomological determinism. Yet such a consequence argument, the charge goes, employs a modal closure principle equivalent with, or sufficiently analogous to, (P). Moreover, incompatibilism is an essential ingredient of libertarianism. So “if the Molinist, along with Jäger, rejects the ... rule’s validity it becomes difficult for him to maintain a libertarian position” (Brüntrup and Schneider 2011, 229). To sum up, the objection is that, contrary to what I argue, Molina cannot in fact coherently reject (P) in order to (dis)solve problems of theological determinism. For (P) – or some closely related principle – essentially figures in the consequence

246

Christoph Jäger

argument for nomological incompatibilism and is therefore an indispensable building block of Molina’s libertarianism. These are interesting objections, and I am grateful to Brüntrup and Schneider for raising them. However, on closer inspection it turns out that, although they are well motivated, both objections fail. Neither do my counterexamples against (P) commit a petitio principii, nor are they pointless, nor does Molina’s rejection of (P) jeopardize his libertarianism. The reason is this. First, contrary to what Brüntrup and Schneider claim, acceptance of a van Inwagen-style consequence argument is dispensable for the libertarian. Second, even if it were not, such consequence arguments for nomological incompatibilism do not employ (P). Instead, they work with a principle concerning what van Inwagen calls “lack of choice”. This operator must not be conflated with causal impact necessity. There are states of affairs which, although they are CI-necessary because we lack causal access to them, would still have been different had we chosen to act differently than we actually did. In other words, there are states of affairs that are CI-necessary for a given agent at a given time but for which the agent still has, or once had, a (non-causal kind of) choice. Somewhat ironically this seems to be quite in line with Brüntrup and Schneider’s constructive proposal for solving the problem of theological determinism. According to this proposal, humans can have “counterfactual power” over what God knows about their free actions. Were we to act differently from how we actually do, God would know that we act in these different ways. So in a sense we do have a choice about what God knows about our actions, even though we do not have any causal choice about anything He knows or believes. However, contrary to what Brüntrup and Schneider maintain it follows from this very point that, since lack of choice is distinct from CI-necessity, the Molinist can coherently deny (P) while accepting that having no choice is closed under entailment. In what follows I shall clarify my Molinist response to the problem of theological determinism by showing that Brüntrup and Schneider’s objections fail.4 I will skip over interpretative questions about whether the his4

I am well aware that the dialectic of my argument puts me in a somewhat delicate position. In honour of Edmund Runggaldier, S.J., I lay out my defence of theological indeterminism cum nomological incompatibilism in the spirit of Luis de Molina, S.J. I do so partly by drawing on examples which, more recently, Mark Ravizza, S.J., has

Scientia Media and Freedom to Do Otherwise

247

torical Molina actually meant to reject (P).5 Here I will mainly focus on my argument that (P) is indeed false, and discuss Brüntrup and Schneider’s systematic worry that if the Molinist rejects (P) he undermines his libertarianism. 2. Scientia media and the closure of CI-necessity under entailment Before examining (P) and its relation to the consequence argument, let us (i) clear away one potential worry regarding the present way of discussing the problem and (ii) see how (P) figures in a second argument against Molinism, i.e., in an argument that questions the consistency of the notion of scientia media. Like Molina’s work on these topics, Runggaldier’s philosophy of action is heavily indebted to Aristotelian metaphysics. Runggaldier also acknowledges traditions in analytic philosophy of language and philosophy of science according to which causal explanations of physical events must not be conflated with intentional or “personal” explanations of human action (Runggaldier 1996, 16-20). However, while he concedes that such distinctions may be philosophically helpful, he also argues that we should resist the temptation to postulate corresponding dichotomies at the ontological level. There is only one reality. Hence we should expect there to be systematic connections between “mental” and “physical” discourse; questions and problems formulated in one of these “language games” should be translatable into the other (cf. Runggaldier 1996, 25). If so, the libertarian cannot escape the free will problem simply by claiming that statements about causally deterministic processes cannot raise problems for claims about (free) human action. Molina concurs. He also agrees with Runggaldier that analogous reasoning applies to “religious language” or, more specifically, to language about the divine, as opposed to the language in put forth against a famous argument for nomological incompatibilism. Moreover, I will argue that the objections against my account levelled by Godehard Brüntrup, S.J., and Ruben Schneider are mistaken, and that their proposal is not in fact a viable alternative for the Molinist. The present contribution is an attempt to clarify the dialectic of the debate and to outline what I take to be the most promising Molinist solution to the problem of theological determinism. 5 I do believe this, however, and extensively argue for this claim in Jäger (2010) and (2011).

248

Christoph Jäger

which we describe and analyse human action. Here, too, we should not think that we can circumvent potential problems facing theological compatibilism simply by arguing that statements and principles formulated in one kind of discourse have no bearing on issues formulated in the other. Argument A attacks theological compatibilism and focuses on postvolitional divine knowledge. Another argument that employs (P) can be levelled against Molina’s famous notion of scientia media. According to Molina, God not only knows postvolitionally what every actual human creature freely does at any point in his or her life. He also knows “prevolitionally”, i.e., (logically) before He has chosen to actualize a particular world, what every possible free creature would freely do in any possible circumstance in which He might place it. Counterfactuals describing such states of affairs are called “counterfactuals of freedom”. (Note however that the name is somewhat misleading. Some “counterfactuals of freedom” describe actions which subjects who will be actual do in fact perform at the respective points in their lives. For such cases, the title “prefactual of freedom” might perhaps be more accurate, at least if we could avoid temporal associations.) Obviously, no human being can have, or could have had, any causal influence on God’s middle knowledge. Before creation no human being even existed. So if the Molinist accepts (P), the objector argues, he must also accept the following argument. Let S be an arbitrary human subject, and A an action which S performs at a given time T: B. The objector’s argument that the Molinist notion of scientia media is inconsistent6 (1)

(2)

(3)

6

At any time in S’s life it is CI-necessary for S that God know (prevolitionally, and from extratemporal eternity) via middle knowledge that S, when placed in circumstance C at T, would freely do A. God’s knowing (prevolitionally, and from extratemporal eternity) via middle knowledge that S, when placed in circumstance C at T, would freely do A, entails that S, if placed in C at T, freely does A. Hence, it is CI-necessary for S that, when placed in C at T, S freely does A.

This corresponds to Argument III in (Jäger 2011).

Scientia Media and Freedom to Do Otherwise

249

Next, consider once more (RF), which rules that an action is performed freely only if it is not CI-necessary. Applying this requirement to (3), we get a contradiction. (3) tells us that S, when put in C, freely does A and yet has no causal influence upon doing so. (RF) rules out that such an action is free. Hence, applying (RF) to (3) we get the self-contradictory statement that: (4)

S is not free in doing A freely when placed in C at T.

Argument A aims to show that Molinism must (incoherently) accept theological determinism, and Argument B maintains that middle knowledge is an inconsistent notion. For reasons similar to those already given in our discussion of Argument A, the Molinist cannot dismiss this argument by rejecting premise (1) or premise (2). Nor can he reject (RF). (4) follows from (3) and (RF). So his only option is to reject (3). (3) follows from (1) and (2) by a certain inference rule, so the Molinist must reject that inference rule. And as it turns out, this rule is identical with the rule that licenses the deduction of (3) from (1) and (2) in Argument A. So once more, what the Molinist must reject is principle (P). By rejecting (P), I maintain, the Molinist can in one elegant fell swoop repudiate both Argument A and Argument B, thereby saving his libertarian reconciliation of divine providence and omniscience with free human action from two crucial objections. So far my argument was that the Molinist’s only option for responding to Argument A and Argument B is to reject (P). However, is (P) really false? I think it is. In Jäger (2011) I presented two counterexamples. Both are inspired by an example from Mark Ravizza (1994) which Eleonore Stump and John Martin Fischer (2000) have developed further. My version of one of the stories goes as follows: Suppose that, necessarily, if the actual laws of nature L and certain conditions C of the world obtain at T2, there will be an avalanche that destroys a camp in the mountains at T3. Let T2 occur some time just before T3, and let C be conditions on which no one has, or ever had, any causal impact. Suppose also that at T1 (some time shortly before T2) Betty detonates explosives, thereby starting an avalanche that is causally sufficient to destroy

250

Christoph Jäger

the camp at T3 as well. Finally, suppose that Betty’s starting “her” avalanche is the result of some suitable indeterministic process.7 In this example it holds that (i) at T1 Betty neither has nor ever had any causal impact on the laws of nature. Nor does she have, or did she ever have, any causal impact on the fact that the conditions of the world at T2 are C. These facts are, in the above terminology, CI-necessary for Betty at T1. Moreover, it holds ex hypothesi that (ii) necessarily, if the laws of nature obtain and the condition of the world at T2 is C, then there will be an avalanche that destroys the camp at T3. Although (i) and (ii) are true, however, it is not true that (iii) at T1 Betty neither has nor has ever had any causal impact on the fact that there is an avalanche that destroys the camp at T3. After all, she also brings about some causally sufficient condition for there being an avalanche that destroys the camp at T3. Because the destruction of the camp is causally overdetermined in this way, we have a counterexample to the principle that if NX,tp and □(pq), then NX,tq. Brüntrup and Schneider’s first complaint is that this counterexample either contains a petitio principii or is “pointless”. They write: Jäger tries ... to refute the transfer principle by making use of Ravizza’s overdetermination objection ... A fact H is assumed to be caused by two independent events E1 and E2. Each of the two events is by itself sufficient for the truth of H. Now let E1 be a free choice of an agent S. With regard to the moral responsibility of S for the occurrence of H, the following inference is possible: (i) S is not responsible that E2 [occurs, C.J.]; (ii) S is not responsible that (E2H). (iii) Thus: S is not responsible that H. However, (iii) is false. Since the inference from (i) and (ii) to (iii) relies on a β-transfer principle, and both premises are true, the transfer principle must be false. ... The ... general problem of these counterexamples to β-principles, regardless of whether they are directed against moral responsibility or the no-choice operator, is this: If H follows deterministically and S is responsible for H, then compatibilism is already assumed. Assuming compatibilism in a refutation of a version of the Consequence Argument, an argument that is meant to defeat compatibilism, is obviously a petitio principii. If, however, H follows indeterministically, then this assumption of indeterminism in a line of reasoning that is – against a version of the Consequence Argument – ultimately meant to prove the compatibility of freedom and determinism is pointless. (Brüntrup and Schneider 2011, 229)

This is an interesting thought. Yet it is misguided in several ways. (I shall ignore minor points, such as “categorical” infelicities concerning the (mis-)characterization of Ravizza-type examples as maintaining that events 7

Cf. Jäger (2011), 94f.; Stump and Fischer (2000), 49.

Scientia Media and Freedom to Do Otherwise

251

cause facts.) To begin with, note that my story about Betty employs neither a no-choice operator nor a moral responsibility operator. The story is explicitly presented as a counterexample to the claim that causal-impact necessity is closed under entailment. That’s an important difference. Brüntrup and Schneider do not explain their petitio-or-pointless argument any further. However, even if that argument were on target with respect to lack of choice and lack of moral responsibility, it does not follow that what holds for these notions also holds for causal-impact necessity. My example would commit the kind of petitio Brüntrup and Schneider describe if it assumed nomological or causal compatibilism, i.e., the view that an action that is nomologically or causally determined can still be free, or that such an action can be one for which its subject is morally responsible. However, what the example assumes is that Betty, by doing something that is causally undetermined, becomes (in part) causally responsible for an event that would in any case be caused by a deterministic process. So, Betty’s detonating her explosives is explicitly stipulated to be an undetermined action, and the kind of responsibility involved is causal, not moral. Hence the example clearly does not assume compatibilism. Nor is it “pointless” for the cases at hand. Brüntrup and Schneider’s idea seems to be this. Suppose that the laws of nature and the condition of the world at T2 indeterministically cause the camp’s destruction. Then my story would not constitute an instantiation of (P), because it would not hold that, necessarily, (L & C) entails that the camp is being destroyed. However, it is unclear to me why the authors think that this could be an objection against my argument. The example explicitly says – also when I present it in Jäger (2011) – that, necessarily, if the actual laws of nature and certain conditions C of the world obtain at T2, then there will be an avalanche that destroys the camp at T3. While Betty’s action is claimed to be undetermined, the process leading from L and C to the destruction of the camp is deterministic. It is precisely for this reason that the story constitutes a counterexample to (P) without either committing a petitio or being pointless. What may raise an eyebrow here is the assumption of a universe with both deterministic and indeterministic processes. However, this assumption is perfectly consistent too and, I believe, not even metaphysically extravagant. As Stump and Fischer explain with respect to one of their moralresponsibility versions of the example, “even in an indeterministic world, some events and states of affairs can be causally determined. ... Even in an

252

Christoph Jäger

indeterministic world there can be ‘pockets of local determination’ ” (Stump and Fischer 2000, 50). I conclude, therefore, that Brüntrup and Schneider’s first charge is mistaken. It is not even clear what exactly the problem is supposed to be for Ravizza-style stories concerning moral responsibility. But in any case the objection does not address my CInecessity version of the Ravizza-Stump-Fischer example. Brüntrup and Schneider’s second complaint, which they regard as even more important, is that Molinists cannot reject (P) without undermining their libertarianism. For (P), they claim, resembles van Inwagen’s “rule β”. This rule plays an essential role in his famous consequence argument for (nomological) incompatibilism, which is an essential ingredient in libertarianism. Now, Brüntrup and Schneider think that as a libertarian (and hence as an incompatibilist) the Molinist must accept a van Inwagen-style consequence argument. The alleged problem then is that, when the Molinist rejects (P), the traditional consequence argument for nomological incompatibilism goes by the board as well, thereby depriving him of a rational basis for libertarianism. “It is our contention”, Brüntrup and Schneider write, “that if Molinism is to be a libertarian position, the Molinist had better consider the Consequence Argument a valid argument”: The conclusion of the Consequence Argument is that nobody has a free choice about anything, assuming that we live in a deterministic world … and [that, C.J.] the past and the laws of nature cannot be changed. We will label this argument “CAComp” (cf. van Inwagen 1983, 96-98, 147). The problem of Divine foreknowledge and human freedom is logically analogous to the problem of freedom and determinism. It is of considerable significance in this context that some arguments against Molinism have the logical structure of the Consequence Argument against compatibilism. (Brüntrup and Schneider 2011, 225)

My argument, the authors argue, has the following crucial drawback: A[n] … important problem in the general rejection of the β-rule seems to be that the Molinist would have to reject not only ... [arguments against middle knowledge, C.J.] but also CAComp ... since in all those arguments the same structure of β-rules is at work (the objection made by Jäger stems from the debate about CAComp) – which is not admissible for the Molinist (as a libertarian), because the assumption of the validity of CAComp is a necessary (if not sufficient) condition for being a libertarian. If the Molinist, along with Jäger, rejects the β-rule’s validity it becomes difficult for him to maintain a libertarian position. (Brüntrup and Schneider 2011, 229)

Scientia Media and Freedom to Do Otherwise

253

Brüntrup and Schneider then opt for an alternative solution to the problem of theological determinism according to which we enjoy what they call “counterfactual power” over divine knowledge and its contents: Another way ... to refute the anti-Molinist Consequence Argument [= Argument B, C.J.] consists not in the rejection of the β-rule but in the repudiation of the premises, that is to deny the statement that we have no choice at all regarding the truth values of the counterfactuals of freedom concerning us. (Brüntrup and Schneider 2011, 229-230)

Again, various aspects of this argument require clarification. I will turn to them shortly. So much however seems clear: Brüntrup and Schneider’s main conclusion is that the Molinist – contrary to my suggestion – had better not reject (P) but should instead opt for an alternative response to Argument B (and Argument A). It will help to have a schematic version of Brüntrup and Schneider’s reasoning before us. It runs as follows: C. Brüntrup and Schneider’s argument (1)

(2)

(3)

If the Molinist rejects the inference rule employed in Argument A and Argument B, he thereby either rejects, or is committed to rejecting, “the β-rule”, which essentially figures in classical (van Inwagen-style) consequence arguments for nomological incompatibilism. Accepting a consequence argument for nomological incompatibilism is a necessary condition for (rationally) accepting nomological incompatibilism, and hence for (rationally) being a libertarian. Therefore, if the Molinist rejects the inference rule employed in Argument A and Argument B, he cannot (rationally) maintain libertarianism.

However, since Molinists wish to – and should – maintain (a rational form of) libertarianism, the overall conclusion is that: (4)

The Molinist cannot coherently reject the inference rule employed in Argument A and Argument B.

What are we to say about this argument? To begin with I would like to point out that the main conclusion of this argument, (4), cannot be right. The idea that the Molinist can coherently accept (P) is clearly mistaken. Witness Argument A and Argument B. As we have seen, in neither argu-

254

Christoph Jäger

ment is it open to the Molinist to reject any of the premises. Premises A (1) and B (1) are direct consequences of the Molinist view that no human being can have any causal impact upon anything that happens in the realm of the divine. Brüntrup and Schneider themselves do not deny that “we are unable to exert any causal power on the knowledge of God and its contents (for this would deny the independence of God)” (Brüntrup and Schneider 2011, 230). If so, premises A (1) and B (1) are not open to dispute. Premises A (2) and B (2) are direct consequences of the concept of knowledge. Finally, (RF) – the claim that X performs an action freely only if the latter is not CI-necessary for X – is an integral part of the Molinist’s libertarianism. Brüntrup and Schneider do not dispute this either. “For the Molinist”, they write, “the rational agent is the cause of his decisions” (2011, 231) “Nothing ... [outside the agent, C.J.] caused the agent to act as he did” (2011, 232). If it is a necessary condition for a libertarian free action that the subject have at least some causal influence on the occurrence of that action, (RF) is true as well. A (4) and B (4) however follow via (RF) from A (3) and B (3), respectively. It follows that the only steps in Argument A and Argument B that Molinists, qua libertarians, can coherently reject are the inferences from (1) and (2) to (3). This means that the Molinist must – contrary to what Brüntrup and Schneider maintain – reject (P). So where exactly does Argument C go astray? Consider premise (2) of Brüntrup and Schneider’s reasoning, i.e., the claim that if one rejects the inference rule invoked in Argument B and Argument A one cannot (rationally) endorse libertarianism, because if one rejects that rule then the consequence argument for incompatibilism collapses. The first thing to note in response is that it is highly controversial that the consequence argument really is “a necessary (if not sufficient) condition for being a libertarian” (Brüntrup and Schneider 2011, 229). It is certainly not a sufficient condition. Libertarianism is the conjunction of (nomological) incompatibilism with the view that free will exists (and hence that causal determinism is false). Hence libertarians must supplement any argument that aims merely to show that incompatibilism is true with arguments to the effect that determinism is false. Only this would yield a full argument for libertarianism. Many incompatibilists doubt that there are successful arguments for indeterminism. Hence they accept, not libertarianism, but “hard determinism”, i.e., the view that determinism is incompatible with free will, but that determinism is true and hence that free will does not exist.

Scientia Media and Freedom to Do Otherwise

255

Maybe what the authors mean is thus that acceptance of a van Inwagenstyle consequence argument is “a necessary (if not sufficient) condition” for being an incompatibilist. But this claim is false, or at least highly problematic and controversial, as well. It may be true that in recent decades many incompatibilists (and libertarians) in fact accepted some version of a van Inwagen-style consequence argument. However, not all incompatibilists are “consequence incompatibilists”. For example, Eleonore Stump, Ted Warfield, and others do not endorse the consequence argument.8 Pressed by objections against the consequence argument, Warfield (2000) has proposed “a new argument for incompatibilism” which, he maintains, “does not employ a ‘transfer of necessity principle’ and so is not simply one more version of the Consequence argument” (Warfield 2000, 168). A discussion of Warfield’s new argument for incompatibilism would extend beyond the scope of this paper. But in any case doing so is unnecessary, for the main problem with Brüntrup and Schneider’s argument is that premise (1) is false. Following Brüntrup and Schneider’s wording, premise (1) in the above reconstruction of their argument says that, if the Molinist rejects the inference rule employed in Argument A and B, he rejects, or is committed to rejecting, “the β-rule”. But when the authors talk about an important problem in “the general rejection of the β-rule” (see the above quotations); when they write about rejecting “the β-rule” and say that “if the Molinist, along with Jäger, rejects the β-rule’s validity it becomes difficult for him to maintain a libertarian position”, what exactly do they mean? As extensively laid out in Jäger (2010) and (2011) and in the present paper, my claim is that Molinists should – and Molina actually did – reject principle (P). However, (P) is not “the β-rule”. It is equivalent neither with the version van Inwagen uses in his original consequence argument nor with the amended version of that rule which Brüntrup and Schneider call (β*) and which, they argue, can be used both in a consequence argument for nomological incompatibilism and in an anti-Molinist argument (along the lines of Perszyk 2003). Their β*-principle runs as follows: (β*)

8

If p, and no one has, or ever had, a choice about making it possible that not-p; and p necessarily entails q; then q, and no one has,

See for example Stump and Fischer (2000), Warfield (2000).

256

Christoph Jäger

or ever had, a choice about making it possible that not-q, symbolically: N*p, □(pq) |– N*q Now, I agree that this rule may be used in a van Inwagen-style consequence argument concerning causal determinism. But (β*) is obviously quite different from (P). Here is a consequence argument that can be constructed with (β*). Let L stand for a conjunction of the laws of nature, H for a proposition describing some complete state of the universe in the remote past (when no human beings existed), and P for a proposition describing some arbitrary later state of the universe. Employing (β*), we can then construct the following consequence argument for the incompatibility of freedom and causal determinism.9 Suppose determinism is true. Then it holds that: D. A consequence argument for nomological incompatibilism (1) (2) (3)

(L&H), and no one has, or ever had, a choice about making it possible that not-(L&H) N*(L&H) Necessarily: (L&H) entails P. □(L&HP) Therefore, no one has, or ever had, a choice about making it possible that P. N*P

(3) follows from (1) and (2) by (β*). Note, first, however that principle (P) employs subject and time indices, while (β*) is a quantificational statement. This may be a minor point, for we may still say, with Brüntrup and Schneider, that bye and large (β*) has “the same structure” as (P). The relevant P-principle that corresponds to (β*) is (with “N” to be read in the sense given in section 1): (P*) Np, □(pq) |– Nq

9

For a detailed systematic reconstruction of consequence arguments of this kind, including a discussion of the most prominent objections to them, see Jäger (forthcoming).

Scientia Media and Freedom to Do Otherwise

257

The crucial problem is still that (P) and (P*) employ an operator that is metaphysically distinct from the one employed in (β*). (P) and (P*) say that causal impact necessity concerning a given state of affairs is closed under entailment; (β*) says that having no choice about a given state of affairs is closed under entailment. It is therefore perfectly coherent for the Molinist to reject (P) (and P*), thereby avoiding the charges of theological determinism as formulated in Argument A and Argument B, and yet to greet (β*) with pleasure, thereby adopting an important tool for a traditional consequence argument for nomological incompatibilism. As argued above, it is doubtful whether the Molinist actually does need a consequence argument in order to accept libertarianism. But even if Brüntrup and Schneider were right that there is such a requirement, rejecting (P) would not cause trouble for the Molinist, because he could still accept (β*) or a similar rule concerning lack of choice. Exactly why, it may be asked, is the concept of having no choice not equivalent with the concept of CI-necessity?10 After all, if I have no choice simpliciter about a given state of affairs, then I have no causal choice about it either. There certainly is entailment in one “direction”. The converse however does not hold. Causal impact necessity with respect to p does not entail that there is a general lack of choice with respect to p. At least, this holds within the Molinist doctrine and given the way Brüntrup and Schneider construe the notion of lack of choice. According to this construal, the notion is to be understood as follows: (No choice) No one has, or ever had, a choice about p if p obtains, and no one has it, or ever had it, within his or her power to perform an action A such that A occurs and therefore not-p obtains. (Cf. Brüntrup and Schneider 2011, 225) Brüntrup and Schneider then argue that the Molinist can coherently, and rightly, claim that humans do have a choice about the counterfactuals of freedom concerning them, and hence that the Molinist can deny that humans have no choice about what God knows, via middle knowledge, about 10

Standardly, equivalence is a relation between statements or propositions. We may say however that concept A is equivalent to concept B iff every statement employing A is equivalent to a corresponding statement employing B. (This definition could still be amended and be made more precise, but for present purposes it will suffice.)

258

Christoph Jäger

them. However, as Brüntrup and Schneider (rightly) emphasize, this does not mean that we can “exert any causal power on the knowledge of God” (2011, 230). (The “therefore” in Brüntrup and Schneider’s construal of the no-choice operator is thus not meant to express a causal relation.) As the authors indicate, this is indeed perfectly in line with what Molina himself says about the issue. “If,” Molina writes, “as is now truly possible, he [Peter] were not going to sin, then that knowledge [that Peter does sin at a certain time, C.J.] would not have existed in God, and so that knowledge, which would not have existed if, as is possible, Peter were not going to sin, does not in any way prevent Peter’s now being able ... not to sin, in just the way he would have been able not to sin had such knowledge not existed”.11 Molina emphasizes that in general, “if created free choice were going to do the opposite, as indeed it can, then God would have known that very thing through this same type of knowledge [i.e., through middle knowledge, C.J.], and not, what He in fact knows”.12 Suppose that – contrary to what these passages suggest – no one has, or ever had, a choice about whether Peter sins at a certain time (for example, about whether he denies Christ in the night before Christ is crucified). Suppose that God is omniscient. Then it would follow, (i) that God knows that Peter sins at the time in question, and (ii) that no one has it, or ever had it, within his power to perform an action A such that A occurs and therefore God does not know that Peter sins at that time. But this is precisely what Molina, along with many other “Scholastic Doctors” he quotes on the issue, denies. (Cf. Molina, Concordia, 4, 52, 19-30) Peter himself, Molina argues, did have it within his power to abstain from denying Christ. Hence according to this view Peter had it within his power to do something such that, had he done it, God would not have known – neither prevolitionally through middle knowledge nor postvolitionally through His “free knowledge” – what He did know, namely that Peter would sin on the night in question. Hence Peter did have a choice about God’s knowledge about him, even though he did not have 11

“At si, ut nunc re ipsa potest, non esset peccaturus, illi non extitisset scientia illa in Deo atque ita nihil omnino impediente scientia quae, si Petrus, ut potest, non esset peccaturus, non extitisset, potest ... non peccare, non secus ac posset, si talis scientia non praeextitisset.” (Molina, Concordia 4, 52, 30, 351/186). 12 “Si namque liberum arbitrium creatum acturum esset oppositum, ut revera potest, idipsum scivisset per eandem scientiam, non autem quod reipsa scit.” (Molina, Concordia 4, 52, 10, 340/168f.)

Scientia Media and Freedom to Do Otherwise

259

any causal access to God and His knowledge about Peter. Brüntrup and Schneider accept this part of the Molinist doctrine. Indeed, their constructive proposal as to how the Molinist ought to solve the problem of theological determinism essentially draws on this very point. What the authors overlook, however, is that precisely these observations show that there is an essential difference between having no choice and causal-impact necessity. While having no choice in general about whether p entails that p is causally inaccessible, the mere causal inaccessibility of p’s obtaining does not entail that there is no choice about whether p. The upshot is that the two operators are different and that because of this difference the Molinist can coherently reject (P) and (P*) while accepting (β*) (and related nochoice principles). “All right”, someone may argue at this point on behalf of Brüntrup and Schneider, “I grant that the two operators are distinct. However, your concession that there is entailment in one direction – that having no choice about whether p entails that p is CI-necessary – suffices for Brüntrup and Schneider’s argument to go through. For it still follows that if having no choice is closed under entailment, then CI-necessity is so closed as well. Hence if, as you claim, CI-necessity is not closed under entailment, having no choice is not closed under entailment either. Brüntrup and Schneider are still right therefore that your proposal for how the Molinist should deal with Arguments A and B causes trouble for him. They are right that the Molinist cannot consistently reject (P) and (P*) while accepting (β*).” Maybe it is this kind of reasoning that underlies Brüntrup and Schneider’s objection. I am not claiming that it does. Yet, similar ideas have been suggested in analogous contexts, so the point is in any case worth examining.13 For example, consider a famous debate in epistemology about the question whether knowledge is closed under known entailment. Antony Brueckner once argued that knowledge is closed under known entailment only if each necessary condition for knowing is so closed (Brueckner 1985, 1994). He then argues that, inter alia, belief is not closed under known entailment and that therefore knowledge is not so closed either. This argument is analogous to the one just sketched concerning the notions of CInecessity and having no choice.

13

For a discussion of this topic see also Jäger (2011).

260

Christoph Jäger

However, both arguments are flawed. If knowledge entails belief but not vice versa, the set of a subject’s beliefs that constitute knowledge is a proper subset of the subject’s beliefs in general. Hence not all elements of the larger set (the set of the subject’s beliefs, including those that do not constitute knowledge) have all the properties of the elements of the subset containing the beliefs that qualify as knowledge. And what turns a belief into knowledge may remove precisely those items from the larger set that are not closed under (known) entailment, in which case knowledge would be closed under (known) entailment, even though belief is not. An exactly analogous argument applies to the relation between having no choice and CI-necessity. In general, it is not true that, if a statement ascribing some modality M to something entails that the thing in question also exemplifies another modality M*, then if M is closed under entailment, M* is closed under entailment. Consequently, just as it is coherent to maintain that knowledge that p entails believing that p, but that knowledge that p is closed under entailment even though belief is not, it is perfectly coherent to maintain that having no choice about p is closed under entailment, and yet that CI -necessity is not. 3. Conclusion Brüntrup and Schneider’s paper touches upon many interesting aspects of the free will problem, in both its theological and nomological guises. Their discussion also contains a number of important observations about Molinism in particular. However, the authors’s objections to my reconstruction and defence of an important part of the Molinist account fail. Neither is it open to the Molinist to accept principle (P), nor is there any need for him to do so in order coherently to maintain a libertarian and incompatibilist position. Moreover, the counterexamples I consider are on target and show that (P) is indeed false. Brüntrup and Schneider should be credited for drawing attention to the question how various modal closure or transfer principles that play a crucial role in the present debate relate to each other. Yet the β-principles they consider and my principle (P) employ different operators and are thus not equivalent. On the other hand, Brüntrup and Schneider’s constructive claim that we enjoy counterfactual power over the truth values of counterfactuals of freedom concerning our actions and decisions is, in my view, not only interesting and important but also true. Moreover, this claim is in line with classi-

Scientia Media and Freedom to Do Otherwise

261

cal Molinism. Yet, as I have argued in this paper, this position cannot be an alternative to the moves I suggest but can only complement them. References Brueckner, A. 1985 “Skepticism and Epistemic Closure”, Philosophical Topics 13, 89–117. -

1994 “The Structure of the Skeptical Argument”, Philosophy and Phenomenological Research 45, 827–835.

Brüntrup, G., 2000 “Der metaphysische Begriff der Willensfreiheit und das Transferprinzip des Keine-Wahl-Habens”, in Wahrheit–Sein–Struktur: Auseinandersetzungen mit Metaphysik, ed. C. Peres and D. Greimann, New York – Zürich: Olms, 102–119. Brüntrup, G., and Schneider, R. 2011 “How Molinists Can Have Their Cake and Eat it too”, this volume. Freddoso, A. (ed.) 1988 Luis de Molina, On Divine Foreknowledge (Part IV of the Concordia), Ithaca: Cornell University Press, 83–286. -

1988 “Introduction”, in id. (ed.), Luis de Molina, On Divine Foreknowledge (Part IV of the Concordia), Ithaca: Cornell University Press, 1–81.

Jäger, C. 2010 “Göttlicher Plan und menschliche Freiheit”, Philosophisches Jahrbuch 117, 299–318. -

2011 “Molina on Foreknowledge and Transfer of Necessities”, in C. Tapp and E. Runggaldier (eds.), God, Eternity, and Time, Farnham – Burlington: Ashgate, 81–96.

-

forthcoming “Das Konsequenzargument”, forthcoming in R. Puster (ed.), Klassische Argumentationen der Philosophie, Paderborn: mentis.

Molina, L. de, S.J. Liberi Arbitrii cum Gratiae Donis, Divina Praescientia, Providentia, Praedestinatione et Reprobatione Concordia, critical edition of the second edition from 1595, J. Rabeneck, S.J., (ed.), Oña – Madrid: Collegium Maximum S. J. 1953 Perszyk, K. 2003 “Molinism and the Consequence Argument – a Challenge”, Faith and Philosophy 20, 131–151. Ravizza, M., 1994 “Semi-Compatibilism and the Transfer of Nonresponsibility”, Philosophical Studies 75, 61–93. Runggaldier, E., 1996 Was sind Handlungen?, Stuttgart et al.: Kohlhammer. Stump, E., und Fischer, J.M. 2000 “Transfer Principles and Moral Responsibility”, Philosophical Perspectives 14, Action and Freedom, 47–55. Van Inwagen, P. 1983 An Essay on Free Will, Oxford – New York: Clarendon Press.

262

Christoph Jäger

Warfield, T. 2000 “Causal Determinism and Human Freedom are Incompatible: A New Argument for Incompatibilism”, in: Action and Freedom (Philosophical Perspectives 14), 47–55. William of Sherwood 1995 Introductiones in logicam, ed. H. Brands and C. Kann, Hamburg: Meiner. Zagzebski, L.T. 1991 The Dilemma of Freedom and Foreknowledge, Oxford: Oxford University Press. -

2002 “Recent Work on Divine Foreknowledge and Free Will”, in R. Kane (ed.), The Oxford Handbook on Free Will, Oxford: Oxford University Press, 45–64

Dualism and Resurrection UWE MEIXNER, Augsburg

What dualism is – I always use the word here as an abbreviation of “psychophysical dualism” – is by no means as clear as it may seem at first sight. The following statement formulates the minimal content of every dualism, that is: the thesis which is logically implied by every dualistic assertion or theory – the thesis, indeed, which has to be logically implied by a theory or assertion in order to be legitimately called dualistic: The Minimal Dualistic Thesis Some mental entities are at least partly nonphysical, in other words: Some mental entities are not completely (or purely) physical. We may define: Definition 1: An assertion or theory is dualistic if, and only if, it logically implies the thesis that some mental entities are at least partly nonphysical. Definition 2: A philosopher is a dualist if, and only if, he maintains an assertion, or a whole theory, that is dualistic in the sense just defined, that is: an assertion, or theory, that logically implies the thesis that some mental entities are at least partly nonphysical. These definitions are helpful for recognizing certain specifications, abounding in the literature, of what allegedly is asserted by the doctrine of dualism, or by the dualists, as being by no means obvious and by no means necessary. Here follows a small selection of more or less implausible theses that are attributed to the doctrine of dualism, or to the dualists, in philosophical and para-philosophical research and teaching: 1) Body and soul, and what belongs to the body and what belongs to the soul, have nothing to do with each other. 2) The brain and nervous system play no causal role at all for psychical processes. 3) Besides physical matter there exists a queer nonphysical matter: mind stuff.

264

Uwe Meixner

4) Body and soul confront each other inimically. 5) There are exactly two substances, the extended thing and the thinking thing. Not a single one of these five theses I consider to be true. Nevertheless, in the sense of the definition given above, I am a dualist. One may, therefore, deny each of these five theses, consider true the negation of each of them, and yet: this does not make one a non-dualist, let alone an anti-dualist, not by a long shot. For the Minimal Thesis of Dualism and the two definitions subsequent to it have their counterparts: The Minimal Anti-Dualistic Thesis No mental entities are at least partly nonphysical, in other words: All mental entities are completely (or purely) physical. Definition 3: An assertion or theory is anti-dualistic if, and only if, it logically implies the thesis that all mental entities are completely physical. Definition 4: A philosopher is an anti-dualist if, and only if, he maintains an assertion, or theory, that is anti-dualistic in the sense just defined, that is: an assertion, or theory, that logically implies the thesis that all mental entities are completely physical. Few philosophers, I presume, would underwrite the Minimal Anti-Dualistic Thesis; few philosophers are, therefore, anti-dualists. But it is also true that few Western philosophers still consider themselves dualists. In reaction to this, I point out that if one is not an anti-dualist, then it is, for logical reasons, not entirely easy to be also not a dualist. Hylomorphists, for example, like to see themselves as non-dualists that, nevertheless, are not antidualists. In this connection, they like to emphasize the unity of body and soul, which the dualists allegedly deny. But when hylomorphists speak of that unity, they surely do not mean to say that body and soul are numerically identical; they intend that body and soul, though different, are inseparable. But what is the foundation of the conceded difference of body and soul? – Although in the entire tradition of hylomorphism the soul does not figure as a substance, in any case, not as a substance in the full sense, it figures there nonetheless as a quasi-substantial concrete individual entity (which, by the way, according to Thomas Aquinas, can exist without the body if it absolutely has to), whose difference from the body primarily consists in its being at least partly nonphysical. Moreover, the soul is also

Dualism and Resurrection

265

in the hylomorphic tradition a mental entity, since also for hylomorphism the soul is the main carrier of everything mental belonging to the person. Considering that hylomorphist have always accepted the existence of human souls, it is a logical consequence of what has been said in the preceding paragraph that in the entire tradition of hylomorphism the assertion has been maintained that some mental entities (namely, human souls) are at least partly nonphysical. All hylomorphists are, therefore, what they by no means want to be: dualists. And therefore: perhaps the vehement defensive reaction from the side of the hylomorphists that I, as a confessing dualist, have been confronted with was less due to the confessed thing itself than to the fact that one does not like the words “dualism”, “dualistic”, “dualist”, even despises them – presumably because whole complexes of negative associations are connected with them, associations which are by no means new but many centuries old. Unfortunately, I do not have any better words at my disposal for the thing in question; I will, therefore, continue using the old ones in speaking about it. The dualism that I maintain and that I have tried to substantiate and defend in many of my publications rests on a phenomenological foundation: on self-experience – self-experience, however, which can be had by everyone any time – and on the conceptual means that ordinary language about the psychical affords to all of us. It is perhaps not entirely unnecessary to point out that even such a simple thesis as the Minimal Dualistic Thesis logically contains a proper part-assertion which, by itself, is not yet of a dualistic nature: Some entities are mental. This is an assertion that one might dub psychological realism. Psychological realism is a logical part of what is stated by any form of dualism, and there is, therefore, the possibility to deny any dualism simply by denying this realism contained in it – without needing to address the question of the physicalness or nonphysicalness of the mental. This road of dualism-denial is travelled by the materialist eliminativists, following Rorty and the Churchlands: merely consider that “Mental entities do not exist” is just a different way of saying what is also said by the sentence „No entities are mental“. But this same road of dualism-denial is travelled also by the many philosophers who have been influenced, with permanent effect, by Wittgenstein’s Philosophical Investigations, to which group of philosophers in

266

Uwe Meixner

recent times also many Christian philosophers are proud to belong. In opposition to this, it needs to be pointed out that if reflexive experience (or self-experience) is respected as a source of knowledge, then there is no way to avoid acknowledging that there are mental events, that, in other words, some events are mental, that therefore some entities are mental (for events are ipso facto entities, considering that the category of event is an ontological category) – which makes psychological realism a experientially verified position. Ordinary language, too, in no way militates against psychological realism, but at all ends allows psychological realism to emerge as implicitly presupposed by it, as part of the ontology required by it: namely, via the rich vocabulary that ordinary language affords for the description of mental events (the conscious ones among which are called experiences) – a vocabulary that evolved in ordinary language entirely without the efforts of misguided and benighted Cartesian-minded philosophers. We speak, without any difficulty, of pains and of other sensations of the most various kinds, intensities, and qualities; of episodes of thinking, of sadness, of resolution; of experiences of conviction and perception, and so on. Can it really be that all of this does not really exist, that our innerpsychological language is a purely mythological or fictional language, which, at best, serves as an indicator of a hidden neuronal reality, very inadequately grasped in it, as for example Daniel Dennett believes? This does not seem to be the case – in view of the testimony of the reflexive experience, given to each and every one of us, of our own psychical states, which in an immediate, non-inferential manner gives to each and every one of us something to describe, with which we, employing our means of description and drawing on our abilities of description, wrestle with more or less success in order to express it descriptively. One can escape psychological realism only if one does not respect reflexive experience, denies its cognitive value. Many philosophers have no scruples to do just this; but I do not see any good reason to do as they do. Psychological realism is half the battle: some entities are mental. Now, if some among those entities are at least partly nonphysical, then one has won the truth of the Minimal Dualistic Thesis and therewith, certainly, the truth of some form of dualism or other. But the conclusion that some mental entities are at least partly nonphysical can hardly be avoided. Consider a mental event: my visual experience that begins with my removing my glasses now, and that ends with my putting them on again now. This experience – let’s call it “L” – had as long as it lasted, as I am told by reflexive

Dualism and Resurrection

267

short-term memory, a certain very prominent property, which I designate as “visual indistinctness”. That L had this property had, in turn, certain purely physical causes, which are with great exactness known to the experts – causes that I put into effect when I removed my glasses, and that I put out of effect when I put my glasses on again. This purely physical causal history, however, does nothing to change the fact that nothing completely physical had, has, or ever will have the property of visual indistinctness. Visual indistinctness is not a property of anything purely physical: neither purely physical micro-objects nor purely physical macro-objects nor purely physical events are ever visually indistinct. One can, in the right circumstances, say that one sees such entities indistinctly; but what is indistinct then are not they themselves, but the visual experiences that one has while one is seeing them. L is, therefore, an experience – hence a mental event – which is at least partly nonphysical, and I can generate other experiences just like L at any time by removing, with my eyes open, my glasses and putting them on again. This shows the Minimal Dualistic Thesis to be true, even if one takes the plural in it seriously. Indeed, regarding its nonphysicalness, L is by no means a singularity among my experiences: a property can be found for each of my experiences which it had, has, or will have, but which no purely physical entity had, has, or ever will have, namely, its intrinsic subject-relatedness: the for-me-character intrinsically belonging to each of my experiences, a character which perhaps emerges most conspicuously in the case of my pain-experiences (it is evident to me in the most immediate manner that they are intrinsically – and sometimes very urgently – addressed to no one else but me). All of these considerations consist in reflexive-phenomenological demonstrations, the analogs of which can be carried out by everyone with respect to himself or herself. Such demonstrations also show what has now already been touched upon: that by no means only mental events are entities that are at least partly nonphysical. My experiences come and go; in all of them, in every single case, there is wholly present as numerically the same entity, and therefore not as an event, not as an occurrence, their subject: I – not I, the entire human being, but I, the subject of consciousness. Just like the experiences whose subject I am, I cannot be found among the purely physical entities – and certainly not as a little man, a homunculus at some place in my brain. Again one has the option simply to throw away the testimony of reflexive experience, to declare that it is illusionary. Some

268

Uwe Meixner

people believe that science leave us no choice than to reject that testimony. In my publications I have repeatedly argued that it not true that science leave us no choice. Our experiences, and we as subjects of consciousness, have a neuronal foundation which is without any doubt a nomological conditio sine qua non of their existence and of ours, qua our being subjects of consciousness. To what extent this neuronal foundation is also a nomologically sufficient condition of the so-being of these experiences and of ourselves as subjects of consciousness – this we are presently, after millennia, finally about to discover en detail. A dualist need not close his eyes in any manner to the progress of the neurosciences, he need not deny any part of it; but he has the rational right to interpret this progress in his own way, and that means: to interpret it in an anti-reductionist way; not to deny phenomena, but to integrate them. Materialism has no monopoly on science; science just is not in itself physicalist-reductionist. The convulsive efforts to force everything into the physicalistic corset, and if for a given item one does not succeed in doing so, to pronounce it nonexistent – hopefully these kinds of philosophical behavior, prevalent today, will at some time in the not too distant future seem as queer as seem today the desperate attempts of many thinkers of the nineteenth century to understand electromagnetic phenomena on the basis of mechanistic conceptions. Once upon a time one sought to reduce all physical phenomena to mechanical phenomena. The ultimate motivation for making that attempt was this: one believed a priori that, in the last analysis, there are no nonmechanical physical phenomena; that it can only seem temporarily as if such phenomena existed, namely, as long as their mechanistic reduction had not been effected. This a priori opinion was an error, and the reductive intention it motivated was misguided. The view that, in the last analysis, there are no nonphysical mental phenomena; that it can only seem temporarily as if such phenomena existed, namely, as long as their physicalistic reduction has not been effected, is likewise an a priori opinion; it is likewise an error, and the reductive intention it motivates is likewise misguided. Behind this a priori opinion – and behind the a priori ignoring, connected to it, of entire regions of experience – there is a heartfelt ontological aversion against everything supernatural. One is deeply convinced that there is nothing “strange (uncanny, weird, …)” going on in the world, and one believes that, if one conceded that some mental entities are at least

Dualism and Resurrection

269

partly nonphysical, then one would concede that there is after all something “strange (uncanny, weird, …)” going on in the world. Being thus disposed, one overlooks the logical fact that with the concession of nonphysical mental entities one would only leave the physicalistic, materialistic position, but by no means already the naturalistic one. One can be a dualist and a naturalist. I myself, however, while being a dualist, am not a naturalist. The word “mystery” is not a sort of philosophical cussword for me. But I do not connect it with spiritistic sessions, ghosts, and the like. Although I am not a naturalist, I have tried to show, in several of my publications, that dualists quite naturally have a purely naturalistic perspective, if they do want to travel that road. Among the more unlikely, but nevertheless not impossible ways of the evolution of matter there is the emergence of conscious processes and of subjects of consciousness – hence of at least partly nonphysical mental entities, which, however, are nomologically dependent on the purely physical, especially on brains (what this dependence is exactly like – this, as I said, we are presently about to discover). Like everything that comes to be in living organisms and does not disappear again soon but is, on the contrary, developed further, conscious processes and subjects of consciousness must have a positive effect for the survival of their carrierorganisms and, consequently, also of their carrier-organisms’ species. The following fact is conspicuous: what subjects of consciousness spontaneously experience to be positive is indeed – not always, but in the main – good for the preservation of the organism and its species; and what subjects of consciousness spontaneously experience to be negative is indeed – not always, but in the main – bad for the preservation of the organism and its species. Is this mirroring in consciousness of vital biological values – in the at least partly nonphysical experiences of more or less great joy, and of more or less great pain – only an idly co-occurring, epiphenomenal accompaniment of the struggle of the organism for survival, or does it have a genuine function for this struggle? The latter alternative has the – by far – greater probability on its side. What, then, is that function? From the experiences of joy and pain the subject of consciousness derives the motivational points of view which, in cooperation with perception, allow it to make a contribution to safeguarding the survival of the organism and of its species: not the main contribution, but often a crucial one. Consciousness and subject of consciousness are, therefore, nonphysical organs of the organism: functionaries for special tasks. Within the framework of the laws

270

Uwe Meixner

of nature, they exist no longer than the organism itself exists, that is: lives, and even no longer than the organism’s nervous system is sufficiently functional. But in order to be able to fulfill their life-guiding, life-saving, lifestrengthening function, consciousness and subject of consciousness must have appropriate causal powers – there is no way around this fact. Physicalistic thinkers are prone to attack at this point, asserting that such causal powers cannot exist. They overlook that physics itself is rather more open for those causal powers than physicalism, which, in contrast to physics, is a special form of metaphysics. This special form of metaphysics is, contrary to its self-image, by no means the only rational extension of the natural sciences. In any case: that in an energetically closed system the total sum of energy neither increases nor decreases – in other words: physics’ law of the conservation of energy – does not preclude the existence of the causal powers in question; nor does the law of the conservation of momentum. And the often invoked principles of causal closure have nothing to do with physics, but everything with physicalism; their invocation by physicalists is, therefore, a glaring case of begging the question. In other places I have explained these matters at great length; I will not talk about them further here. Now, what I have just outlined concerning the biological role of consciousness and subjects of consciousness can be – in outlines: omitting the details – the whole story of consciousness and subjects of consciousness, and for a naturalistic dualist it is the whole story. But even it is in fact not the whole story, as I believe, it must nevertheless be maintained that though this story then needs to be complemented – and given the complementation, part of what has already been said will appear in a new light – yet, given the complementation, nothing of what has already been said is completely rescinded. Even if what has already been said is not the whole story, it remains true that there cannot be a naturally continued existence of the subject of consciousness after the demise of its organism. As far as the field of the natural – the field that is staked out by the laws of nature – reaches, the captain of the ship, to use an old simile, must go down with his ship. If he nevertheless survives its destruction, if he even becomes the captain of a new ship, then this is a miracle in the true sense: a breaking of the laws of nature (which laws –to avoid falling into inconsistency here – must of course be conceived of in such a way that they, though laws of nature, are nevertheless not unbreakable). This much was intuitively clear to the witnesses of the resurrection of Jesus from the dead, also without

Dualism and Resurrection

271

their having a fairly distinct idea of the laws of nature or being even acquainted with a fairly clear concept of law of nature. Hence they intuitively took the resurrection of Jesus as the truly overwhelming ultimate proof of His divine origin – for only who is Lord over the laws of nature can break them – and derived from their apperception a good part of the immense spiritual force that is necessary for founding a new world-religion. Now, all Christian believers are hoping for their own resurrection. That is, they hope for a miracle, a miracle greater than all those so-called miracles of which it is commonly said – in the media, for example – that they are being hoped for. How is that miracle – aside from the fact that, qua miracle, it is a breach of the laws of nature – to be thought of, if in modest presumption (if I may say so) one does undertake to think about these last things? They are divine things, supernatural things, but they do concern us if we are believers. Like all people that speak positively about these things, I now speak – this is in the nature of the case – less of how these things are, and more of how it seems to me that they would have to be like if they were truly the fulfillment of existential human hopes. I take myself as an arbitrary example. For my resurrection to be the fulfillment of my hope, I myself – numerically the same person that once died – must be resurrected, and not some copy, not some follow-up model of me. For my resurrection to be the fulfillment of my hope, I must moreover be resurrected to a life in the full sense, not to an existence which is like that of the shadow-souls in Hades. Prima facie this seems to suggest that the resurrection of the dead would have to be in such a way that what we find written in Rainer Maria Rilke’s poem “Resurrection” 1 turned out to be – as far as the essentials are concerned – a true description: The count hears the sounds, he sees a luminous crack; he wakens his thirteen sons in the family tomb. He greets his two wives respectfully from afar – ;

1

What follows below is my prose translation into English of Rilke’s poem “Auferstehung [Resurrection]” (Rainer Maria Rilke, Die Gedichte, Frankfurt a. M.: Insel Verlag 41990, p. 470).

272

Uwe Meixner

and all, full of trust, rise to eternity and wait only yet for Erich and Ulrike Dorothea, who, at the age of seven and thirteen, (sixteen hundred and ten) died in Flanders, and who today, in front of the others, walk unwaveringly.

This is one way to imagine bodily resurrection, and the poem irresistibly suggests that the bodies with which the resurrected persons died are numerically the same bodies with which they are also resurrected. But this would mean that the life of the resurrected, although it is life in eternity, is more or less like the life that they led before their deaths (with certain cutbacks perhaps, in the poem tenderly alluded to by the “he greets his two wives / respectfully from afar”). In opposition to this I would like to maintain: First, if my resurrection is to be the fulfillment of my hope, then it must be a resurrection to an eternal and, by the grace of God, glorified life, that is, to the completion of the way that is called theosis in the GreekOrthodox Christian tradition: the divinization of Man by divine grace. Second, my present body is not only presently – as it is now – incapable of an eternal life glorified by the grace of God, it is also essentially unfit for such a life: it is metaphysically impossible that I lead such a life and still have, or have again, this body, my present body. Therefore, if I am bodily resurrected, then not with this body, but with a new body essentially different from it; which new body, for example, appears (as is told of the body of the resurrected Christ) without nearing and disappears without moving away; but which new body has, as long as it is present, all empirical marks of a normal material object, that is: carries out the usual physical actions on other material objects, and suffers such actions from them; but suffers them, none the less, only in a very peculiar manner, namely, in a manner excluding all injury. What do you think would have happened if somebody had tried to impale the resurrected Christ on a lance, or throw Him to the ground? What do you think would have happened if the resurrected Christ had emptied, like Socrates, the cup of hemlock?

Dualism and Resurrection

273

If one wishes to deny that my present body is essentially unfit for eternal life, given the way that life ought to be, then one must proceed on the assumption that the validity of the laws of nature for my present body is metaphysically contingent. Then one must proceed on the assumption that it is possible that this body, numerically the same body, acts in ways that are radically contrary to the laws of nature. One must consider this to be a metaphysical possibility, but, of course, not also a nomological one. But even as a mere metaphysical possibility, the possibility in question does not seem plausible to me. But if, when I am resurrected, my present body is a thing of the past, a no longer useable garment, and I am clothed in a new body, a new garment that fits the new earth and the new heaven of which the Book of Revelation speaks, then my present body obviously cannot be the carrier of my personal identity, cannot be what guarantees that I myself am resurrected, I the same who once died – and not a copy or follow-up model of me. Then my present body cannot be what guarantees my personal identity in the Resurrection, even if Almighty God did preserve, as some believe, that body or a piece of it somewhere without interruption – through all cosmic catastrophes, including the earth being swallowed by the sun – so that it did indeed exist continuously (though dead) until the Last Day. Even if the old body is still there – no matter whether after a short, after a long, or after no time – it is for my glorified existence not the ontologically appropriate body, and therefore, a fortiori, it does not guarantee that it is I who is resurrected to the life everlasting. What is it, then, that guarantees my personal identity in the Resurrection? Clearly not I, the entire human being, for it is only in the Resurrection that I again become an entire human being: a transfigured and glorified entire human being; rather, it is, with the miraculous help of God, I, the former subject of consciousness and agent for my old body, in other words, I, the soul of that body, with my memories and dispositions, who guarantees that identity. Though the new body and the old body are essentially different: though the new body exists as my body under the conditions of resurrection to an eternal life in glory, while the old body is essentially incapable of this, the two bodies are certainly not unrelated to each other. For one thing, it is the same human soul that is united with them, one after the other, and this brings along with it many things. The soul has a history, a human history with the old body, a history of experience, expression, and action, depo-

274

Uwe Meixner

sited in its memories and acquired dispositions, and if my resurrection is to be the fulfillment of my hope, then this history together with its center, the soul, finds a true new home in the new body, which therefore must also be a human body, though a glorified and transfigured one. And no negligible continuity between the old and the new must already be preserved for the mere reason of making it apparent to other people that they are dealing with the same person they once knew. This is also the reason why the new body of the resurrected Christ carries the stigmata of His old body and quite generally looks like the old body. Only in this way could Doubting Thomas become Thomas the Apostle. The new body stands in continuity to the old one, and yet it is a radically new and other body. The relations of power have been reversed. Whereas the soul served the old body as a steward of its vital interests in the contingencies of this-worldly material life – this was a position that was nomologically open and which biological evolution filled, with the strange dependability characteristic of it, in the manner in which that position could be filled under the conditions of nature: by the coming into being and further development of central nervous systems; and whereas the soul had, in its organism-serving role, a little free time at best, the new body is, on the contrary, the vehicle of the soul’s glorious freedom. In summing up, the following can be said about the nature of the human person from a dualistic perspective, with special consideration accorded to the Christian hope of resurrection. The human person is, properly speaking, the human soul, which is called “human” for the reason that it is essential for it to have a human body. However, it is no more essential for it to have its present body, its body in via, as it is essential for it to have its coming body, its body in patria. Now, is it the same human being – the human being that lives after the resurrection, and the human being that lived before the resurrection? One thing is clear: before and after the resurrection it is the same human person who lived and lives, albeit with different human bodies. If a human being is a unified whole out of body and soul, then the conclusion can only be that after the resurrection it is a new human being that lives, numerically different from the old human being – the latter human being having definitively died. I will then say: I was that human being who died, but I am now this human being who lives. Speaking in that way, I will be making use, as I already do in this life, of the possibility to employ the word “I” not only for designating the human person – that is: the human subject of consciousness and human agent, the human soul – but

Dualism and Resurrection

275

also to employ it analogically for designating the intimately united pair of human soul and human body, that is: for designating the entire human being. Given this possible (and very usual) analogical employment of “I”, I certainly will not do anything more unusual from the logical point of view in saying that I was that human being but am now this human being than I would do in saying that the Ship of Theseus was 200 years ago that ship (pointing to a ship in the dry dock) but is now this ship (pointing to a ship afloat). References More on my views on the nature of the human person and the human mind – and what I did not offer here: the detailed discussion of the opinions of others – can be found in the following English-language publications of mine: Meixner, U. 2004 The Two Sides of Being. A Reassessment of Psycho-Physical Dualism, Paderborn: mentis. -

2006a “Consciousness and Freedom”, in A. Corradini, S. Galvan, and E.J. Lowe (eds.), Analytic Philosophy Without Naturalism, London: Routledge, 183-196.

-

2006b “Classical Intentionality”, Erkenntnis 65, 25-45.

-

2006c “The Indispensability of the Soul”, in B. Niederbacher and E. Runggaldier (eds.), Die menschliche Seele. Brauchen wir den Dualismus?, Heusenstamm: ontos, 19-40.

-

2007 “Three Tasks for (Hard Interactionist) Dualists”, in F. Castellani and J. Quitterer (eds.), Agency and Causation in the Human Sciences, Paderborn: mentis, 49-74.

-

2010a “Materialism Does Not Save the Phenomena – and the Alternative Which Does”, in R.C. Koons and G. Bealer (eds.), The Waning of Materialism, Oxford – New York: Oxford University Press, 417-437.

-

2010b “The Emergence of Rational Souls”, in A. Corradini and T. O’Connor (eds.), Emergence in Science and Philosophy, New York – London: Routledge (Taylor & Francis), 163-179.

Living like God in Switzerland? Soul, eternal present and beatitude in Jean-Jacques Rousseau’s fifth promenade NIKO STROBACH, Münster

Dicit enim Boethius quod nunc fluens facit tempus, nunc stans facit aeternitatem. Thomas Aquinas, ST I q10 a2 obj1 La mer Qu’on voit danser le long des golfes clairs A des reflets d’argent La mer Des reflets changeants Sous la pluie [...] Charles Trenet (1946)

1. Introduction The aim of the present paper is to provide a philosophical reading of JeanJacques Rousseau’s fifth promenade from his rêveries du promeneur solitaire. Although the text is well-known among scholars of French literature, it might not be quite as well-known among analytic philosophers or philosophers of religion.1 Why should they, too, care about this text? Historically, the fifth promenade is a remarkable text, since it takes up traditional ideas on divine life and its occupation, eternity, consciousness of time, the soul, and supreme happiness which have been around since Aristotle’s Metaphysics XII and were developed in the Middle Ages. No doubt, these are topics which have played an important role in Edmund 1

The “or” is inclusive, of course. I confess that I had never come across the text before my attention was drawn to it by Patricia Oster-Stierle and Karheinz Stierle in the course of a joint seminar „Les lumières – textes philosophiques et littéraires du XVIIIeme siècle“ at Saarland University during the winter semester 2010/11. So my thanks go to both of them and to Svantje de Silva for developing the idea and for bringing the group together.

278

Niko Strobach

Runggaldier’s research and continue to do so. Rousseau makes heavy use of all those elements. In this sense, the fifth promenade, is a surprisingly conservative text. Neither is it far-fetched to say that it is a religious text. Systematically, the text is interesting because of certain tensions it contains. They might be the result of blending Aristotelian and medieval ideas with Hellenistic ideas. A critical mind will wonder: Don’t those elements contradict each other at some deep point? Does Rousseau want to have the cake and eat it? It matters systematically if he can. Personally, the fifth promenade has a certain potential of being a moving text. It is, at the same time, a theoretical and an autobiographical text. Its theoretical content has to be extracted from the bits of autobiography it is naturally blended with. There are more such texts in the history of philosophy: Augustine’s confessions (which provide the model for Rousseau’s confessions), Marcus Aurelius‘ Meditations, Petrarch’s letter on the ascent of Mont Ventoux – to mention just a few. After describing the context I will go into some detail about the most clearly theoretical and philosophical passages of the fifth promenade and present them in such a way that its tensions become visible before indicating what we might learn from them. 2. The structure of Rousseau’s fifth promenade The fifth promenade is a text of about ten printed pages in a standard edition of Rousseau‘s complete works.2 Rousseau (1712 – 1778) wrote his rêveries du promeneur solitaire during the last years of his life, between 1776 and 1778. Not primarily intended for publication, they were published posthumously in 1782. The book contains ten promenades the last of which is unfinished. The fifth promenade looks back to a period of a little less than two months which Rousseau spent on St. Peter’s island, a small island in the middle of lake Bienne in Switzerland in 1765. He fled there after his house in nearby Môtiers, situated in the Prussian territory of Neuchâtel, had been attacked on Sept 6 (he describes the event as la lapidation de Môtiers, cf. 1041). He had fled from Geneva after a passage on 2

In what follows, page numbers without further reference refer to this text. I have retained the original spelling, which is sometimes inconsistent and often differs from contemporary usage, in accordance with the edition by Gagnebin and Raymond.

Living like God in Switzerland?

279

religion at the end of his Emile had caused trouble. It seems that the months he spent on St. Peter’s island were September and October 1765, before he went on to England. As Google informs us, the title of the book, rêveries du promeneur solitaire, is difficult to translate into English. Sometimes it is rendered as Reveries of the (or: of a) Solitary Walker (Rousseau 1979), sometimes as Musings of a Solitary Stroller (Hadot 1995, 262), or even Dreams of a Solitary Stroller (Adeleye et al. 1999, 347). The translation of the title determines the translation of promenade: either „walk“ or „stroll“. But how, more importantly, should one translate „rêverie“? The English word „reverie“ is no translation at all, although, being used for describing of a state one is lost in, it is a relatively precise loanword. „Dream“ is clearly wrong, since „rêverie“ is related to „rêve“ as „Träumerei“ is to „Traum“ („Traum“ would be an absurd heading of Robert Schumann’s Träumerei). All this matters, since the fifth promenade is particularly self-reflexive: It is not just a rêverie, but an attempt at describing the phenomenology, at defining and theorizing about the very speical mental state Rousseau calls rêverie. „Musing“, especially in connection with „stroll“, is certainly understated (even if the original French title should contain some deliberate understatement). Nothing less than supreme happiness is achieved in the state of rêverie. A more traditional term, not unfitting, is „beatitude“. Rêverie is contemplation of a special kind due to its special object. But since „contemplation“ would not work as a translation, I shall leave „rêverie“ untranslated, reminding the reader that this is just what is to be defined in the text of the fifth promenade. What is the structure of the text? Rousseau describes (1) St. Peter’s island (1040f. from the beginning of the text up to „...durant les vendanges“), (2) why he got there (1041: „C’est dans cette isle ... la mienne aussi“.”) (3) and his life on the island (1041: „On ne m’a laisser...“ up to 1045: „... élans du désir“). The description of life on St. Peter’s island is followed by (4) a purely theoretical passage (1046: „J’ai remarqué dans les vicissitudes...“ to 1048 „...pu rêver agréablement“), which smoothly leads back to

280

Niko Strobach

(5) a reminiscence of the state of rêverie on St. Peter’s island which is associated with heaven (1048: „Mais il faut avouer ... peu de tems“), concluding with (6) a description of the state of meta-rêverie, which takes place in 1777 and has as its object the states of rêverie in 1765 (1049: „Les hommes se garderont ... j’y étois reellement“) before suddenly awaking into (7) a lament for, as a bodily being, not being able to sustain the state of rêverie for long (1049: „Le malheur ... offusqué“). Believe it or not, the fifth promenade is a reasoned text: There is a claim, and reasons are provided which are to persuade the reader of its truth. The claim is that the two months on St. Peter’s island were the happiest time of Rousseau’s life („le tems le plus heureux de ma vie“, 1042) – not a very general claim, but one which occasions the rest of the text as an explanation why that was so. The core of the answer is: because the circumstances during those two months quite frequently allowed for the state of rêverie. One might ask: Why would that make them happy months? Because, and this is the second, much more general claim, the state of rêverie just is the happiest state a human being (not just Rousseau) can ever achieve. Formally, the criterion for seeing that it is really supreme happiness that is described here, is a traditional one since the days of Plato’s Philebus (20d-e): Supreme happiness is a state which leaves no desires open. The state of supreme happiness is the highest good just because someone who has attained it will not ever wish for anything else. The first statement of this principle is colloquial, when Rousseau remarks on St. Peter’s island: [J]‘y aurois passé deux ans, deux siècles, et toute l’éternité sans m’y ennuyer un moment [...] (1041)

Similarly, he states: [Le temps sur l’isle] m’eut suffi durant toute mon existence sans laisser naitre un seul instant dans mon ame le desir d’un autre état. (1042) [O]n s’alloit coucher content de sa journée et n’en désirant qu’une semblable pour le lendemain. (1045)

There is a lot of things Rousseau likes about the island. What is most important about it, though, is that life there facilitates attaining the very special state of rêverie, which is to an even greater degree a state of fulfilment than any other activity on St. Peter’s island.

Living like God in Switzerland?

281

3. The description of the state of rêverie St. Peter’s island is a place Rousseau can retreat to and leave his troubles behind. The first half of the text as a whole conveys the impression of a deeply exhausted man (somehow great writers know how to do this). St. Peter’s island is a place where Rousseau can isolate himself (cf. 1040: „le bonheur d’un homme qui aime à se circonscrire“, 1048: „une Isle ... naturellement circonscritte et séparée du reste du monde“). It is a microcosm which allows for a playful immersion in botany, admiring the order and beauty of the world of plants and rehearsing it by classification during walks which are planned systematically, but not too seriously. How nice to be dreaming of writing books about botany without actually having to write them! The few people around are friendly, but do not demand any deeper interest (1048). Thérèse Levasseur, the mother of his children and probably - his wife from 1768 onward, is there, too (1043), but does not play any role. Some opportunity arises for dressing up in order to play creation by, again not too seriously, placing some imported male and female rabbits on an even smaller island nearby (1044). Rousseau has his books sent and enjoys leaving the boxes shut („Un de mes plus grands délices étoit surtout de laisser toujours mes livres bien encaissés...“, 1042). There are no duties. A few unavoidable letters are written with a borrowed pen on the desk of a stranger – who would not be reminded of quickly checking e-mail in a holiday resort’s internet café? What did Rousseau do that made him so happy? First of all: nothing („le precieux far niente“, ibid.), and without regretting his laziness („oisiveté“, ibid.). Among the description of life on the island (section 3), two scenes are especially prominent. They contain two descriptions of the state of rêverie which both motivate the theoretical discussion (section 4) („soit couché ... soit assis....“, 1047): relaxing in a boat in the middle of the lake in the afternoon and sitting at the shore of the lake at nightfall. [J]’allois me jeter seul dans un bateau que je conduisois au milieu du lac quand l’eau étoit calme, et là, m’étendant tout de mon long dans le bateau les yeux tournés vers le ciel, je me laissois aller et dériver lentement au gré de l’eau, quelquefois pendant plusieurs heures, plongé dans mille rêveries confuses mais délicieuses, et qui sans avoir aucun objet bien déterminé ni constant ne laissoient pas d’être à mon gré cent fois préférables à tout ce que j’avois trouvé de plus doux dans ce qu’on appelle les plaisirs de la vie. (1044) Quand le soir approchoit [...] j’allois volontiers m’asseoir au bord du lac [...]; là le bruit des vagues et l’agitation de l’eau fixant mes sens et chassant de mon ame

282

Niko Strobach

toute autre agitation la plongeoient dans une rêverie delicieuse où la nuit me surprenoit souvent sans que je m’en fusse aperceu. Le flux et reflux de cette eau, son bruit continu mais renflé par intervalles frappant sans relâche mon oreille et mes yeux, suppléoient aux mouvemens internes que la rêverie éteignoit en moi et suffisoient pour me faire sentir avec plaisir mon existence sans prendre la peine de penser. De tems à autre naissoit quelque faible et courte réflexion sur l’instabilité des choses de ce monde dont la surface des eaux m’offroit l’image: mais bientôt ces impressions légéres s’effaçoient dans l’uniformité du mouvement continu qui me berçoit, et qui sans aucun concours actif de mon ame ne laissoit pas de m’attacher au point qu’appelé par l’heure et par le signal convenu je ne pouvois m’arracher de là sans effort. (1045)

These passages deserve a close look, since they are very precise and, although at first sight they look like pure phenomenology, they are heavily theory-laden. (a) Stretched out in a boat on calm water in the middle of the lake, his eyes turned towards the sky, sometimes for hours, Rousseau is immersed („plongé“) in „mille rêveries confuses mais délicieuses“ – delicious even without order. Rousseau contrasts them, implicitly and explicitly. The implicit contrast is with mental states like thinking, reasoning or defining. In contrast to them, the mental states called rêveries have no well-determined and constant object („aucun objet bien déterminé ni constant“). The explicit contrast is with the most enjoyable instances of what is usually called the pleasures of life („ce qu’on appelle les plaisirs de la vie“). Rousseau claims mental states called rêveries to be far more pleasing than the instances of usual pleasures. The theoretical section of the fifth rêverie will attempt to give a precise answer to the question why. (b) Then again, in twilight, at the shore of the lake, the sound of the small waves and the motion („agitation“) of the water would immerse Rousseau in delicious rêverie („dans une rêverie délicieuse“). The motion of the waves serves a triple function. (i) Function number one: The motion of the waves gets hold of Rousseau‘s senses and so, in a way, brings them to a halt („fixant mes sens“); even more importantly, the motion of the waves chases all motions („toute agitation“) from his soul. An alternative description of the same happening is: the state of rêverie extinguishes those motions („...que la rêverie éteignoit en moi“). Which motions? The soul’s internal motions („mouvemens internes“), i.e. the soul‘s usual reaction to a varied and inconstant perceptual environment; but also, pretty literally, the soul‘s troubles due to worries, unresolved questions, conflicts and the like. Good they are gone. But

Living like God in Switzerland?

283

something is missing now, namely precisely what usually makes the soul feel its own presence to itself: the internal motions of the soul that the motion of the waves has cradled away („berçoit“). However, the state of rêverie is not to be identified with having fallen asleep. It is a conscious state, i.e. a state in which the soul is conscious of its own existence. If this is so, something must have stepped in for its internal motions that have been calmed away. (ii) Function number two: The motion of the waves steps in for the soul’s internal motions and keeps it awake – not in the way of fighting against its losing consciousness, but supplying it with just enough input to feel its pure existence instead of its affections. The visible and audible regular flapping of the waves is a supplement for the internal motion which the rêverie has extinguished. But it is nothing to be thought about: It suffices for making Rousseau feel, with pleasure, his own existence without the effort of thinking („sentir avec plaisir mon existence, sans prendre la peine de penser“). (iii) Function number three: the motion of the waves provides Rousseau with an image of the instability of things during short intervals of receding rêverie which are, however, soon over again. Staying in the state of rêverie, sitting at the shore for hours, requires no active participation of the soul of willing to do so (it happens „sans aucun concours actif de mon ame“). 4. Theory The narration is suddenly interrupted and we get several pages of pure theory. Rousseau develops a conception of supreme happiness in order to be able to subsume under it the state of rêverie he has been describing. The argument is by contrast: Certain mental states are excluded as plausible candidates for supreme happiness. What remains as a plausible candidate is characteristically unlike them. The candidates which are discounted are states of lively („vif“) exaltation and passion („délire et passion“) (a, b) and then even external things („choses exterieures“) in general (c) for the following reasons: (a) The intrinsic liveliness („vivacité“) of states of exaltation and passion implies that they are not the kind of thing that lasts for long. Their nature excludes their being stable states and their being realized often and regularly. This is, however, not supposed to mean that the state of rêverie is

284

Niko Strobach

not a state of being alive. On the contrary! Just as it is very important to Aristotle, in Met. XII 7, to establish his Prime Mover’s mode of being, permanent actuality, as something which may be called a life, so Rousseau wants the state of rêverie to be a state of life (cf., e.g., p.1047: „Sans mouvement la vie n’est qu’une lethargie.“ „Un silence absolu porte à la tristesse. Il offre une image de la mort.“) So „vivacité“ is not a characteristic of all life. (b) While passing pleasure is rarely experienced, lasting pleasure is even hard to imagine. Concerning the usual candidates for our most lively pleasures („nos plus vives jouissances“) it is phenomenologically implausible that one would seriously want them to last forever even while experiencing them. But if that is so, they are not truly fulfilling but leave something else to be desired. They leave one’s heart troubled and empty („un état ... qui nous laisse encor le coeur inquiet et vuide“).3 (c) The affections of our soul depend on external things. Those things are unstable, and so are our affections which are about them. Having got hold of some external object cannot be supreme happiness since you can never be sure to keep it („il n’y a rien là de solide à quoi le coeur se puisse attacher“).4 By contrast, supreme happiness should be a fulfilled state which leaves no desire open („...d’un bonheur suffisant, parfoit & plein, qui ne laisse dans l’âme aucun vide qu’elle sente le besoin de remplir“, 1046). It should have the potential of being a permanent state. It cannot be about external things, since their mutability would threaten its permanence. More than that, it should be a state which is such that being in it includes the awareness of its permanence, which is appreciated as part of the pleasure: [L]e bonheur que mon coeur regrette ... [est] un état simple et permanent ... dont la durée accroit [= augments] le charme au point d’y trouver la suprème félicité. (ibid.)

3

Rousseau anticipates the crucial formula in Goethe’s Faust „werd’ ich zum Augenblicke sagen: Verweile doch! du bist so schön!“ (lines 1699f., cf. also lines 1158111584), „Je voudrois que cet instant durât toujours“, 1046). It is probable that the “troubled and empty heart” was influenced by Augustine, conf. 1.1, „Inquietum est cor nostrum donec requiescat in te“, which should be related to conf. 13.38.53 “tua quies tu ipse es”. 4 Augustine, conf. 4.4.7 to 4.10.16 contains an extensive statement of this traditional view – and, at the same time, movingly challenges its plausibility.

Living like God in Switzerland?

285

Each of these points may be disputed: It is not clear to me why supreme happiness should be a state whose permanence or non-permanence I must even be able to consider or appreciate while being in that state. Granted that I am considering the point while experiencing some lively pleasure, I am not sure that I would not seriously want it to last forever. One problem here is the underspecified scenario: Am I supposed to know I am going to be thirsty after a while? Does that exclude my wishing to last that state forever, while I am not yet thirsty? And finally, I am not sure why supreme happiness should not be about a transitory object if „supreme“ means „the best that is to be had (for humans)“. Anyway, Rousseau spells out the conditions for supreme happiness in positive terms as follows: Mais s’il est un état où l’âme trouve une assiette assez solide pour s’y reposer tout entiére et rassembler là tout son être, sans avoir besoin de rappeler le passé ni d’enjamber sur l’avenir; où le tems ne soit rien pour elle, où le présent dure toujours sans néanmoins marquer sa durée & sans aucune trace de succession, sans aucun autre sentiment de privation ni de jouissance, de plaisir ni de peine, de desir ni de crainte que celui seul de notre existence, & que ce sentiment seul puisse la remplir tout entier[e]; tant que cet état dure celui qui s’y trouve peut s’appeler heureux, [...] d’un bonheur suffisant, parfait et plein [...]. (1046)

So supreme happiness is a state (a) in which the soul is assembled in its entirety, not reaching out into the past or the future (b) in which time is nothing to it (c) in which the present lasts forever without even making its duration or any succession felt (d) in which it feels no fear, lack or pain (e) in which it feels no other pleasure than that of being pleased at its own existence. And Rousseau states that this was indeed the state he was frequently in during his solitary rêveries („dans mes rêveries solitaires“). This is a state which is far from the state one would usually associate with the word „rêverie“. In fact, this is a state which is not sufficiently characterised by the lack of a well-defined object and by the lack of thinking in the sense of being mentally occupied with a well-defined object. The question is whether or not at the deepest point of rêverie there is any object of awareness left. Musing about nothing in particular is just the slow entry into this

286

Niko Strobach

state. It is this very special state, which has been called the state of rêverie in this paper, because this is the theoretically important part of the mental journey called rêverie in the broader sense. Rousseau makes (e) precise by saying that the soul in this state is stripped of („depouillé“) every other affection („toute autre affection“). Its object, that which the soul rejoices about, is nothing external to it but is the soul itself and its own existence („de soi-même et de sa propre existence“, 1047).5 This is an affection of the soul, too (namely a case of selfaffection), but none that requires anything else. If the soul were totally affectionless there would be no reason why it should be aware of anything. This explanation comes as no surprise after the theory-laden description of the twilight at the lake-shore. Recall that the two important functions of the motion of the waves were (1) extinguishing the motions of the soul which are affections of external things or troubled reactions to them (2) stepping in for those motions as minimal input, so there is something to keep consciousness going. The following question about the coherence of the twilight-scene and the part of the theory labelled as (e) might arise: Are the motions of the waves the intentional object of the mental state of rêverie? Clearly, that’s not what they are intended to be by Rousseau. And it is plausible enough that that motion is not consciously registered in the sense of the thought that there is a motion of waves on lake Bienne. That would spoil the whole story. However, the motions cause some new internal motion (after having extinguished all others) which, in a way, mirrors the motion of the waves. And it is that motion which makes the soul aware of its own existence, but, being very slight, of nothing else. This internal motion is the immediate object of consciousness in the state of rêverie (cf. also 1048: „Le mouvement qui ne vient pas du dehors se fait alors au dedans de nous“). This is coherent with the theory because it is a motion of the soul itself which informs the soul of its own existence. This internal motion, the only remaining one, is not being interpreted as a sign of the presence of anything external. In the sense of Schopenhauer ([1813 / 1847], §21) this is a retreat from applying the principle of sufficient reason (in this case: cause) to concious states. I think that Hadot (1995, 260) is wrong to claim: 5

Analytical ontologists might complain that this is categorically vague. But note that a soul as the bearer of the state of its existence is never called into question.

Living like God in Switzerland?

287

[T]he feeling of existence of which Rousseau spoke is precisely this feeling of identity between universal being and own existence.

Rousseau’s description is too precise to allow for such speculation. Neither is any universal being mentioned, nor any act of identification or awareness of identity of the soul with anything (except, possibly, itself). Rousseau is more ego-centric than Hadot would think. No loss of self, no Entgrenzung takes place, at least not according to the official description. For a point where it is nevertheless suggested (in a passage which is of dubitable coherence with the theory) see below. Point (d), freedom from pain etc. logically follows from all internal motions except one pleasant one having vanished. Points (a) to (c), however, are much more problematic. One wonders a bit why Rousseau includes them at all. This is particularly interesting, since they are about time and timelessness. Here is a guess, why, which, I think, is not far-fetched: Rousseau is well aware of the tradition, initiated by Aristotle, that being in a state of total fulfilment, while being occupied with nothing but one’s own existence, i.e. being in a state of complete self-sufficiency is associated with God’s mode of being: [T]ant que cet état dure on se suffit à soi-même comme Dieu. (1047)

Now, if self-sufficiency is a feature which the state of rêverie shares with God’s mode of being, why not assume also timelessness to be a feature of the state of rêverie, especially if tradition has closely associated both as features of God’s mode of being? It is probable that Rousseau thought so, whether fully conscious of this move or not. How plausible is this move? Answering this question really requires answers to two questions: (1) Is Rousseau right to assess what he has been describing as a state of complete self-sufficiency similar to the traditional idea of God’s mode of being? (2) Is he right to assess what he has been describing as a timeless state in the sense of (a) to (c)? It is impossible to go into detail about the metaphysical and theological tradition at this point. Let us focus on the most basic elements of the Aristotelian tradition just insofar as we need them for a rough comparison. But since there is something which moves while itself unmoved, existing actually, this can in no way be otherwise than as it is. [...] On such a principle, then, depend the heavens and the world of nature. And it is a life such as the best which we enjoy, and enjoy for but a short time (for it is ever in this state, which we cannot be),

288

Niko Strobach

since its actuality is also pleasure. [...] And thought thinks on itself [...] and the act of contemplation is what is most pleasant and best. If, then, God is always in that good state in which we sometimes are, this compels our wonder; and if in a better this compels it yet more. And God is in a better state. And life also belongs to God; for the actuality of thought is life, and God is that actuality; and God's selfdependent actuality is life most good and eternal. We say therefore that God is a living being, eternal, most good, so that life and duration continuous and eternal belong to God; for this is God. (Aristotle, Metaphysics XII 7, 1072b7 – 1073a30, tr. Ross)

Aristotle’s God is an unmoved mover of the universe. He is always in the same state, and necessarily so, neither in circular motion nor in any kind of self-motion, contra Plato’s Phaedrus (245c-246a). „Being always in the same state“ is, later on, interpreted not as being stable throughout time, but as existing in timeless eternity.6 Thinking about changing objects, Aristotle would say, would – absurdly – make him change his mind.7 And it would make him dependent on something else (Met. XII 9, 1074b15 – 1075a5). But if his state is to differ from sleep (!), he must be thinking about something (Met. XII 9, 1074b18). So the one thing which is the content of his thinking is his own thinking. This is the complete self-sufficiency Rousseau alludes to.8 Although changeless, God’s mode of being, pure actuality, may be called life;9 it is a state of contemplation (theoria); and it may be called a state of supreme happiness. It matters to Aristotle, contra Plato’s Philebus (33b), that this happiness is felt to be pleasant. How near can humans get to this? They, too, can theorize, even about unchanging objects, and that’s also their happiest and most pleasant state. But neither can their theorizing have the same content as God’s (i.e. God’s thinking), nor can 6

Cf. Thomas Aquinas ST I q10 a1-4. Cf. esp. ST I 10 a2 c.: “ratio aeternitatis consequitur immutabilitem”. Clearly, he holds that succession is excluded from eternity, ST I q10 a1 c.: “ipsa aeternitas successione caret, tota simul existens”. 7 Later on we find the refined idea that if he can see the changing objects as not changing, but all at once, there is no threat in this respect. It is not possible to follow the implication of this idea here. 8 For this point in Thomas Aquinas cf. ST I q10 a1 ad3 (“aeternitas dicitur tota, non quia habet partes, sed inquantum nihil ei deest”) in combination with a2 c. (“Deus nec solum est aeternus, sed est sua aeternitas”). Interestingly, this variant of the idea leaves some room for something else being on God's mind than himself, as, e.g., ST I q15 a2 c. makes clear. 9 Cf. for the same view in Thomas Aquinas ST I q10 a1 ad2: “illud qoud est vere aeternum, non solum est ens, sed vivens.”

Living like God in Switzerland?

289

they persist even in their comparatively low-level theorizing – they must sleep in between and die in the end. Humans, according to this picture, are never self-sufficient. Even when they theorize they don’t just think about themselves. Even this very crude summary suffices to make clear the boldness of Rousseau’s claim, but also its creativity. Like Aristotle, Rousseau would say that while God is forever self-sufficient, humans are capable of being so only for short periods of their specific supreme happiness.10 Rousseau suggests that this state is the state of rêverie. If theoria means thinking about well-determined objects, then this claim stands in sharp contrast to Aristotle’s. However, Aristotle sees clearly that the subject – object distinction is problematic at least in the case of divine theoria (possibly also in special cases of human theorizing, cf. Aristotle, de an. III 4 + 5). Aristotle, unlike Rousseau, does not claim that the mental activity which constitutes supreme human happiness is entirely self-reflexive. Only God’s is. How about the pure feeling of one’s own existence vs. thinking one’s own thinking? Both ideas are less far apart than one might think at first sight. If some being’s existence is thinking and even, in a way, objectless thinking, one might say that for this being thinking its own thinking must be the same as the pure feeling of its existence. However, it is remarkable that Rousseau would not have been able to use the word „thinking“ in this context, as he sharply distinguishes between „sentir“ and „penser“. This verbal point aside, one may state: Rousseau thinks that he analyses supreme human happiness as being of the same structure as divine happiness, as the Aristotelian tradition describes it. Is he right to think so? There is a point against identifying the selfsufficiency of the Aristotelian God with the self-reflexiveness of the state of rêverie, which is quite independent of the question of timelessness: The state of rêverie depends on the change of certain material objects, i.e. the motion of waves on lake Bienne or something else of the sort, although its content may be nothing but the soul’s existence. Things get even more problematic when time and timelessness are considered. It has become clear that at least changelessness is essential to the Aristotelian God, perhaps even timelessness. If such a God takes notice of 10

On Thomas Aquinas’ complicated view of different levels of participation in (God's) eternity cf. ST I q10 a3 c.

290

Niko Strobach

the world, His self-sufficiency seems to require being entirely at every occasion of being. Interestingly, Robert Pasnau has made clear that this „holochronicity“ does not imply atemporality (Pasnau 2011). So let us see if the state of rêverie has the kind of changelessness which would be required of an Aristotelian God throughout time (who may well be necessarily unaware of time’s passage). Since time passes around Rousseau on St. Peter's island, this is what his description of rêverie there should be compared with first. Something more about rêverie in a timeless environment will be said towards the end of this paper. Rousseau has a „phenomenological“ inner perspective on this: (a) The soul in the state of rêverie being assembled in its entirety in the present does not have to recall the past or envisage the future („sans avoir besoin de rappeler le passé ni d’enjamber sur l’avenir“) – and so it doesn’t; (b) time is nothing to it („le tems ne soit rien pour elle“), i.e. the soul is not aware of time’s passage, which it has forgotten. So its present lasts forever („le présent dure toujours“), not because it is felt to have any duration, but, as far as one can tell, just by not being opposed to anything that is over or yet to come. So the present (rather: time) does not make its duration felt („sans ... marquer sa durée“) and there is no consciousness of any succession. As far as I can tell, this is not an implausible inner phenomenology of a holochronic being. What I doubt is that, even for a while, someone in the state of rêverie, as Rousseau describes it, is such a being. It is true that a person in the state of rêverie has forgotten what the time is on a scale of hours. So in this respect Rousseau „wakes up“ upon the call for dinner (1045). So it is also true that such a person after a while does not know for how long the flapping of the waves has already been going on in her presence. It is true that in that state the soul does not reach out into the past or the future in the mode of caring about it.11 However I cannot persuade myself that there is no temporal consciousness whatsoever, no consciousness of any succession in whatever small window of awareness involved in rêverie, as Rousseau describes the state.

11

At this point, the text foreshadows the connection between retention and protention in Husserl’s phenomenology (Husserl 1985) and the concept of „Sorge“ in Heidegger (1967 [1927]).

Living like God in Switzerland?

291

(1) If there is to be a contrast with pleasures from external causes with respect to the Faust-test, it should be possible to say about a moment of rêverie „Je voudrois que cet instant durât toujours“ while being in that state (note the demonstrative: „cet instant“). How do you do that without being time-conscious at all? (2) If some internal motion mirrors the slight motion of the waves, there is indeed some change going on, and just this is required in the twilight scene for making the soul aware of its own existence. I don’t see how one could be aware of oscillation without some minimal consciousness of time and succession. How much of it? Rousseau himself knows best how to tell in a sentence which is rightly admired for its sound and rhythm: Le flux et reflux de cette eau, son bruit continu / mais renflé par intervalles frappant sans relache / mon oreille et mes yeux [...] (1045)

It is difficult to imagine just nothing to be happening. I don’t see why the twilight-scene should be an example of that. It is true that Rousseau discusses a case of extreme sensory deprivation in a dark prison („cachot“). He claims that the state of rêverie would still be attainable in that case (1048). But, in my view, this contradicts his more plausible claim that „repos absolu“ will prevent the state of rêverie (1047). 5. Complications towards the end of the text Remarkably, timelessness plays no more role (at least not in connection with the soul’s pure awareness of itself) towards the end of the fifth promenade. The remark on the divineness of rêverie is followed by a short digression (1047) on the relation of vita contemplativa to vita activa. There is a serious moral problem: It would not be good if everyone, especially those not mature for rêverie, started detesting active life („la vie active“) and its duties at the prospect of supreme happiness in rêverie. It is interesting to compare the different answers given by different monastic orders, say, Franciscans and Carthusians, to this problem. Rousseau’s quick way out is to apologize that, being kicked out off society and being of no use any more anyway, he has the right to indulge in rêverie and find happiness in contemplation. It is remarkable that he feels the need for some excuse.

292

Niko Strobach

After this digression we get a kind of recipe or practical manual for inducing the state of rêverie: Neither is a complete still („repos“) any use, it will make you sad and recall death; nor agitated or irregular motion. What you need around you is some uniform and moderate motion. Imagine something in a good mood („une imagination riante est necessaire“). You have to achieve some compromise between motion and rest, thinking and not thinking: slightly stir up the surface of your soul without troubling its deep regions, just enough for remembering yourself while forgetting your trouble.12 The change of sound is amazing. And it keeps changing, in quick succession addressing the topics of meta-rêverie and of the liberation of the soul from the body: (1) Even better than a first-order rêverie which is threatening to be a bit boring after a while („abstraite et monotone“) is the enhanced multi-colour second-order rêverie of a rêverie, which has a first-order rêverie as its central content, but is embellished by mental landscape painting. Since already the first-order rêverie is essentially self-reflexive, the second-order one is doubly self-reflexive. It is, by the way, structurally close to Aristotle’s “thinking of thinking” (noesis noeseôs, Met. XII 9, 1074b34). Even the reason for assuming it, i.e. independence, is the same, since it matters to Rousseau that he can always retreat to meta-rêverie, so nothing will lack him even though his enemies (by whom he feels to be constantly surrounded) can prevent his return to a place like St. Peter's island. The constant availability of his memories makes him independent of good luck and self-sufficient in any circumstances. If this is an advantage of meta-rêverie over simple rêverie, it shows that a certain ambiguity had been unresolved in the description of simple rêverie: Although simple rêverie leaves nothing to be desired while being enjoyed there is always a threat of its termination. So, as to fugitiveness, simple rêverie is not so much better off than usual pleasure as Rousseau would like his beatitude to be. (2) Rêverie is a state of liberation from earthly passions and enables the soul to imagine flying above the sky and to be already conversing with the celestial intelligences („les intelligences celestes“) like in times soon to 12

Sure enough, extinguishing desire in this way does not always work. This is part of the profound sadness of Fernando Pessoa’s last poem “The happy sun is shining...” (Pessoa 1989, 199).

Living like God in Switzerland?

293

come. Rousseau was, by the standards of his time, an old man when he wrote the Rêveries. In this respect, rêverie is a good start of leaving one’s mortal body („dépouille“). This is an amazing new point, since nothing of what was said before implied that rêverie could be independent of bodily preconditions. It raises an important topic: Obviously, the solitary walker appreciates social interaction, the right kind of company provided. This, however, is in principle excluded from the state of rêverie in the sense of pure self-consciousness described before. There is no hint of timelessness here. I shall not try to follow the train of thought of the last two pages of the fifth promenade, if there is any. Instead I will suggest a – somewhat speculative – answer to the question why, for about half a page, the text changes into a therapy manual; and consider the topic of self, before, very briefly, state a systematic lesson which might be learned from Rousseau’s text. 6. Calmness of the waters – calmness of the soul Why the „user’s guide to rêverie“? In my opinion, it brings to light a feature of the theory-laden description of rêverie which has been there all the time, but which is superseded by the quasi-theological interpretation which Rousseau himself provides. One might ask: Why waves? Certainly, because Rousseau experienced them. But his description of them is theoryladen. The external waves produce slight internal waves which replace their opposite: the troubles of affection. This points back to Hellenistic philosophy. Epicureans, Stoics and Skeptics, the three therapeutic „schools“13 describe supreme happiness as ataraxia, untroubledness, or as galênê (tês psychês), calmness of the soul.14 What matters here is that 13

In connection with the Skeptics, the word “school” must be taken cum grano salis. Cf., e.g., Sextus Empiricus, PH 1.4.10: ataraxia de esti psychês aochlêsia kai galênotês (tr. Bury: “ ‘Quietude’ is an untroubled and tranquil condition of the soul”); Marcus Aurelius, Meditations 12.22: ...hôsper kampsanti tên akran galênê, stathera panta, kai koplos akymôn ([you shall be] like someone who walks around a high bend [and now sees the calm water in a little bay], everything calm, and the bay without waves); Epicurus, letter to Menoeceus, §128, directly after introducing the term „tês psychês ataraxia“: hotan de hapax touto peri hêmas genêtai, lyetai pâs ho tês psychês cheimôn (when this once happens to us all the winter-storm of the soul vanishes). 14

294

Niko Strobach

“Hellenistic” waves, wherever they come from, are no atemporal phenomenon. Their function is contrast with storm, not the introduction of timelessness. Rousseau’s observation that „repos absolu“, if experienced, is a rather unpleasant state fits well the Epicurean imagery of reflections of sunlight on the calm sea, a little bit of glittering going on, but not unpleasantly much. 7. Loss of self? It is remarkable that Rousseau hardly explores an option of describing the special state he describes in a way that would seem commonplace today: as a loss or dissolution of the self, if only temporary. What we rather find is the (Augustinian) metaphor of descending into oneself („[L]e mouvement [...] trop fort [...] nous arrache d’au dedans the nous [...]“, 1047) and, of course, the soul as the stable bearer of internal motions. Throughout the fifth rêverie we get robust categorial soul-talk. There is only one passage where Rousseau is close to the „temporary loss of self“ metaphor: En sortant d’une longue et douce rêverie, [...] enfin ramené par degrés à moi-même et à ce qui m’entouroit, je ne pouvais marquer le point de séparation des fictions aux réalités. (1048)

Is it not strange that Rousseau, after having reached the state of feeling nothing but his own existence, has to be led back to himself? Should one say that this is a point where descriptions break down? Perhaps, but let us try, at the price of sounding terribly scholastic: A way of avoiding contradiction might be to distinguish between a minimal self and an ordinary self. So what Rousseau regains is his ordinary self, while he had been reduced15 to his minimal self. The German word „Entrückung” would, as nice as it is, not be totally apt, since it supposes transport without qualitative change of what is transported. Rather, a formula which captures the state a lot of the state of rêverie is „vollkommene Entbundenheit und Selbstgegebenheit” (Stierle 1980, 202). There might be an advantage of Rousseau’s manner of description over the talk of a loss of self: It is not clear (although this might be a point where one has to say „That’s the way it is“) why a dissolution of the self is 15

No negative connotation is intended here: think of the words “reduction” and “essence” when talking about cooking and sauces.

Living like God in Switzerland?

295

not at the same time the loss of consciousness. For isn’t consciousness always someone’s consciousness? The point is important, because it is related to the consciousness of time. An important point of Rousseau’s theory-laden description of the twilight-scene is that the small waves have to step in for the affection to provide some minimal input for the soul to make herself aware of herself. The minimal self, it seems, must retain some minimal consciousness of time. It is hard to see how it could in a timeless environment. 8. A very brief systematic result Rousseau’s description of the special state of rêverie is precise enough to appear incoherent if this state is supposed to be as atemporal as he himself would like it to declare. This leaves open several options for beatitude: (1) Since beatitude would have to be timeless, but in that case is indistinguishable from being unconscious and therefore no beatitude, the whole concept of beatitude is incoherent. (2) Beatitude is not a timeless state after all. There is enough temporality in it to allow for some glittering and gleaming. Who would not appreciate that? The minimal self needs minimal time. What kind of time? Possibly a circular time. Not the huge cycles of recurrence, which are usually suggested as an illustration, but rather a tiny time bubble of, say, three or four waves in diameter. Tense logics for circular models exist. Circular models of tense logic do have successive instants, and there is something before and after each instant. But one will have to say either that the distinction between past and future breaks down, so there is neither past nor future; or that, while there are past and future, an event token's being future does not exclude its being past, but, on the contrary, implies it; and its being past implies its being future (cf., e.g., Burgess 1979, 572). Would it be boring? No, if your thoughts, including your meta-thoughts, were not numerically the same, your consciousness’ environment wouldn’t be circular time, since something would by qualitatively and thus, by indiscernibility of identicals, be numerically different from cycle to cycle. The fact that such models can be described does not provide an answer to the question whether they deserve to be called models of time. So much is clear: While for every wave which hits a certain point of the shore of lake Bienne on a certain day there is a certain n such that it is the

296

Niko Strobach

n-th wave that does so on that day, there would be no such n for the analogue of a wave in the time bubble - if not precisely that would prevent calling it a time bubble. Or are there (3) states we simply cannot (yet) imagine? In any case, for the time being it seems reasonable to distinguish carefully between minimal, possibly circular, time on the one hand and a completely changeless present on the other. References Adeleye, G. et al. 1999 World dictionary of foreign expressions: a resource for readers and writers, Wauconda, IL: Bolchazy-Carducci. Aristotle 1907 De anima with translation, introduction and notes by R.D. Hicks, Cambridge: Cambridge University Press [= de an.] -

1908 Metaphysica translated by W.D. Ross. The Works of Aristotle vol. VIII, Oxford: Clarendon Press [= Met.].

Augustine 1912 Confessions, 2 vols. translated by W. Watts (Loeb), Cambridge, MA: Harvard University Press [= Conf.]. Burgess, J.P. 1979 “Logic and Time”, Journal of Symbolic Logic 44, 566-582. Epicurus 1973 Letter to Menoeceus, Greek text in G. Arrighetti (ed.), Epicuro Opere, 2nd edition, Turin: Einaudi. Goethe, J.W. 1991 Faust. Der Tragödie erster und zweiter Teil, E. Trunz (ed.), München: Beck. Hadot, P. 1995 Philosophy as a Way of Life: spritual exercises from Socrates to Foucault, translated by M. Chase, Oxford: Blackwell. Heidegger, M. 1986 [1927] Sein und Zeit, Tübingen: Max Niemeyer. Husserl, E. 1985 Texte zur Phänomenologie des inneren Zeitbewusstseins (1893 1917), Hamburg: Meiner. Marcus Aurelius 1916 Meditations translated by C.R. Haines (Loeb), Cambridge, MA: Harvard University Press. Pasnau, R. 2011 “On Existing All at Once” in E. Runggaldier and C. Tapp (eds.), God, Eternity, and Time, Farnham: Ashgate, 11-28. Pessoa, F. 1989 Esoterische Gedichte. Mensagem / Botschaft. Englische Gedichte, G.R. Lind (ed.), Zürich: Ammann. Petrarch 1933 Familiarum rerum IV 1. Francesco Petrarca, Le Familiari, V. Rossi (ed.), vol. 1, Florence: Sansoni. Plato 1925 Philebus, in Plato, Statesman – Philebus – Ion, translated by H.N. Fowler and M.R.M. Lamb (Loeb), Cambridge, MA: Harvard University Press.

Living like God in Switzerland? -

297

1914 Phaedrus, in Plato, Euthyphro – Apology – Crito – Phaedo – Phaedrus, translated by H.N. Fowler and M.R.M. Lamb (Loeb), Cambridge, MA: Harvard University Press.

Rousseau, J.-J. 1959 Œuvres complètes, B. Gagnebin and M. Raymond (eds.), vol. 1, Paris: Bibliothèque de la Pléiade. -

J.-J. [1782] Les rêveries du promeneur solitaire, Cinquième promenade, in Rousseau 1959, vol. 1, 1040-1049.

-

J.-J. [1782 / 1789] Confessions, in Rousseau 1959, vol. 1, 1-656.

-

J.-J. 1979 Reveries of the Solitary Walker, translated by Peter France, London: Penguin.

Schopenhauer, A. 1988 [1813 / 1847] Über die vierfache Wurzel des Satzes vom zureichenden Grunde. Schopenhauer, A. Werke in fünf Bänden, Ludger Lütkehaus (ed.), vol. 3, 9-168, Zürich: Haffmans. von See, K. (ed.) 1980 Neues Handbuch der Literaturwissenschaft vol. 13: Europäische Aufklärung III (ed. J. v. Stackelberg), Wiesbaden: Akademische Verlagsgesellschaft Athenaion. Sextus Empiricus 1933 Outlines of Pyrrhonism translated by R.G. Bury (Loeb), Cambridge, MA: Harvard University Press [= PH]. Stierle, K. 1980 “Theorie und Erfahrung. Das Werk Jean-Jaques Rousseaus und die Dialektik der Aufklärung” in v. See (ed.), 159-208. Thomas Aquinas Summa Theologiae. Sancti Thomae Aquinatis doctoris angelici Opera omnia iussu Leonis XIII. P.M. edita, cura et studio fratrum praedicatorum, Rome 1888 [= ST].

299

The Authors

Lynne Rudder Baker is Distinguished Professor of Philosophy at the University of Massachusetts Amherst, USA. Her fields of research include metaphysics, philosophy of mind, and philosophy of religion. Recent publications: The Philosophy of Everyday Life (Cambridge 2007); Persons and Bodies (Cambridge 2000); Explaining Attitudes (Cambridge 1995). Johannes L. Brandl is Assistant Professor at the Department of Philosophy at the University of Salzburg, Austria. His main areas of research are in philosophy of psychology, philosophy of language, epistemology, and the history of philosophy in Austria. Recent publications “The Unmysteriousness of Consciousness: A Case Study in Naturalistic Philosophy”, in: G. Gasser (ed.), How Successful is Naturalism? (Frankfurt 2007); “Information, Intentionality, Experience”, in: A. Hieke / H. Leitgeb (eds.), Reduction. Between the Mind and the Brain (Frankfurt 2009); “What is Prereflective Self-Awareness? Brentano’s Theory of Inner Consciousness Revisited”, Revue Roumaine de Philosophie 55 (2011). Godehard Brüntrup is Professor of Metaphysics and the Philosophy of Mind and Language at the Munich School of Philosophy, Germany. His areas of interest include philosophy of mind and philosophical psychology, metaphysics, and philosophy of religion. Recent publications: “Panpsychism and Structural Realism”, in: M. Blamauer (ed.), The Mental as Fundamental. New Perspectives on Panpsychism (Frankfurt 2011); “3.5Dimensionalism and Survival – A Process-Ontological Approach”, in: G. Gasser (ed.), Personal Identity and Resurrection (London 2010). Antonella Corradini is Professor of Philosophy of the Human Sciences at the Department of Psychology of the Catholic University of Milan, Italy. Her main interests are in philosophy of psychology, in philosophy of mind, and in philosophy of the neurosciences. She is co-editor of Emergence in Science and Philosophy (New York – London 2010) and co-author of Modelli epistemologici in psicologia. Dalla psicoanalisi al costruzionismo (2nd enlarged edition Rome 2011).

300

Sergio Galvan is Full Professor of Logic at the Catholic University of Milan, Italy. He works mainly on Gödel’s theorems, on philosophical logic and on the ontology of abstract objects. Recent publications: “The Principle of Deontic Reflexivity and the Kantian Axiom”, Logique et Analyse 44 (2001); Einführung in die Unvollständigkeitstheoreme (Paderborn 2006); “Time and Reality of Phenomenal Becoming”, in: F. Stadler / M. Stoeltzner (eds.), Time and History (Heusenstamm 2006), Ontologia del possibile (Milan 2009); “The Emergence of the Intuition of Truth in Mathematical Thought”, in: A. Corradini / T. O’Connor (eds.), Emergence in Science and Philosophy (New York – London 2010). Christoph Jäger is currently Heisenberg Research Fellow at the Department of Christian Philosophy, University of Innsbruck, Austria. His areas of specialization include epistemology, metaphysics, late medieval philosophy, and philosophy of religion. Publications include: Selbstreferenz und Selbstbewusstsein (Paderborn 1999), “Göttlicher Plan und menschliche Freiheit: Vorsehung und ‚mittleres Wissen’ bei Luis de Molina”, Philosophisches Jahrbuch 117 (2010); “Reliabilism and the Extra Value of Knowledge” (with Wayne Davis), Philosophical Studies 2012, forthcoming. Ludger Jansen teaches Philosophy at the University of Rostock and the RWTH Aachen University, Germany. His research interests include ontology, philosophy of social and biomedical sciences, and ancient philosophy. Book publications: Tun und Können (Frankfurt 2002), Biomedizinische Ontologie (ed. with Barry Smith, Zürich 2008); Philosophische Anthropologie in der Antike (ed. with Christoph Jedan, Frankturt 2010). Christian Kanzian is Associate Professor at the Department of Christian Philosophy at the University of Innsbruck, Austria. His fields of research include metaphysics, ontology and the history of philosophy. Recent publications: Ding – Substanz – Person (Heusenstamm 2009); Soul, ed. with M. Legenhausen (Heusenstamm 2010). Gerhard Leibold is professor at the Department of Christian Philosophy at the University of Innsbruck, Austria. His fields of research include history of philosophy, especially medieval philosophy and the philosophy of Ludwig Wittgenstein, and theory of knowledge. Main publications: The edition of the Expositio in libros physicorum Aristotelis (libri II and VI) of William of Ockham (St. Bonaventure, N. Y. 1985); Johannes Duns Scotus,

301

Über die Erkennbarkeit Gottes (with H. Kraml and V. Richter, Hamburg 2000); Wilhelm von Ockham (with H. Kraml, Münster 2003). Winfried Löffler is Associate Professor at the Department of Christian Philosophy at the University of Innsbruck, Austria. His fields of research include metaphysics, philosophy of science and philosophy of religion. Recent publications: Einführung in die Religionsphilosophie (Darmstadt 2006); Einführung in die Logik (Stuttgart 2008). E. J. Lowe is Professor of Philosophy at Durham University, Great Britain. He is specializing in metaphysics, the philosophy of mind and action, the philosophy of logic and language, and the history of early modern philosophy. Some of his books include The Possibility of Metaphysics (Oxford 1998); The Four-Category Ontology (Oxford 2006); Personal Agency (Oxford 2008); More Kinds of Being (Chichester – Malden 2009). Uwe Meixner is a professor of philosophy at the University of Augsburg, Germany. He is the author of numerous books and papers on many topics in the field of theoretical philosophy. Main publications in English: Axiomatic Formal Ontology (Dordrecht 1997); The Two Sides of Being: A Reassessment of Psycho-Physical Dualism (Paderborn 2004); The Theory of Ontic Modalities (Heusenstamm 2006); Modelling Metaphysics (Heusenstamm 2010). Otto Muck is Emeritus Professor at the Department of Christian Philosophy at the University of Innsbruck, Austria. His fields of research include the dialogue between science and religion, metaphysics and its method, cf The Transcendental Method (New York, 1968) and the collection of articles: Rationalität und Weltanschauung (ed. W. Löffler, Innsbruck 1999). Josef Quitterer is Associate Professor at the Department of Christian Philosophy at the University of Innsbruck, Austria. His fields of research include Philosophy of Science, Philosophy of Mind and Bioethics. Recent publications: Agency and Causation in the Human Sciences (ed. with F. Castellani, Paderborn 2007); Die Aktualität des Seelenbegriffs (ed. with G. Gasser, Paderborn 2010).

302

Friedo Ricken is Emeritus Professor of Ethics and History of Philosophy at Munich School of Philosophy, Germany. His fields of research include ancient philosophy, general ethics, and philosophy of religion. Recent publications: Platon, Politikos. Übersetzung und Kommentar (Göttingen 2008); Warum moralisch sein? (Stuttgart 2010). Ruben Schneider is a fellow of the John Templeton Foundation and research assistant at the Munich School of Philosophy, Germany. His fields of research include analytic metaphysics, philosophy of religion and classical scholasticism. Recent publication: Kant und die Existenz Gottes (Münster 2011). Niko Strobach is Professor of Logic and Philosophy of Language at the University of Münster, Germany. His main areas of research are logic and metaphysics, in particular the philosophy of space and time, and also ancient philosophy. Recent publications: Alternativen in der Raumzeit (Berlin 2007); Einführung in die Logik (Darmstadt 2005, 22011). Christian Tapp is Assistant Professor of Interdisciplinary Studies in Philosophy and Theology at the University of Bochum, Germany. His research interests include logic, philosophy of science, metaphysics, and philosophy of religion. Recent publications: God, Eternity, and Time (ed. with Edmund Runggaldier, Farnham 2011); “Infinity in Mathematics and Theology”, Theology and Science 9 (2011); “Gotteswahn? – Die Philosophie zwischen Atheismus und Theologie”, in: A. Franz / C. Maaß (eds.) Diesseits des Schweigens. Heute von Gott sprechen (Freiburg 2011). Franz von Kutschera is Emeritus Professor at the University of Regensburg, Germany. He has published in almost all major fields of philosophy (from logic to the philosophy of religion) and its history (from Plato to Frege). Recent books: Philosophie des Geistes (Paderborn 2009); Wert und Wirklichkeit (Paderborn 2010). Paul Weingartner is Emeritus Professor at the Department of Philosophy at the University of Salzburg. Austria. His fields of research include logic, philosophy of science, and philosophy of religion. Recent Publications: Basic Questions on Truth (Dordrecht 2000, Russian tr. Moscow 2005); Laws of Nature (with P. Mittelstaedt, Berlin 2005); Omniscience, from a

303

logical point of view (Heusenstamm 2008); God’s Existence. Can it be Proven? (Heusenstamm 2010). Timothy Williamson is Wykeham Professor of Logic in the Faculty of Philosophy at the University of Oxford, Great Britain. His fields of research include logic, philosophy of language, metaphysics, and epistemology. Recent publications: The Philosophy of Philosophy (Oxford 2007); Contributions to P. Greenough and D. Pritchard (eds.), Williamson on Knowledge (Oxford 2009); “Necessitism, contingentism and plural quantification”, Mind 119 (2010).