Ontological Investigations: An Inquiry into the Categories of Nature, Man and Soceity 9783110329865, 9783110329698

Citation preview

Ingvar Johansson · Ontological Investigations

REPRINT PHILOSOPHY Modern Classics of Analytical Philosophy

Edited by Rafael HUntelmann · Erwin Tegtmeier · Käthe Trettin Volume 4

Ingvar Johansson

Ontological Investigations An Inquiry into the Categories of Nature, Man and Society

|VJ ontos verlag

Frankfurt « Lancaster

Bibliographic information published by Die Deutsche Bibliothek Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliographie; detailed bibliographic data is available in the Internet at http://dnb.ddb.de

c

2004 ontos verlag P.O. Box 15 41 »63133 Heusenstamm nr. Frankfurt www.ontosverlag.com ISBN 3-937202-42-0

2004

All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publisher. Printed on acid-free paper (TcF-Norm). Printed in Germany.

CONTENTS

Foreword to the second edition

VII

Preface

XV

1. Ontology

1

2. Irreductive Materialism

22

3. States of Affairs and Qualities

30

4. Exclusive and Inclusive Qualities

42

5. Actions and Functions

58

6. Patterns, Changes, and Pure Gestalten

85

7. Self-Sustaining Gestalten and Gestalten Causa Sui

95

8. External, Internal, and Grounded Relations

110

9. Existential Dependence

124

10. Container Space and Relational Space

145

11. Tendency

161

12. Efficient Causality

171

13. Intentionality

196

14. Nature: Parts and Wholes Without Intentionality

227

15. Man and Society: Nested Intentionality

258

16. Epistemological Positions

315

Notes

334

Bibliography

360

Index

367

Appendix 1: An aphoristic summary of Ontological Investigations

379

Appendix 2: Determinables as Universals

389

Appendix 3: Ontologies and Concepts. Two Proposals

411

Foreword to the second edition After fifteen years, a second edition of Ontological Investigations will now appear. It contains three appendices: First, a summary of the conclusions of the book in aphoristic form; second, a piece on universals which provides a more elaborate defence of my realist point of departure; and, third, an appendix on ontology in information science, a topic which is also addressed in this Foreword. 1. Ontology and information science At the time when Ontological Investigations was being written, the term 'ontology' was slowly entering the world of information science. It is easy to understand why. A philosopher ontologist is someone who tries to capture large chunks of the abstract structure of the world. In doing so, of course, he (or she) has to use concepts. But he is using these concepts only in order to say something about the world. He is not looking at concepts in order to find conceptual connections between them. When I say that the English word "cat," the German word "Katze," and the Swedish word "katt" are synonymous, then I am looking at the concept that the three words have in common. But when I make an assertion by using the words "The cat is on the mat," then I am using the concept cat in order to speak about a cat in the world. The assertions in this book are like "The cat is on the mat," though more general and abstract. Knowledge engineers in information science look at concepts and conceptual models. But they need classifications of very abstract concepts for their work, and, rather naturally, they have chosen to call these abstract classifications "ontologies," too. I once gave a brief talk at an ΑΙ-workshop about the difference between using concepts to denote entities in the world and viewing concepts as entities in themselves. It is here published as appendix 3. Out of every philosophical ontology, one can construe an ontology for knowledge engineers. Conversely, every information science ontology can (if only for the peculiar purposes of the philosopher) be regarded as an hypothesis about the abstract structure of the world and as such serve as an object of philosophical discussion. Now, just as ordinary engineers can often reach their practical goals more efficiently with theoretically obsolete

theories (such as Newtonian mechanics) drawn from the natural sciences rather than with more truthlike ones such as quantum mechanics, so it might well be the case that knowledge engineering can be done more efficiently with bad philosophical ontologies than with truthlike ones. At the moment, however, there seems to be nothing in particular that speaks in favour of such a view. (I find it nice that I am, while writing this foreword, working in a place - the Institute for Formal Ontology and Medical Information Science at the University of Leipzig - where the relationship between philosophical ontologies and information science ontologies has been put on the agenda.)

2. Ontology and political correctness At the beginning of sub-chapter 15.7, I report that "This theory of categories of mine is marxist in inspiration and (to a lesser extent) as regards the problems dealt with, but somewhat analytic in its way of handling these problems. The resulting ontology, however, is most closely related to the positions of two philosophers who belong neither to marxism nor to analytic philosophy: Aristotle and Hussserl." When I wrote the book I was a Marxist and a so-called Euro-Communist (that is, I believed in the combination of planned economy, democracy, and independence from the Soviet Union), and I did as much as I could to stress overlaps between my ontological views and those of classical Marxism. In particular, I did three things: I wrote the sub-chapter 15.7 on "Marxist dialectics and intentionality"; I stressed (in chapter 2) the fact that Friedrich Engels put forward a kind of level ontology; and (in chapter 16.1) that not only Karl Popper, but also Lenin, has claimed that truth can admit of degrees. I even coined the expression "the Lenin-Popper position on truth." Retrospectively, I can now see that I am almost certainly the only one who has ever found this expression funny. As far as I can see, one can without inconsistency combine the realist ontology of my book with both a conservative, a liberal, a socialist, and a feminist political philosophy. This notwithstanding the fact that during the 90s of the last century many philosophically-minded socialists and feminists were radical social constructivists who denied the existence of a mind-independent world about which knowledge can be obtained. Note that I am not saying that each and every part of an ontology has to be

VIII

ideologically neutral, parts of the ontology of human nature are probably not. Furthermore, if the belief that the whole of nature is imbued with mind or spirit is regarded as a political ideology, as a kind of "ecologism," then not even this book is ideologically neutral. It denies such a view. 3. Ontology and natural theology Most contemporary Western theologians seem to think that talk of God can have no ontological import whatsoever. Religion, they say, is "merely a language game." However, there are some religious philosophers and some theologians who think otherwise. They think one can rationally argue about the existence of God, even though, for them, the question is both normatively and emotionally loaded. That is, they believe in what the medievals called natural theology. Now, so do I; I regard it as part of ontology. My conclusion, however, differs from theirs. In my considered judgement, the spatiotemporal whole in which we live is for its existence not dependent on any "higher" and non-spatiotemporal existent. In my opinion and in this book, space-time is at the end of the ontological dependence line, though I do not anywhere in the book argue that this has to be the case. This fact, I want all prospective natural theologian readers to notice. 4. Ontology and analytic philosophy As a student in philosophy in Gothenburg, Sweden, I was trained and educated within analytic philosophy. This tradition, though, has been very heterogeneous. For instance, Cambridge analytic philosophy (in particular, Bertrand Russell) and Austrian analytic philosophy (i.e., followers of Franz Brentano such as Alexius Meinong) affirmed ontology, whereas logical positivism and Oxford analytic philosophy after "the linguistic turn" denied it. I wrote my book in the wake of the appearance (1978) of David Armstrong's two-volume book Universals and Scientific Realism, in which an anti-Platonist immanent realist position on universals is defended. I thought, quite mistakenly, that Armstrong's Aristotelianism would progressively conquer the Anglo-American and Scandinavian strongholds of analytic philosophy. Therefore, I took very lightly the defence of my realist point of departure, which comprises just a few pages in chapters 1.3

IX

and 1.4. I hope that in this edition this mistake is somewhat remedied by appendix 2 and by the three paragraphs below. Obviously, we regard our everyday mesoscopic world as perceivable and as containing numerically different things with many different properties. Often, several things are perceived as having the same property. In such perceptions there seems to be an immediate revelation to the effect that the world contains universals. If one perceives, let's say on a tennis court, three exactly similar spherical balls, one may find it obvious that, independently of language and of relations of resemblance, there are three different things that are qualitatively identical with respect to their shape. Apart from numerical identity and difference, the world seems to contain qualitative identity and difference; and where there is qualitative identity there is a universal. Every non-Aristotelian view of universals seems to contain some flagrantly non-commonsensical stain. The non-realisms of conceptualism and predicate nominalism imply a complete denial of mind-independent properties such as shape; everything that is universal is said to belong to concepts or predicates. The essence of the non-realism of trope nominalism is the astonishing view that every property is necessarily only a once-andfor-all occurrent. Platonist realism, on the other hand, posits the existence of universals outside space and time plus a peculiar relation of "participation in" between the spherical tennis balls in this world and a universal sphericality outside this world. Immanent realism, with its view that universals have a repeatable existence in - and only in - the ordinary spatiotemporal world, may seem to lack a corresponding remarkable peculiarity, but this is not quite true. When worked out in its details, such a realism has to claim that one single universal can be wholly present in several different things that are spatially or temporally scattered. Prima facie, this is a bit odd, too, which means that all positions on the problem of universals are odd. But they are not equally so. Immanent realism is the least odd position, and the existence of universals in re is the best explanation of the kind of property perceptions mentioned. According to Plato, all instances of a certain universal participate in "their" universal. According to my brand of immanent realism, it is rather the other way round: a universal participates in all its instances. Here come five views of mine:

X

(i)

there are universale, but there are no universale outside the spatiotemporal world; (ii) since a universal can have many instances, it cannot be identical with any of its instances; (iii) since a set does not exist in any of its elements (a set has elements), a universal cannot be identical with the set of its instances (or e.g. with a set of tropes); (iv) since a mereological sum does not exist in any of its parts (it has parts), a universal cannot be identical with the mereological sum of its instances (or with such a sum of tropes); (v) since a spatiotemporal aggregate does not exist in any of the aggregated units, a universal cannot be identical with the aggregate of its instances (or with any aggregate of tropes). Besides Armstrong's book, there is another great analytic-philosophical book that I would like to mention, namely John Searle's Intentionality (1983). Here, language is ontologically embedded. I read it in 1984, but, unfortunately, I only had time to comment on it in footnotes in the first edition of this book. Had it appeared some years earlier, I would have tried to stand on its shoulders in the way I tried to do with Armstrong's book (see chapter 1.3). Having paid this truly earnest homage to Armstrong and Searle, I want to make another point perfectly clear. Armstrong has defended immanent realism with respect to universale, but he has also (mistakenly) tried to reduce intentional states to ordinary material states. Searle, on the other hand, has (rightly) argued that intentional states are non-reducible entities, but he regards (mistakenly) the problem of universals as merely a conceptual muddle not worth the attention of serious philosophers. There is, though, one great philosopher who has defended both a kind of immanent realism for universals ("species") and the non-reducibility of intentional states, namely Edmund Husserl. Furthermore, Husserl laid bare the existence of different kinds of "relations of existential dependence," and the latter, as will become clear, are absolutely crucial to my ontology. In spite of his "transcendental turn," I regard Husserl as the greatest philosopher of the twentieth century. Back to analytic philosophy. I know neither who has coined the term analytic metaphysics nor exactly when it came into common use. But I

XI

know that it was not available to me when I wrote Ontological Investigations. Had it been, then I would have made clear that, while each chapter in the book can be regarded as a piece of analytic metaphysics, it would be nonetheless misleading to give the whole book such an epithet. Why? Because the book is an attempt at systematic metaphysics, too. In my opinion, analytic and systematic metaphysics are too distinct endeavours which should however be approached in tandem. Systematic metaphysics ought to be done the way analytic metaphysics is done, and systematic metaphysics is needed even by those who have no urge for a complete ontology; they need it as a consistency test of all the fragmentary presumed truths they have come to accept when visiting various corners of analytic metaphysics. If one has taken definite stands on the problems of, say, universale, causality, and the mind-body relation, then one ought also to ask oneself if the three positions taken are consistent with each other. If one concludes that they are, then one has entered the domain of systematic metaphysics. Armstrong, Searle, and Husserl, are systematic metaphysicians in this sense, and so also were Aristotle and Plato. 5. Ontology and epistemology Leibniz claimed (and I agree) that even an omnipotent God would have to stay within the framework of logical truths. Analogously, I would like to claim that even an omnipotent epistemology (which I do not believe exists) would have to stay within the framework of ontology. Just take a look at Descartes and Kant. Descartes claimed that philosophy ought to start with an all-embracing systematic doubt. But when he tried to doubt everything, he came to the conclusion that, for sure, a thinking enduring substance exists. Some of his contemporaries complained, rightly, that he should have confined himself to the conclusion that necessarily a "thought" or a "something" exists. My point is that one cannot rationally claim that nothing at all exists. Every epistemology has to start with some ontology. In Kant, this fact comes out in the following way. He claimed that philosophy ought to start with an investigation of the faculty of knowing; but, then, he presupposed the existence of such a faculty. As is clearly stated in the last chapter of the book, I regard substantive epistemological investigations as regional-ontological investigations, and I

XII

regard all regional-ontological investigations such as those made in the book as fallible. 6. Ontology and ontological thinking Since the publishing of the first edition of the book in 1989,1 have in some articles been able to make some of the ideas put forward more clear and distinct. These articles are listed below and related to different chapters. Chapter 1.1: • Hume's Scottish Kantianism, in the Polish journal Ruch Filozoficny LIX (2002), 421-53. (Here, in contradistinction to the book, I make a real attempt to unravel David Hume's ontology.) Chapters 1.3 and 1.4: • Determinables as Universale, The Monist 83 (2000), 101-21; see appendix 2. Chapter 2: • Hartmann's Nonreductive Materialism, Superimposition, and Supervenience, Axiomathes XIII (2001), 1-21. • Critical Notice of Armstrong's and Lewis' Concepts of Supervenience, SATS - Nordic Journal of Philosophy 3 (2002), 118-22. • The Asymmetries of Property Supervenience, in S. Lindström & Po. Sundström (eds.), Physicalism, Consciousness, and Modality, Umeä Preprints in Philosophy 2002:1,95-124. (In chapter 2, I ought to have mentioned the fact that the proposed ontology contains relations of supervenience, but I didn't.) Chapters 4.1, 4.5, and 11.2: • Physical addition, in R. Poli & P. Simons (eds.), Formal Ontology, (Kluwer: Amsterdam 1996), 277-88. • Presuppositions for Realist Interpretations of Vectors and Vector Addition, in U. Meixner & P. Simons (eds.), Proceedings of the 22nd international Wittgenstein Symposium 1999, (öbv&hpt: Wien 2001), Chapter 5: • Functions, Function Concepts, and Scales, The Monist 86 (2004), Chapter 6: • Pattern as an Ontological Category, in N. Guarino (ed.), Formal Ontology in Information Systems, (IOS Press: Amsterdam 1998), Chapter 8: • Marty on Grounded Relations, in K. Mulligan (ed.). Mind, Meaning and Metaphysics, (Kluwer: Amsterdam 1990), 151-56.

XIII

Chapter 13.5: • Intentionality and Tendency: How to make Aristotle up-to-date, in K. MuMijgan ^ed^), Language, Truth and Ontology, (Kluwer: Amsterdam Chapter 13.7: • Perception as the Bridge Between Nature and Life-World, in C. Bengt-Pedersen & N.Thomassen (eds.), Nature and Life-World, (Odense UP: Odense 1998), 113-37. Chapter 14: • The Unnoticed Regional Ontology of Mechanisms, Axiomathes* VIII (1997), 411-28. Chapter 15: • Stearic's Monadological Construction of Social Reality, in D. Koepsell & L.S. Moss (eds.), John Searle's Ideas About Social Reality, (Blackwell: Oxford 2СГОЗ), 233-55. 7. Ontological gratitude First, I want to thank the publisher and ontologist Rafael Hüntelmann for taking the risk of publishing this volume. Second, I want to thank Kevin Mulligan and Barry Smith for having helped me, supported me, encouraged me, and discussed with me for more than twenty years now, and for having urged me to try to get a second edition published.

Leipzig in February 2004, Ingvar Johansson

XIV

PREFACE

This book has a long history. It started in Lund in 1975 as a project called 'Explanations in the social and the historical sciences'. My hypothesis at that time was that philosophers of science are tackling problems on too abstract a level. The right solutions were to be found, I thought, in the concrete details of actual scientific theories. Gradually, however, I came to the opposite conclusion, namely that the philosophy of science was not abstract enough. I began to see that our most abstract presuppositions need rethinking. Whilst taking this 'ontological turn' I moved to Umeä, where the book was substantially completed in December 1985. Since then I have only added some minor changes and a few footnotes. Many acknowledgements are appropriate. First of all I want to thank Kevin Mulligan. Since we met at a conference in the summer of 1983, we have had the opportunity to spend all in all about five weeks together discussing ontology. These weeks have turned a sketch of an ontological system of mine into at least something like a unified whole. Kevin has an unusual ability to combine incisive criticism with helpful comments. His urge for clarity and rigour lives in peaceful coexistence with his delight in new and still loose philosophical ideas. There are two chapters where my debts are especially large: chapter 6 on patterns and chapter 9 on existential dependence. Neither would have been written at all without Kevin. He has also suggested some key terms, for example, 'nested intentionality'. And last but not least, he was kind enough to take on himself the work of checking my English. Svante Nordin read all my first manuscripts in Lund and made

PREFACE

many very helpful comments. In U m e i , Bo Dahlbom read my first complete manuscript in 1982-3, and showed me just how incomplete it was. Barry Smith has made a lot of detailed comments, some of which are very important, on the whole book. The same goes for Peter Simons with respect to chapters 1-8. Dick Haglund has read and given useful comments on chapter 10 and Anders Pettersson on chapter 13. Sören Hallden in Lund encouraged me to start working with my ideas on explanations in the social sciences. Martin Edman and his seminar in Umeä provided many useful comments. I have a special indebtedness to the late Ivar Segelberg, who introduced me to the ontological way of thinking in the mid-1960s in Gothenburg. Parts of my system have some affinity with his 'quality-instance' ontology (cf. note 13 to chapter 1); Segelberg's writings have now at last been translated and will probably soon appear in English. Thanks are also due to Sylvia Benckert and Stig Lindgren for kindly answering questions about what is happening in physics today, and to Craig Dilworth for help with the translation. The Swedish Council for Research in the Humanities and Social Sciences financed the start of the project mentioned above, as well as the work of checking my English; they also provided a publication subsidy. Anders Karitz's Foundation provided me with a travel grant. I am very grateful to Karin Knecht for accurate and fast typing, but most of all for some intense work when, for different reasons, I was in a real hurry.

XVI

1

ONTOLOGY

This book is a book about the world. I am concerned with ontology, not merely with language. Many ontological treatises concentrate largely on the nominalist-realism problem or try to pin down in minute detail some ontological category. Such undertakings are necessary, and will be met with below. Equally important, however, is the need to develop a coherent system of all the most abstract categories needed to give a true description of the world. Whether we are nominalists or realists this should be regarded as one of the central tasks of philosophy. At any rate it is the task of this book. 1.1 T H E O R I E S O F CATEGORIES My use of the word 'category' is intended to remind the reader of Aristotle's famous book Categories. Originally, the Greek word corresponding to the English 'category' meant 'predicate'. Predicates describe properties. A category, however, is not just any old property. Properties can be arranged according to their degree of abstraction. To take a simple example, the properties light red and dark red, which seem to be concrete properties, are nevertheless abstract in relation to all the different shades of colour which are called 'light red' and 'dark red' respectively. But, of course, taken together, light red and dark red constitute a more abstract property, namely red. And this property in turn, together with properties like blue and yellow, constitutes the property colour, which is more abstract still. Climbing up the ladder of abstraction, we will at the end reach the property of just being a quality. If this really is the end, then we have found a category. In this manner we can also 1

ONTOLOGICAL INVESTIGATIONS

find other categories. By means of predicates like 'chair', 'stone', 'man', 'horse', 'electron', it is possible to arrive at the category of substance. Concepts like 'longer than' and 'father of' indicate in the same way the category relation. Aristotle lists the following ten categories: Substance Quantity Quality Relation Place Date Posture State Action Passivity 1 T h e difference between 'posture' and 'state', and between 'action' and 'passivity', might be in need of some explanation. T o be sitting and to be lying are instances of posture, while having shoes on and being armed are instances of states. T h e former relate to the body in itself, the latter take something more into account. To cut is an instance of action, but being cut is an instance of passivity. Aristotle's Categories is in my opinion more of a Kategorientafel (table of categories) than a Kategorienlehre (theory of categories). This book is intended to be a Kategorienlehre in the strict and narrow sense, that is a realist theory of categories regarded as real aspects of being. This is the view of Aristotle, and the view subscribed to here. In a loose sense of the word, however, a Kategorienlehre can be extracted from the works of every great philosopher. In the case of a nominalist like David Hume the table of categories might be: Impressions Ideas Principles of associations: Resemblance Contiguity Causation 2 In order to get such a list from Kant, we simply disregard the distinctions between the noumenal world, the world of experience, and the realm of the transcendental. Both his 'forms of intuition' 2

ONTOLOGY

and his 'categories of understanding' can then in a loose sense be considered categories of the world. We then obtain the following list: Forms of intuition: Space Time Categories of understanding: Quantity Unity Plurality Totality Quality Reality Negation Limitation Relation Inherence and subsistence (substance and property) Causality and dependence (cause and effect) Community (reciprocity between agent and patient) Modality Possibility - impossibility Existence — non-existence Necessity - contingency3 A comparison of the category systems of Aristotle and Kant is illuminating first of all because it shows that a category system is not necessarily confined to our most abstract concepts. It might, as in the case of Kant, also include somewhat less abstract concepts. For some of Aristotle's categories (substance, action, and passivity) seem, in Kant's list, to have become subcategories of the category of relation. Another interesting point is that two of Aristotle's categories (posture and state) are quite absent from Kant's list. This is pardy due to the fact that Kant may be said to make a very sharp distinction between subject and object, a distinction which is not to be found in Aristotle. Subjects act and are in some sense free, in contradistinction to objects, which fall under a deterministic natural scientific causality principle. The category of subject is not listed by Kant, because according to him, subjects cannot be found in the world of experience: here determinism reigns supreme. 3

ONTOLOGICAL INVESTIGATIONS Subjects exist, as K a n t sees it, only in the so-called n o u m e n a l world, a world about which we can have no knowledge. N o t h i n g like this is to be found in Aristotle, who in a sense regards all objects as subjects. Every object obeys a teleological activity principle. But j u s t as some of Aristotle's categories are not regarded as categories by K a n t , so also some of K a n t ' s categories are not regarded as categories by Aristotle. Especially noteworthy is the category of being or reality. According to Aristotle, categories are aspects of being, and so being itself cannot be one category a m o n g others. F r o m this we can see that a table of categories need not list all the abstract concepts of the metaphysical system of which it is part. A system of categories may allow relations to exist between the categories, (i.e., relations apart from the relation of simple differences). K a n t , for instance, holds that there is a specific relation hidden behind his triplets of subcategories. T h e first two subcategories taken together are supposed to yield the third. Totality, to take an example from the category of quality, is to be regarded as constituted by a synthesis of unity a n d plurality. T h i s line of thought is taken to its extreme in another famous category system, t h a t of Hegel in whose Logik all the categories, subcategories, a n d sub-subcategories unfold in a long chain, which m a y be represented as follows: Being (quality —» quantity —» measure) —» Essence (essence as reflection into self—» appearance—» reality) —* C o n c e p t (subjectivity —* objectivity —» the idea) 4 K a n t ' s categories, relation, substance, causality, a n d reciprocity, have in Hegel's system become sub-subcätegories, that is, subcategories of the subcategory of reality. The arrows are m e a n t to indicate that the categories derive from each other in both a logical a n d a n ontological sense. Hegel is a thoroughgoing idealist, so his categories can merge with reality in a way that is not open to Aristotle (and not even to K a n t ) . I t is this idealism which makes it possible for him to treat concept as an ultimate category. O n e way of obtaining a deeper u n d e r s t a n d i n g of theories of categories is to look at Gilbert Ryle's concept o f ' c a t e g o r y - m i s t a k e ' . His example, by now classical, is of a foreigner visiting Oxford, w h o is shown the colleges, libraries, a n d so on. 5 W h e n the sight-

4

ONTOLOGY

seeing is over the foreigner asks 'But where is the University?'. The question has as its presupposition the idea that the University is some kind of building among the others. The man does not understand that the University is the functional totality of what he has been shown. In effect he regards universities as instances of the category of substance instead of instances of the category of function. Another example of a category-mistake is the question 'What colour has the number five?'. Numbers are not instances of things, and cannot have sensible qualities such as colours. A category-mistake represents an absurdity or an a priori falsehood rather than something which is a posteriori false. If you see something as a category-mistake, you do not feel in need of empirical evidence for your view. You just understand that something is wrong. A category-mistake represents the unthinkable. If you are to convince someone that he has made a categorymistake, you must show him that he has to think in another way. A theory of categories — as a by-product - draws the line around the thinkable. Such a line, however, should not, many philosophers notwithstanding (among them Ryle), be regarded as immutable. It can change, and it has changed through history. In this respect it behaves like science and common sense. (Such changeability is discussed in chapter 16.) 1.2 C O N T E M P O R A R Y O N T O L O G I C A L P R O B L E M S The traditional ontological problems, going back to the philosophers of ancient Greece, are by no means resolved. This book can, for instance, be seen as an attempt at a resuscitation of Aristotelian thinking on ontology. But the problem which has been the driving force leading up to the theory of categories to be presented, is of very modern origin. In my opinion, the outstanding ontological problem of today is how to situate society in nature. Modern physics, together with the theory of evolution has, in an irreversible way, undermined all idealist interpretations of nature. The emergence of the social sciences, on the other hand, has given rise to the idea that society is some kind of'second nature'. The latter term is explicitly used by Marx, but implicitly presupposed by most social scientists. As long as ontologists do not take the problem of the relation between nature and society seriously, I think we will have to live with an unsolved conflict between those 5

ONTOLOGICAL INVESTIGATIONS

who regard man and society as something distinct from pure nature and those who, for reasons of science, think that at bottom everything is actually bundles of elementary particles. I n order to get a better understanding of nature, man, and society — and especially of the relations between them — I think we have to rethink our most abstract categories. Now, of course, many philosophers even in our time have proposed new categories, a n d some of these (for example, tendency and intentionality) will have a prominent part to play in this book. But what is usually missing is the insight that the time is ripe not merely for modifying or reintroducing some category, but for creating a new categorial system. Some philosophers have been interested principally in the social sciences and have neglected the natural sciences, or have perhaps used a preconception of the latter merely as a foil for discussing the former. Others have taken things the other way round. In my opinion we have simultaneously to revise both our view of nature and our view of man and society. O n e has to try to grasp the real categories needed by both the natural a n d the social sciences, even if it looks like an undertaking which no m a n can perform alone. It is today impossible to have a firm grasp even of the essential ideas in all the sciences, but the attempt to create a new categorial system must none the less be made. Traditional theories of categories can be classified along at least two different dimensions. They are, first, either atomistic or holistic; and, second, they either contain or lack the category subject. T h e thesis of ontological atomism is essentially the proposition that facts are not existentially dependent on each other. ' E a c h item can be the case or not the case while everything else remains the same', to paraphrase 1.21 of Wittgenstein's Tractatus.6 Ontological atomism is usually accompanied by an epistemological presupposition to the effect that the understanding of the world as a totality is secondary to the understanding of its parts. As the parts can exist independently of each other, so they can be understood independently of each other; and then it is as if the understanding of the totality equals the sum of the understanding of the parts. Holistic views, in contrast, maintain that any particular item can be what it is only if all other items are what they are. T h e part is existentially dependent on all other parts. It cannot disappear or change without something else being changed.

6

ONTOLOGY

Because of this, it is also impossible truly to understand a part without at the same time grasping everything else, i.e. the totality. Truth, to cite Hegel, is totality. An atomistic ontology imposes some categories and outlaws others. Now, of course, the same is true of holistic ontologies. Different things become thinkable in different ontologies. The opposition between atomistic and holistic ontologies is, however, as has been said, not the only relevant opposition, especially not when discussing man's place in and relation to both nature and society. The other great opposition is that between traditions which allow no place anywhere in science for the subject category, and those which maintain that the subject category captures precisely what is specific to the social sciences and humanities. On the one hand we have the philosophers of the French enlightenment, nineteenth-century positivism, logical positivism, behaviourism, and structuralism; on the other we have the Romantic philosophers, neo-Kantianism, existentialism, and hermeneutic philosophy. Kant could not think of a subject as inhabiting the world of experience, and, in consequence, he could find no place for the subject in science either. Ryle could not think of a subject as a substance. The hermeneutic philosophers cannot think of a world at all without subjects. That the subject category has become enigmatic is easy to understand, even independently of calling in popular reductionist views within specific sciences, such as behaviourism within psychology and epiphenomenalism within neurology. Obviously, we have two different intuitions tending in opposite directions. Our presently available knowledge of history makes certain patterns appear to us irresistibly from changes of fashion to wars and revolutions. Such patterns cry out for explanation, and even for deterministic explanation. History does not appear to modern man to be merely a sum of chance events. At the same time, however, we all seem to know that behind these patterns with their multitude of actions there are people - people capable of free action. It is difficult not to ascribe some kind of freedom to humans making them differ in kind from at least inanimate nature. But the subject category, which incorporates this peculiar freedom, seems impossible to reconcile with the thought of completely lawlike patterns in history. The theory of categories to be presented here might be seen as 7

ONTOLOGICAL INVESTIGATIONS

an attempt to surpass both the traditional opposition between atomism and holism and the opposition between pro-subject and anti-subject category systems. In neither case am I the first to try, but I do not know of any philosopher who has taken as his task finding a true via media between both oppositions at once. 7 I regard the ontological problems just mentioned as analogous to the problem of motion viewed from the perspective of Zeno. Zeno's paradoxes were intended to show that although we do see movement, we cannot think it. And Zeno assumed what most category systems presuppose: that what cannot be thought cannot really exist. He thereby reached the conclusion that movement is an illusion. One of his paradoxes, the so-called Race Course, can be stated in the following manner. Assume that you intend to walk from one place, A, to another, B, along a straight line between them. In order to come to В you have to pass the midpoint between A and B, but, in order to do this, you have first to pass the midpoint between A and the aforementioned midpoint. A distance can be halved infinitely many times. There is no end to the division, because when you divide a distance, however small, you will always get new distances. This means that you can construct infinitely many 'midpoints' between A and B, which, in turn, means that in order to be able to walk from A to В you have to be able to traverse infinitely many intervals. According to Zeno, however, this is impossible. In order to traverse infinitely many intervals you need an infinitely long time, which, of course, means that you will never arrive at B. It is impossible to think movement. Zeno's conclusion is not inevitable. T o see this, one has only to maintain that Zeno's reasoning power was insufficiently great and that, in contradistinction to his views, it is possible to think movement. But in order to have any force, such a claim also has to show in some detail what the true thinking of movement is like. This was not done until Newton and Leibniz constructed the differential and integral calculus, about 2,000 years after Zeno. In this kind of mathematics, it is possible to handle infinitely small magnitudes; and Zeno's paradoxes deal with infinitely small distances. With the differential and integral calculus it became possible to think, without any contradiction, that a body at a particular moment can be in a certain place, and, notwithstanding this, also at exactly the same moment, have a certain velocity. Moment here literally means point of time. Were it to mean a 8

ONTOLOGY

period of time, however small, then a contradiction would actually arise. A body cannot during a period of time be both in movement and at a particular place. In order to be able to think movement man had to be able to think new categories; a completely new concept of point had to appear. A real effort was needed; as I said, it took 2,000 years. Newton and Leibniz taught us to think what we perceive. As I see it, this is what we still have to learn with regard to the relations between nature, man, and society. From the point of view of the category system here put forward, the emergence of modern physics and the decline of the Aristotelian world view had, for the historical and the social sciences, two unfortunate consequences. Modern physics, first, removed the subject category from the natural sciences. In a long tradition that begins with the French Enlightenment this has been taken to imply that the subject category should be banished from all the sciences. T o take seriously the category of the subject was regarded as involving a kind of prescientific and anthropomorphic way of thinking. Second, modern physics taught us that science, in order to develop, sometimes has to go beyond the limits of common sense; Aristotelian physics seems to be closer to the perceptions of children than is the mechanistic world picture. As the psychologist Jean Piaget has argued, children seem to pass through an Aristotelian stage before they internalize our world view. 8 The criticism of common-sense conceptions, however, has been carried to such extremes that very few philosophers have noted the importance of a point made above: that scientific progress may sometimes consist in learning to think what we perceive. This, in spite of the fact that even the emergence of modern physics contained an example of just this phenomenon, viz. the development of the concept of movement. Sometimes science and philosophy have to transcend common sense, sometimes they have to defend it. There is, however, no foundational transhistorical common sense. Post-Galilean differs from pre-Galilean common sense. There is an interaction between science, philosophy, and common sense, but very seldom does this interaction cancel all conflicts between these perspectives. In particular, there are conflicts today. I think today's common sense embraces both a naive realism and the view that there are subjects, views which are partly in conflict with today's science realistically

9

ONTOLOGICAL INVESTIGATIONS

interpreted, but views which I shall later defend below (see sections 13.7 and 15.2, respectively). One aim of this book is to create an ontological system which does justice both to the subject category and to naive realism without making science a mere instrument of prediction. In other words, I am trying to effect a specific interaction between today's philosophy, science, and common sense. An interaction which I think takes us closer to truth. Now some words in anticipation. In order to understand man and society we shall have to understand a category of intentionality, but in order to understand intentionality it is necessary to contrast it with a concept of tendency. This latter concept has been neglected in modern philosophy because it has been wrongly assumed that it is excluded from Newtonian physics. There is a growing literature on both 'intentionality' and 'tendency' into which my discussion will neatly fall.9 The concepts of intentionality and tendency are necessary for understanding the concept of action, but they are not sufficient. One of my main theses is that it is not possible to obtain a philosophical understanding and acceptance of our concept of action until we have rethought the concept of the momentary instant, which plays such a prominent role in the whole of modern physics. To anticipate once more, I think the concept of the momentary is ontologically inevitable, and I also believe that, for many reasons, it has come to swallow up other equally necessary concepts referring to phenomena extended in time. The latter kind of concept is necessary for understanding action, but it is also necessary for a proper understanding of nature and the natural sciences. In order to understand the part-whole relation needed in the social sciences, we first have to understand the kind of categorial part-whole relations presupposed by physics. When I talk about modern physics, I am using this expression in the sense in which it contrasts with medieval or Aristotelian physics. Thus 'modern physics' includes relativity theory and quantum mechanics, though it is not confined to these. Speaking about relativity theory and quantum mechanics might seem frightening, especially in view of the long discussions of parts and wholes in quantum mechanics. I should therefore say at once that in spite of my occupation with the problem of parts and wholes, I shall in this book totally neglect quantum mechanics and its 10

ONTOLOGY

philosophical problems.10 My justification for this is my belief, argued for later on, that not even the part—whole problem within Newtonian physics has hitherto been correctly understood, which means that, like the discussions within the philosophy of the social sciences, the philosophical discussions related to quantum mechanics take place against a background of false assumptions. With regard to relativity theory, things are a little different. The category system I put forward tries to rehabilitate a 'container concept' of space and time. As many people, scientists and philosophers alike, take relativity theory to imply the negation of such a concept, I have to give some arguments for my contrary view. However, in the first chapters a container conception of space and time is merely presupposed, and those who accept such a concept may simply skip those parts where relativity theory is discussed (parts of chapter 10). 1.3 UNIVERSALS IN RE The theory of categories to be put forward takes its departure from a realist conception of universals. More specifically, it is the Aristotelian conception of universals as universals in re. This means, to quote David M. Armstrong, that 'Universals are nothing without particulars. Particulars are nothing without universals'.11 Everything that exists in the world has both a particularity-aspect and a universality-aspect, and there are no universals outside the world. In contradistinction to Plato's transcendent realism, this is an immanent realism. The view that universals exist in particulars is, to say the least, not the commonly accepted conception nowadays.12 In spite of this, however, I do not intend to argue at length for my position. There is, even in the perennial disputes within philosophy, a certain sort of development which makes it possible to take a stand on other people's shoulders. In the case at hand, I am going to rely on Armstrong's revival of universals in re, that is to say, on the two volumes of his Universals and Scientific Realism. As Armstrong rightly points out, the opposite of realism about universals ought really not to be labelled 'nominalism' but 'particularism'. Realism says that there are universals, particularism denies this and says that there are only particulars. This denial is compatible with different versions of what the seeming universality 11

ONTOLOGICAL INVESTIGATIONS

consists in. Nominalism in the true sense is one such version. Armstrong, however, follows tradition and retains the label 'nominalism' but distinguishes between five kinds of nominalism: predicate nominalism (i.e., true nominalism), concept nominalism, class nominalism, mereological nominalism, and resemblance nominalism. I regard all these nominalisms as false, and refer to Armstrong for detailed arguments. I shall merely make three comments which I myself regard as fundamental. First, if one is a materialist and thinks that there is a world independently of all consciousness, then it seems incredible to maintain that a property such as extension did not exist until man created the corresponding predicate or concept. If there is a world independently of man there must be universals independently of man. Second, if all the things which we believe have the property of extension do not have such a universal property in common, but are merely brought together by the general predicate or concept 'extension', then such a bringing together must rely on the relation 'falling under the predicate (or concept) extension'. This relation, however, must then be a universal in its turn. And so the problem has just moved one step. The nominalists have to face an infinite regress here. Third, many attempts to get rid of universals rely on a presumed fundamental relation of similarity or resemblance. Many nominalists try to define all universals by means of particulars and the relation of resemblance. Extension, for instance, is then nothing else than a class of particulars resembling one another. Such a view contains at least two obvious difficulties. First, 'resemblance' seems to refer to a universal, and so one has not got rid of all universals; and if one universal is accepted one must have specific reasons not to allow more than one universal. Second, resemblance must be resemblance of something, and these somethings cannot be undifferentiated particulars since then the resemblance is inexplicable. Resemblance is resemblance of something in some respect - that is, surely, in some universal aspect. So much for ordinary nominalism. Since transcendent realism seems not to be a live option today, I shall pass it by in silence; it is, however, discussed and criticized in Armstrong's books. But before taking the existence of at least some universals in re for granted, yet another position with regard to universals should be 12

ONTOLOGY

mentioned; a position which can be easily conflated both with the kinds of nominalism earlier mentioned as well as with immanent realism. The position might be called 'the property-instance* or 'the quality-instance point of view'. It is a particularist position, but the particulars are not here regarded as bare or propertyless particulars. The particulars are property-instances. 13 According to this last-mentioned view, two instances of exactly the same colour hue do not have, as I maintain, a universal in common. They are merely two numerically distinct propertyinstances, which means that there is specific particularity without universality. In immanent realism qualitative identity is as fundamental as numerical identity, but in the 'quality-instance point of view' the concept of 'qualitative identity' is a derived notion. Now, the big problem here, as of course Armstrong has noted, 14 is to explain how the particulars are to be classed and sorted. How, in the case of the two colour hue instances can we find ontological reasons for saying that they have exactly the same colour hue? There seems to be something universal lurking in the back of the quality-instance position too. I want to stress that in one sense I am not in opposition to 'quality-instances'. On the contrary, the conception of universale as existing in re imply a conception of quality-instances. But my quality-instances have a universal aspect; they are not mere particularity. This makes it possible for me to speak about both universals and instances of universale. So far, I have only subscribed to a 'minimum conception' of immanent realism — that there are at least some universal properties, namely the most specific properties possible. With some reservations, this is Armstrong's position. I myself, however, shall later on in several respects enlarge the number of different kinds of universals which are to be found in the world (see below, pages 15, 33, 34). Having outlined my view on universals, some words are needed about the relation between universals on the one hand and terms in our language on the other. Instances of universals can be referred to by means of proper names, universals in their universality can be expressed by means of general terms. However, the former fact does not imply that all proper names refer to some definite universal-instance, and the latter fact does not imply that all general terms refer to some definite universal. Those general terms 13

ONTOLOGICAL INVESTIGATIONS

that actually do refer to one universal I shall call 'realist general terms', and those which do not so refer I shall call 'conventional general terms'. Let us now look at our ordinary general property terms and ask ourselves which are realist and which are conventional. Take 'blue'. This term dpes obviously not refer to one universal, but to a disjunction of universals, i.e. all the different blue hues. To say that something is blue, is not to describe that thing in terms of one universal but to say that the thing instantiates one of several possible universals. T w o instances of the same universal are qualitatively identical, but two blue things need not be qualitatively identical. The same argument goes also for terms like 'dark blue', 'medium blue', 'light blue', 'turquoise', etc. It does not take much time to realize that we actually lack realist general terms for the colour hues; at least, if we do not bring in the physicists' talk of colour hues as frequencies of electromagnetic waves. In the latter case each frequency refers to one universal. If universals exist in re, independently of man and language, it is not at all an odd fact that there might be universals without any corresponding realist general terms. Language has to satisfy many pragmatic requirements which, from an ontological point of view, are totally irrelevant. It is not strange that many of even our most specific property terms are conventional and refer to a disjunction of universals. Notwithstanding this, I hope the context will make it clear when I am referring to one specific universal and when I am not.

1.4 D E T E R M I N A T E S AND D E T E R M I N A B L E S Property concepts can be ordered into levels of abstraction. If there were concepts referring to the most determinate colour hues some of these would be subsumed under the concept 'dark blue', some under the concept 'medium blue', etc. The concepts 'dark blue', 'medium blue', 'light blue', and 'turquoise', in turn, are subsumed under the concept 'blue'; and 'blue' under 'colour hue'. Another way of speaking of such abstraction hierarchies is to say that 'colour hue' is a determinable concept for all the subsumed concepts, which then are determinates for 'colour hue'. This means that 'blue' is a determinate with respect to 'colour hue' but a determinable with respect to 'dark blue', 'medium blue', 'light blue', and 'turquoise', and their subsumed concepts. 14

ONTOLOGY

The determinable-determinate distinction is only of fundamental interest in ontology if it applies to universals themselves and not mere terms, 15 All immanent realists agree that the most specific properties possible are universals, but diverge on the question whether there are any determinable universals. Armstrong, for instance, is of the opinion that there are only determinate universals, whereas I maintain that there are determinable universals as well. This view, it must be stressed, does not at all amount to saying that all general terms are realist terms, that is, terms which refer to universals. I regard the term 'colour hue' as a realist term, the terms 'blue', 'green', 'yellow', etc., as conventional terms. In my opinion 'colour hue' does refer to a determinable universal, but the latter terms do not. They refer to a disjunction of determinate universals. Most conventional general terms hover, so to speak, between realist ones. I shall put forward some arguments for my thesis by considering Armstrong's main counterarguments. He has two objections to the view that 'the property redness has the property being coloured': In the first place, it is not the way that we talk. We do not say that redness is coloured. Particulars may be coloured, but redness, we say, is a colour. Ordinary language does not attribute a property to redness, but only membership of a class. And as we have constantly emphasized in this work, membership of a class does not automatically bestow a property upon the member. In the same way, triangularity is not said to be shaped. Particulars are shaped. Triangularity is a shape. It is a member of the class of shapes. 16 I think Armstrong misrepresents the view that there are determinable universals; at least his account does not fit the view I want to defend. Redness (i.e., a determinate hue of redness) does not have the property of being coloured. Both being red and being coloured are properties which substances have (cf. chapter 3). Substances have at one and the same time both determinate and determinable property-instances. On the other hand, a denial that colour is a property of redness does not imply that to be coloured is merely to be a member of a class. The determinate-determinable relation is a quite specific relation which is identical neither with 'being a property of nor with 'being a member o f . I shall return to this relation later (in chapter 9), but to get an idea of what is involved, 15

ONTOLOGICAL INVESTIGATIONS

one can notice that if there are determinable universale then such universale cannot be instantiated without some determinate being instantiated, and likewise the determinates in question cannot possibly be instantiated without the determinable being instantiated. What is Armstrong's second objection to determinable universale? In the second place, not only does redness resemble the other colours in being a colour, it also differs from them in colour. Redness and blueness differ as colours. The property in which they are supposed to resemble each other is the very thing in which they differ. This seems impossible. Things cannot differ in the respect in which they are identical. The conclusion must be that the resemblance of the colours is not a matter of their having a common property. Similarly, not only does triangularity resemble the other sorts of shape, it also differs from them in shape. Triangularity and circularity differ as shapes. So their resemblance cannot be constituted by a common property. 17 The kernel of the argument is contained in the sentences, 'The property in which they are supposed to resemble each other is the very thing in which they differ. This seems impossible. Things cannot differ in the respect in which they are identical.' Armstrong regards the belief in determinable universals as contradictory, and goes on to say that only absolute idealists, who do regard the world as ultimately contradictory, can accept this notion of identity-indifference. But matters are almost the other way round. The existence of determinable universals, first, does not entail a contradictory reality; and, second, provides the best explanation of the way we speak of similarities and dissimilarities in the world. What I have said about determinate and determinable universals is meant to imply that things cannot differ with respect to determinates if they have qualitatively identical determinates, and similarly that they cannot differ with respect to determinables if they have qualitatively identical determinables. 'Things cannot differ in the respect in which they are identical', as Armstrong rightly says. But, of course, if the determinates are not identical, then the things differ with respect to determinates. And this in no way is in conflict with or contradicts the fact that the things simultaneously instantiate the same determinable universal. 'Identity-in-difference', if understood correctly, just means identity 16

ONTOLOGY

of determinable and difference of determinate, and no contradiction is involved. Actually, the criticism presupposes what is to be proved, namely that there are only determinate universals. Only then will it become impossible to explain in what way a red thing and a blue thing are identical. They are just different, and so this difference must do the impossible and explain the presumed identity. Another modern philosopher who subscribes to universals in re but restricts his realism to determinates, is Brand Blanshard. He has an objection which is of a different type from that of Armstrong: Not only is it hard to extract an identical genus from such different species; one cannot even be sure of the species from which one is supposed to extract it. According to the theory, blue is blue, and remains identical through all the changing hues of blue. But where does blue end as it passes through violet and purple towards red on the one side, and through peacock blue towards green on the other? If the concept of blue is definite, there should be no doubt where it is present and where not, but there is doubt on both sides. Indeed, one could quite well start with some intermediate quality like purple and include under one's new genus a set of qualities that had previously been assigned to blue and red. Thus qualitative universals, if they exist at all, will be so blurred as to melt away into each other. 18 The general term which Blanshard discusses, 'blue', should also in my opinion be criticized in the way Blanshard criticizes it. 'Blue' is a conventional term which refers to a disjunction of universals capable of being ordered in an ordinal scale. The latter possibility is a possibility independent of our language. But how the ordinal scale is divided conceptually, that is, how the colour terms are delimited, does not tell us anything about the world. The subsumption of the determinate colours under different terms is conventional. Blanshard's mistake lies somewhere else. It consists in taking for granted that if 'blue' is a conventional general term, then all the determinable terms 'above* it must be conventional terms too. Blanshard makes a false assumption. As a matter of fact his own line of argumentation can be used to show this. According to Blanshard, if 'blue' refers to a universal it should be 'definite, there should be no doubt where it is present and where not'. This is not 17

ONTOLOGICAL INVESTIGATIONS

true of'blue', but it really is true of'colour'. If'colour hue' is taken in the phenomenological sense as perceived colour hue, then it is impossible to find the kind of conventional ingredient which is obvious with regard to colour terms like 'blue', 'green', etc. T h e precise limit between blue and green is a limit within a language, whereas the limit between colour and, for example, extension is a limit in the world. Things are, however, just a bit more complicated. For Blanshard could reply in the following way. Colour hue ought to be regarded not as a phenomenological property, but as a property in the nonperceivable part of the world. And then 'colour hue' is obviously not a term with definite limits. Like 'blue', 'colour hue' represents a conventional delimitation of a scale. T h e theory of electromagnetic radiation revealed 'colour hue' to be a conventional term. What looked like a realist term was shown up, by the advance of science, to be a conventional term. Both light and heat radiation as well as radio waves are today regarded as electromagnetic waves of different wavelengths between 0.4 and 0.8 μπι, heat radiation exists between 0.8 μτη and 1 mm, and radio waves have longer wavelengths. The corresponding general terms cut out relatively conventional parts of the scale of wavelengths wherein all electromagnetic radiation may be placed. T h e divisions made are grounded in our experience of the radiation. As a division of the radiation in itself, it is as conventional as our division of the visible spectrum. Such processes of 'conventionalization' occur now and then in the history of science, and they perhaps deserve more attention than they get from philosophers of science, but I shall not dwell on them here. Instead I want to maintain that a process of conventionalization can never be complete in the sense that all determinable terms become conventional terms. The claim that each particular determinable term can in principle turn out to be a conventional term, is a quite different claim than that all determinable terms whatsoever could be conventionalized. The conventionalization of 'colour' presupposes that other determinable terms are realist general terms. It presupposes, first, a non-conventional distinction between primary and secondary qualities. T h e conventional term 'colour' refers to a disjunction of primary determinates, the different wavelengths of light, and so we need a new term which refers to perceived colour hue. This distinction between primary and secondary colour cannot be regarded as a distinction 18

ONTOLOGY

merely in our language, it must refer to a distinction in the world. Second, the traditional theory of electromagnetic radiation presupposes that electromagnetic phenomena are different in kind from (for example) gravitational phenomena. The term 'electromagnetic phenomena' must therefore be regarded as a realist term. If in the future a field theory is established which unites electromagnetism and gravitation, that is to say, if 'electromagnetic radiation', were to be conventionalized then this theory would contain other determinable terms which should be regarded as realist terms. The arguments put forward in favour of the existence of determinable universals supply us with two criteria for the detection of such universals. First, determinable universals have to have absolute limits; there must, metaphorically speaking, be real gaps between different determinables. Between (for example) volume, shape, and colour there are such gaps, but between blue and green there are not. Second, determinables delimit their determinates in such a way that a determinate cannot be determinate to more than one determinable at a specific level. A determinate red shade can only have colour as determinable, not volume or shape. In this sense determinables constitute a kind of lawlikeness. This is an important point, but it will not become fully clear until chapter 9. So far I have only mentioned three determinable universals: volume, shape, and colour. These are the first determinables with regard to their species infimae, the most specific determinates possible of volume-determinates, shape-determinates, and colour-huedeterminates. But they are not last determinables, the summa genera or categories. Volume, shape and colour in turn are determinates of the determinable property. And, as will be argued in chapter 3, substances are determinates of the determinable quality; the latter being one of the categories. Qualities as well as the other categories and subcategories are, it has been claimed, universals in re. There are absolute limits or gaps between the categories, and they appear in a kind of law (the 'existential dependencies' discussed in chapter 9). Such a realist stand with regard to determinables and categories means that I intend my category system to be a Kategorienlehre in the full sense of this word. In the chapters to follow I try to speak about the world, not - or not only - about our most abstract concepts or about idealizations and purely fictitious entities. The latter play a very important role in science, and even more in 19

ONTOLOGICAL INVESTIGATIONS

technology and social engineering, but not within ontology. Ontology is concerned with what has real existence. In the next chapter I shall introduce the category of ontological level and explain one characteristic of the present theory of categories, its irreductive materialism. The other categories, subcategories, and distinctions needed will then be introduced and argued for in what I hope is the right pedagogical order. At the beginning of this chapter I made a distinction between a Kategorienlehre (theory of categories) and a Kategorientafel (list of categories). For some chapters it will look as if I were merely presenting a Tafel, but later on the categories will be connected with each other. In order to facilitate reading what follows, I shall at once list the categories to be dealt with, along with some subcategories. Here they are, in ontological order of dependence in so far as this can be represented linearly: Space-time State of affairs substance Quality property External relation Grounded relation Inertia Spontaneity Tendency presentational

fictional Figure 1.1

20

ONTOLOGY

This list must at once be supplemented by two comments. First, there is also the 'category' of existential dependence (chapter 9). I put 'category' here within scare-quotes because this is a special kind of relation which connects the other categories, and so is not really a category. Like 'exists', 'identity', and 'number', the relation 'is existentially dependent upon' applies to all the categories; it is transcategorial. Second, the crucial distinction between exclusive and inclusive qualities (chapter 4) is not a distinction between different categories either. This distinction conceptualizes different relational properties which qualities may have with regard to space and time. With the help of these categories and 'categories', other less fundamental categories can be analysed and defined, categories like ontological level, action, function, pattern, pure gestalt, efficient causality, machine, organism, subject, and nested intentionality. And once these are available, it becomes possible to understand what our intuitive distinctions between nature, man, and society really amount to.

21

2

IRREDUGTIVE MATERIALISM

T h e term 'level ontology' signifies a theory according to which the world is layered; its objects form levels which stand in certain dependence-relations. T h a t the world is layered in this way means that universale belonging to one level cannot be reduced to, or replaced by, universale belonging to another level. T h e different levels can contain completely different types of laws - and these cannot be reduced to each other either. I n a level ontology one can distinguish between overlying and underlying levels. A higher level is dependent for its existence on those lying under it, but a lower level can exist without the levels above it existing. Biological organisms cannot exist if there are no molecules and atoms, but molecules and atoms can exist without there being biological organisms. Society cannot exist if there are no biological organisms, but biological organisms can exist even if there is no society. Level ontologies exist in both idealistic and materialistic versions. In a materialistic level ontology the lowest level — that which must exist in order for the other levels to have the possibility of existing — is material in the philosophical sense of the word. In what follows I shall speak sometimes of 'irreductive materialism' and sometimes of 'materialistic level ontology': in both cases I mean the same thing, that the higher levels cannot be reduced to the lowest material level. Plato's theory of forms expresses an idealistic level ontology. According to Plato, everything is dependent for its existence on the idea of the Good, without, however, being reduced to that idea. Roughly speaking, Plato can be said to distinguish four levels: ideas, mathematical objects, things, and representations of things. An irreductive materialism forces itself upon us today, not

22

IRREDUCTIVE MATERIALISM

mainly for philosophical reasons, but because of the theory of evolution. At first there were only atoms and molecules, afterwards simpler organisms, then more complex, and finally organisms with consciousness. According to the definition of irreductive materialism, the upper levels cannot come before the lower; the lower can, but need not, come before the upper. The theory of evolution says that the lower levels have in fact come before the upper. A level ontology can in general allow temporal differences between levels, but it does not prescribe such differences. T h e theory of evolution and cosmology implies a tiered structure having approximately the following form (beginning with the lowest level): elementary particles, atomic nuclei, atoms, molecules, bodies, passive organisms, active organisms, pre-psychic organisms, organisms with a sensorial psyche, organisms with a perceptual psyche, and organisms with an intellectual psyche - human consciousness. 1

2.1 T H E C A T E G O R Y O F O N T O L O G I C A L L E V E L Irreductive materialism easily gives rise to associations with the idea of geological strata or bricks laid on top of one another. These associations are, however, misleading. The central line of thought behind a level ontology is best understood when one looks at language, or more precisely, when one looks at how a text phenomenologically appears to us. What I am going to say is not meant to imply a mentalist position. In linguistics it is necessary to distinguish between two levels of a word or sign. It can be said, for instance, that on one level the sign 'bachelor' has eight letters and on another that it is synonymous with the concept 'unmarried man'. Louis Hjelmslev called the former level the sign's expression-part, and the latter its content-part. Ferdinand de Saussure called the former the signifier and the latter the signified. Similarly to many others, I shall call the former the linguistic sign and the latter the concept. Linguistic signs can be visual (writing), auditory (speech), or tactile (e.g., braille). Concepts cannot, I maintain, exist without some sort of linguistic sign which sustains them. O n the other hand one can very well conceive that visual linguistic signs, for example, exist without any concepts. It is in just this way that one sees a text in a foreign language not completely unrelated to one's own. The concept level is dependent for its existence on the linguistic sign level; but on the concept level there are characteristic and specific 23

ONTOLOGICAL INVESTIGATIONS

law-governed relations which are described with the help of words which do not refer to the sign level. The levels as conceived by irreductive materialism do not exist beside or above one another in a spatial sense. Irreductive materialism stands closer to a monistic than to a pluralistic ontology, though neither characterization exactly fits it. The levels in question do not constitute distinct substances in the way in which corporeal and spiritual substances are distinct in Descartes' ontology. The levels in a level ontology must contain emergent qualities, but not all emergent qualities need necessarily represent new levels (see chapter 9). Neither should the levels in irreductive materialism be identified with a spatial part-whole relation (see chapter 9 and section 15.1). If one considers elementary particles, atoms, and molecules as constituting three different levels, this confusion can easily arise. Atoms are indeed parts of molecules and elementary particles parts of atoms. But the linguistic example above shows that the entities on the underlying level cannot always be seen as parts of the overlying. Concepts and linguistic signs have, phenomenologically, approximately the same spatial extension. The same relation reveals itself when we look at the societal level. This level cannot exist without there existing a level of nature which sustains it. But this does not mean that the nature-level is in any sense a spatial part of the societal, or, for that matter, that the societal is a purely spatial part of the nature-level. In a level ontology the references of many notions are in a certain sense ambiguous. The notions can, at one and the same time, refer to two different levels. If one is speaking of a certain particular teacher, one is not only speaking of the societal object, but also of a certain biological organism. The question of what it is that has the property of being a teacher has two different answers. One answer is that it is the biological organism which has the property. The other is the tautological one that it is the teacher that has the property. It is actually the second answer which is correct, for the biological organism cannot as a biological organism have the quality of being a teacher. It can only, together with other biological organisms - and of course other natural objects constitute a basis for the societal level where teachers might exist. The peculiarity indicated here regarding the reference of certain concepts appears at various points in the history of philosophy. 24

IRREDUCTIVE MATERIALISM

One place is in Hegel's Encyclopedia of Logic, where Hegel writes the following section (216, supplement): The single members of the body are what they are only by and in relation to their unity. A hand e.g. when hewn off from the body is, as Aristotle has observed, a hand in name only, not in fact. 2 A hand removed from the body, and therefore also from the organism-level, is not actually a hand. The hand loses its identity as a hand when it is cut free from the body, at the same time as, in another sense, it of course retains its identity. The same point can be made not only with regard to all parts of organisms and instruments, but also with regard to societal phenomena. Marx has, for example, in his Grundrisse exchanged the hand for a railway: Ά railway on which no trains run, hence, which is not used up, not consumed, is a railway only potentially, and not in reality.' 3 A railway which falls outside of the societal level is no real railway. Both Marx's and Hegel's examples presuppose the existence of identity on two different levels, that is, they presuppose a level ontology. And even if the entities on the higher level must be united with something on the lower level in order to exist, the levels are nevertheless so separated that one can formulate theories about the higher level without mentioning the lower; the latter are merely unexpressed presuppositions. The dual character of reference indicated here is the same dual character as lies behind the distinction functionalists usually make between a thing's structure and its function. A railway has a structure in and of itself (the shape of the rails and ties, etc.), as well as a potential or actual function (that of bearing trains). This way of seeing the situation, however, usually misses the very point being made here, namely that the structure belongs to one ontological level, and the function to another. Actually, it is a mistake to say that it is the same thing which has both the structure and function in question. I shall later return to the problems involved here (see chapter 14). Irreductive materialism is an ontological position according to which the world contains an existential hierarchy. It must not be conflated with theories about control hierarchies which abound in biology and the social sciences. That an upper ontological level is 25

ONTOLOGICAL INVESTIGATIONS

existentially dependent on a lower one implies in itself nothing at all about whether or not the former controls and monitors the latter. 2.2 A R I S T O T L E T o put a little more meat on the concept of a 'level ontology', as well as to create a somewhat wider frame of reference for what follows, I shall present some brief historical examples of level ontologies and views related to them. According to Aristode every truly existing thing has two aspects: matter and form. A certain combination of form and matter can constitute matter for a higher-level form, and that combination can be matter for an even higher-level form, and so on. This line of thought is exemplified by Aristotle's theory of souls or activity principles. All inorganic objects are, according to Aristotle, a combination of the four elements earth, water, air, and fire. Plants of course also consist of these four elements, but also of something more. These four elements in certain combinations are only the matter of plants, but plants must also have a form. And the form is constituted by an activity principle ('plant-soul') which represents the intake of nourishment and growth. According to Aristotle one cannot understand animals if one does not assume yet another and higher activity principle, an 'animal-soul' which represents the ability to move and perceive. This principle (which thus constitutes the zoological level) is a special form which, in order to exist, must be united with matter. And this matter is itself nothing other than the matter and form which constitutes plants. But that which is specific to man, reason, exists on a level lying one step higher. The form or activity principle that represents the ability to think can, however, only exist together with the lower levels which constitute animals. Aristotle's level ontology leads, however, to a quite special problem. As is clear from the above, 'form' and 'matter' are relative concepts: something is matter in relation to a certain form, and vice versa. Thus, if every concrete existing object is a combination of form and matter, the question arises as to whether there can then exist a lowest level. I f so, the lowest level too must be a combination of form and matter. At the same time this matter cannot exist in itself, since there does not exist a still lower level. 26

IRREDUCTIVE MATERIALISM

Aristotle solved the problem by distinguishing between potential and actual existence. The lowest level (called prima materia by the scholastics) has only potential existence. It can never have real or actual existence, but it must nevertheless be postulated for logical reasons; it exists as potentiality. The four elements, too, come from prima materia, through the forms represented by the contraries cold/warm and dry/wet. These contraries constitute the form of the elements: earth is cold and dry; water is cold and wet; air is warm and wet; and fire is warm and dry. Modern physics is not, as one has perhaps been led to believe, completely free from an Aristotelian type of form-matter metaphysics. Nowadays the notion of energy plays the role of prima materia. Energy can never exist in and of itself; it must always assume some form. It can be manifest as momentum, potential energy, electromagnetic energy, or thermal energy, as well as in the form of mass. That even mass is a form of energy, a consequence of the theory of relativity, implies that, ontologically speaking, everything has energy as a substratum. Pure energy exists, as does Aristotle's lowest substratum, only potentially. The problem regarding the existence of the lowest level is of minimal interest to most of science, since the separate disciplines normally study levels which are a good way up the existence hierarchy. But it is nevertheless of interest to have this problem pointed out. I should also like to take the opportunity here to clarify a distinction which Aristotle sometimes employs, and which I shall make use of here. He says that the underlying level (the matter) is a substratum for that lying above it (the form). A combination of form and matter can be a substance, that is, something which exists in and through itself, but it can at the same time be a substratum in relation to an overlying form. It is only prima materia which is necessarily only substratum. One must carefully distinguish between substance and substratum, a point to which I shall return in the next chapter. 2.3 POST-SCHOLASTIC L E V E L O N T O L O G I E S Level ontologies have been rare amongst the post-scholastic philosophers. And when they have appeared, they have been most often in the form of idealistic theories about different levels of consciousness (Leibniz and Hegel). The only major philosophical 27

ONTOLOGICAL INVESTIGATIONS m o v e m e n t to a r g u e for a n irreductive materialism in later times is dialectical materialism, but even there this aspect of the theory h a s not been particularly emphasized. T h a t it is dialectical materialism in particular which advocates a n irreductive materialism has its roots partly in Friedrich Engels' w a a t i n g to s t a n d Hegel on his h e a d , and transform his idealistic philosophy into a materialistic one. Hegel's levels of consciousness came in that way to be transformed into levels of existence with a materialistic basis. T h a t I mention Engels here is due to his clearly formulating a line of thought which lies latent in every materialistic level ontology. Engels not only holds that the existence of the lower levels is a condition for the existence of the higher, b u t also that the higher levels only a p p e a r when particular conditions are fulfilled on the lower. T h u s there is conformity to law not only within b u t also between levels. Engels has called this the transformation of quantitative changes into qualitative changes. O n e of m a n y weaknesses in Engels' sketchy discussions is that, like m a n y others, he does not clearly distinguish between the cases where emergent qualities constitute a new, higher level, a n d those cases where they are only emergent within a given level. 4 Relatively few philosophers in this century have explicitly advocated an irreductive materialism. A m o n g those who have done so, I think one should mention first a n d foremost Samuel Alexander, Nicolai H a r t m a n , R o m a n I n g a r d e n , Michael Polanyi, a n d Mario Bunge. 5 All are generally regarded as eminent philosophers, b u t also, all are lone figures within their respective philosophical cultures. 6 I t is quite clear to t h e m all t h a t science has for a long time had irreductive materialism as an implicit ontological hypothesis. Bunge expresses this in the following way: A n ontological hypothesis involved in and encouraged by m o d e r n science is that reality, such as is known to us today, is n o t a solid homogeneous block b u t is divided into several levels, or sectors, each characterized by a set of properties a n d laws of its own. T h e m a i n levels recognized at present seem to be the physical, the biological, the psychological, a n d the sociocultural ones. Every one of these may in t u r n be divided into sublevels. F o r instance, the m a i n sublevels of the physical level are the physical p r o p e r and the chemical sublevels, a n d the main sublevels of the sociocultural level are the economic, the social

28

IRREDUCTIVE MATERIALISM

proper, and the cultural sublevels. Finer divisions can be introduced and none are clear-cut and rigid. A second, related presupposition is that the higher levels are rooted in the lower ones, both historically and contemporaneously: that is, the higher levels are not autonomous but depend for their existence on the subsistence of the lower levels, and they have emerged in the course of time from the lower in a number of evolutionary processes. This rooting of the higher in the lower is the objective basis of the possibility of partially explaining the higher in terms of the lower or conversely. The two basic ontological hypotheses just referred to are built into the contemporary outlook to the point that they underly the usual classification of the sciences and they dominate more and more our system of higher education. Thus, the scientific psychologist is required to learn more and more biology, and even chemistry and physics, because it is increasingly being recognized that psychical events are rooted in these lower levels; but the psychologist is also required to become conversant with sociology, because we have come to realize the reaction of the sociocultural level on the immediately lower ones: thus, we recognize the influence of religion on food habits and the reaction of the latter on food production. Only physicists are entitled to ignore the upper levels.7 Bunge holds, as I do, that such an ontology has consequences for both epistemology and methodology, and he has tried to bind together ontological hypotheses with epistemological as well as methodological rules. 8 I shall, however, confine myself to ontology.

29

3

STATES OF AFFAIRS AND QUALITIES

The distinction between things and properties of things is an obvious distinction within common sense or everyday ontology. The theory of categories here presented will, I hope, stay close to common sense, but this twofold division of things and properties is still too crude for our present purposes. In spite of the fact that Aristotle had discovered some of the real complexities involved in the division, many philosophers, especially today, think they can erect a whole ontology on its basis or on the basis of a similar distinction — such as that between particulars and universals or elements and sets. Aristotle realized that things are not just particulars and properties are not just universals. With regard to properties, that is, Aristotelian accidents, one has to distinguish between the universal property and the individual property which exists in a specific particular. Similarly, Aristotle maintains, a man or a thing like a piece of gold are not only particulars. They involve a Table 3.1

Said of a subject (universal, general)

Not said of a subject (particular, individual)

Not in a subject (substantial)

In a subject (accidental)

(Second substances)

(Non-substantial universals)

man

whiteness, knowledge

(First substances)

(Individual accidents)

this individual man, horse, mind, body

this individual whiteness, knowledge of grammar

30

STATES OF AFFAIRS AND QUALITIES

universal aspect as well; there are also the universale man and gold. For this reason Aristotle distinguished between 'first substances' and 'second substances'. The individual man and the individual piece of gold are first substances whereas man and gold are second substances. Aristotle does not merely have a twofold division between substance and accident, he employs a fourfold division of the sort shown in Table З.1. 1 The clearer the ideas, the clearer the problems they bring with them. This goes for Aristode, too. In Table 3.1 a defect is immediately apparent: the fourfold division, which is ontological, is justified by a grammatical distinction, that between what is said and what is not said of a subject. If one tries to remedy this defect, one discovers a problem in the Aristotelian view. The ontological distinction corresponding to the grammatical one is obviously the one between particulars and universals as expounded in chapter 1, that is to say universals are universals in re and do not stand over and above particulars. The fourfold division then rests on two ontological distinctions, universal-particular and substanceaccident. This means, and now the problem alluded to becomes apparent, that first substances are merely the union of particularity and substantiality and neither of these bring with them what was for Aristotle the defining characteristic of first substances, namely independent existence. We need a third ontological distinction, one between dependent and independent existence. This distinction paves the way for the notion of 'fundamental state of affairs'. If I put the view I am going to argue for in a table like the preceding one it looks like this:

Table 3.2 Independent Universals

Particulars

Dependent

Fundamental states of affairs

Instances of fundamental states of affairs

31

Qualities

..substances £ ^ ^ properties

Quality- ι instances* 4 ^

substanceinstances propertyinstances

ONTOLOGICAL INVESTIGATIONS

The distinction between dependence and independence will be explained and defined in chapter 9. For the moment it has to be understood with the help of the ordinary connotations of the words. Why the terms 'state of affairs' and 'quality' have been chosen will be explained in the subsequent paragraphs. My 'instance of fundamental state of affairs' corresponds to Aristotle's first substance; both represent in a way independent existence. Similarly, 'substance' corresponds to second substance, 'property-instance' to individual accident, and 'property' to non-substantial universal. The other concepts in my table ('state of affairs' and 'substanceinstance') do not have counterparts in Aristotle.

3.1 SUBSTANCE AND STATE O F AFFAIRS To every level in a level ontology there correspond particular subject concepts and (monadic) predicate concepts. The level of society can be described with subject concepts like 'firm', 'household', 'commodity', 'money', and 'salaried worker', and predicate concepts like 'preference system', 'income', and 'price'. The physical level is described with subject concepts like 'hydrogen', 'helium', 'iron', and 'gold' and predicate concepts like 'shape', 'colour', and 'mass'. The logical distinction between subject and predicate reflects at first glance an ontology in which properties are always properties of something, i.e. it reflects some kind of substance-property ontology. The theory of categories I shall present includes such an ontology, and does so in the special sense that every level has its own special substances and properties. This last-mentioned view has certain consequences which are not always recognized, the most important being the necessity of drawing a sharp distinction between substratum and substance. A property (e.g. being of a price) cannot exist without something which sustains it (e.g., a commodity). A price is always the price of a commodity. Properties require for their existence property bearers or substances. But the opposite also holds: substances cannot exist without properties. Substances cannot have independent existence. A commodity is not a commodity if it does not have a price. A piece of iron cannot exist if it does not have such properties as colour and shape; and colour and shape must themselves be sustained by something in order to exist. 32

STATES OF AFFAIRS AND QUALITIES T h e thesis of substance and properties has had its critics, the most important being the empiricists. T h e latter claimed that substances, as distinct from properties, are not observable, that as a consequence they do not exist, and that this implies that substance concepts can in principle be reduced to predicate concepts. Empiricists have never managed, however, as they themselves admit, 2 to turn this 'in principle' into 'in fact'. A n d this, I would argue, is not merely a minor problem, for the empiricist's claim is based on a mistaken analysis of our perceptions. W h e n we perceive properties such as shape or colour w e perceive them as a matter of fact as properties of something. Normally we do not perceive merely such and such properties; we also perceive something, which cannot be reduced to the properties which it has. Modern perceptual psychology 3 and phenomenological philosophy have shown that the substance concept cannot be dispensed with when it comes to characterizing our perceptions. T h e empiricist critique of the substance concept is untenable. A n d other ways of defending the view that substances are nothing other than conjunctions of properties often have a hidden empiricist premise. T h i s is the case, for example, with the view that, for reasons of simplicity, one should consider substances as complexes of properties. T h e correct view concerning how our perceptions appear, implies that it is in fact simpler to speak about substances than about complexes of properties. It is simpler to postulate what we in fact directly perceive than something we do not perceive. In Armstrong's terms I defend 'irreducibly substantival universale', 4 which he rejects. (This is the second respect in which I deviate from Armstrong's views on universals. See above, page 13f. I do not go into this discussion about irreducibility in detail since my view on this topic is of minor importance for the main theses of the present work.) Substances and properties presuppose one another; neither can exist without the other. A substance-instance cannot exist without some property-instance, and a property-instance cannot exist if it is not a property-instance o f a substance-instance. Substances and properties which presuppose one another make up a unity of which the substance and properties in question are aspects. Such a unity is also a universal, and a universal of the kind I shall call state of affairs. W h e n the corresponding substance-instance and propertyinstances exist there is an instance o f a state of affairs. Since substances and properties presuppose one another, it 33

ONTOLOGICAL INVESTIGATIONS

follows that neither of them can have independent existence. An instance of a fundamental state of afTairs on the other hand can have an independent existence. Not in the sense of absolute independence, because it depends on its parts, the substanceinstance and the property-instances, but as a unity it has a relative independence. With these parts or aspects 5 it is something which can exist in and of itself in space and time. The fact that the unity of substance-instances and property-instances, in contradistinction to these instances themselves, can have independent existence, is a fact which shows that states of afTairs cannot be reduced to substances and properties. Like substances, states of affairs are special kinds of irreducible universals. (This is the third respect in which I deviate from Armstrong.) O n the view I am setting forth substances and properties are both different and irreducible to each other, but in spite of this they have the same kind of relation to states of affairs. They are in a sense parts (abstract parts or moments, for those who know Husserl; see chapter 9) of instances of states of affairs. And as they are parts of what has independent existence but have themselves not such an existence they may well be called qualities. The most peculiar (but also the most important) point to notice with regard to the distinctions put forward is that what can have independent existence must be structured. States of affairs cannot be simple; they must be complex. They necessarily have qualities, i.e. substances and properties. In my opinion ordinary things necessarily have both substance(s) and properties, which means that I claim that ordinary things are states of affairs. A thing like a gold ring is a state of affairs which contains both a substanceinstance (being of gold) and some property-instances (e.g. being ring-shaped) . 6 If we disregard holistic philosophers like Spinoza and Hegel, it seems that many philosophers have a strong temptation to connect what is simple with what has independent existence, and connect what is complex with what has dependent existence. States of affairs, as I have described them, contradict these intuitions. The concept of (fundamental) 'state of affairs' now introduced is very restricted compared with ordinary usage, which takes as paradigmatic examples two things plus a relation. Facts like 'a is smaller than b', 'a is placed upon b', and so on, are not states of affairs in the fundamental sense here intended. (Although, as we 34

STATES OF AFFAIRS AND QUALITIES

shall see in chapters 8 and 9, such facts are indeed a kind of complex states of affairs.) 7 In anticipation of chapter 13 ('Intentionality'), I should at once say that one main line of division within the set of all states of aiTairs is that separating states of affairs which contain intentionality from those which do not. The former are or involve subjects. Among the latter we find ordinary things and for many chapters to come I shall confine my discussion to these alone. 3.2 SUBSTANCE AND PROPERTY A universal can subsume other universals, and a universal term ('colour') can subsume other universal terms ('blue', 'green', etc.). Both relations reflect the determinable-determinate relationship mentioned in chapter 1. With regard to the division of qualities into substances and properties, it should be noted that the subsumption relation is sometimes held to have a somewhat different character in the two cases. When a substance is subsumed under another, for example human under animal, or iron under element, then there obtains the relation species-genus. In such cases one can normally characterize the species by giving the genus and certain distinguishing properties, precisely as is done in the classical Aristotelian definition of human as rational animal. If, on the other hand, a property is subsumed under another, for example yellow under colour hue, then it is impossible to find anything like an Aristotelian definition of the subsumed universal. Yellow is a particular colour; there is in a certain sense nothing more to add. The difference just described was once used by W.E. Johnson to distinguish what he called the determinable-determinate relationship from the species-genus relationship. I shall however use the terms 'determinate' and 'determinable' as, so to speak, determinable terms. They cover both Johnson's original determinatedeterminable relationship and the species-genus relationship as well as similar relationships in other categories. Determinate-terms can be logical subjects. There is neither a logical nor a grammatical mistake in either of the sentences 'Yellow is a colour' or 'Iron is an element'. Of course 'yellow' and 'iron' can also be logical predicates, as in the sentences 'The statue is yellow' and 'The statue is (of) iron'. 35

ONTOLOGICAL INVESTIGATIONS T h u s it is not only terms which refer to things that can be logical subjects, b u t also terms which refer to substances or properties. A thing ( = a state of affairs) like a piece of gold has both a substance a n d several properties such as volume, mass, colour, a n d s h a p e . T h e difference between substances and properties is not to be found in their relations to the things which have them as qualities. Substance and property j u s t are different categories w h i c h at the categorial level are two-sidedly dependent. But between the subsumed determinates, for instance between the s u b s t a n c e gold and the properties of gold, there m a y be n-sided relations of dependence between substance and n - 1 properties. T h e properties d e p e n d both on each other and on the substance for their existence, as well as vice versa. I think it is the fact that a substance can be dependent on several properties while all these properties are dependent on the same substance, which explains w h y substances are and should be regarded as property bearers whereas properties should not be regarded as substance bearers. I n this way the relation of inherence is cut loose from the requirement t h a t one of its relata (substance) should have independent existence. M y bringing together of substances and properties into the category of quality is a m o n g other things m e a n t to imply that w h a t I h a v e m a i n t a i n e d a b o u t realism with regard to properties is also true for substances. M y belief in 'irreducibly substantival universals' is a belief in natural kinds. As I have said earlier, I shall not argue for this irreducibility. But it is w o r t h pointing out that there are different ways of introducing n a t u r a l kinds. D u r i n g the 1970s n a t u r a l kinds enjoyed a renaissance within analytical philosophy. 8 B u t as usual within this tradition the a r g u m e n t s were primarily of a linguistic n a t u r e . It was argued that language necessarily is such t h a t some users of referring expressions presuppose the existence of n a t u r a l kinds. T h i s turns things upside down. I think that both o u r o r d i n a r y everyday language, as well as ordinary scientific language, silently presuppose an ontology embracing n a t u r a l kinds. T h i s ontology has now been rediscovered in the disguise of a necessary t r u t h a b o u t language by philosophers such as Kripke.

3.3 S U B S T A N C E A N D S U B S T R A T U M A c o m m o d i t y together with its price constitutes a state of affairs; 36

STATES OF AFFAIRS AND QUALITIES the commodity is the substance of this state of affairs, and the price is its property. But this societal state of aiTairs presupposes the natural level; this natural level, in contrast, does not presuppose the societal. T h e commodity must assume some form in a material object. T h i s underlying level constitutes the higher level's substratum. T h e r e are, I claim, two sorts of property-bearer, substances and substrata. T h u s the property price is always the property of both a commodity (the substance property-bearer) and a material object (the substratum property-bearer). I f the commodity is an iron b e a m , the substratum is this piece o f iron; if the commodity is one kilogram of potatoes, the substratum is the potatoes as such. T h e substratum as distinct from the substance, is a l w a y s a state of affairs (i.e not merely an aspect of a state o f affairs). O f course the substratum can in turn be divided u p into aspects of substance and property, but then the latter substance or substances belong to the underlying natural level. 'Substratum' is a relative term. T o be a substratum is not to be a substratum in itself, but to be a substratum in relation to a given ontological level. T h e iron beams and potatoes referred to may well exist without being substrata o f commodities or a n y t h i n g else. T h e distinctions between substratum, substance, property, and state o f affairs are distinctions w h i c h refer to the world and not to language. T h e distinction between proper names and universal terms, in contrast, as well as that between logical subject and logical predicate, have reference to language. A few words should be said about the relations a m o n g all these distinctions. A universal term is a term whose reference can be found in m a n y different places simultaneously; a proper name is a term whose reference can be found only in one place at any one time. T h e term 'property', as well as the terms it subsumes ('colour', 'shape', etc.) are universal terms. T h e same holds, however, for the term 'substance' and the terms it subsumes ('commodity', 'iron', etc.), and even the term 'state of affairs' and its subsumed terms; all these are terms whose reference is multiple or plural. T h e concepts introduced so far c a n be graphically illustrated by using them to restate the position o f irreductive materialism. A double-headed arrow indicates that the universale related are m u t u a l existence conditions for one another. A single-headed arrow indicates that the existence of that toward w h i c h the arrow points is necessary to the existence of that from w h i c h the arrow points.

37

ONTO LOGICAL INVESTIGATIONS

Figure 3.1, by way of illustration, is an ontology with four levels. The higher levels are emergent in relation to the lower ones. The things on the highest level are of course not substrata for any other level. The overlying levels cannot be used to explain the existence of the underlying; thus there is no question of a 'pure form' in Aristotle's sense. On the other hand, this ontology, when it comes to the lowest substratum, has a problem in common with that of Aristotle, namely: does such a lowest substratum exist? If all levels in the ontology must have a substratum, then the lowest level, i.e. that on which states of affairs) exist, must also have a substratum (= substratumo). But if everything which really exists is a combination of substances and properties, then it is impossible that substratumo exist. One is faced with a dilemma: either the lowest level lacks a substratum, or there exists some state-of-affairs-species whose instances do not consist of substance and properties. This dilemma was mentioned in the previous chapter, as was the fact that Aristotle believed himself to have solved it by distinguishing between actual and potential existence. T o substratumo {prima materia) he allows a necessary though merely potential existence. The same holds, as was also pointed out, for pure energy as it is understood in contemporary physics. Before Einstein's formulation of the principle identifying mass as form of energy, it was possible to consider mass, shape, and volume as properties of bodies. With the introduction of the equivalence principle, properties such as shape and volume have to be regarded as properties of the energy form mass. The ontological status of mass has been changed from that of a property to that of a substance. Energy without form has

(Substance (Substance

properties) = state of affairs 4

Ψ

properties) = state of affairs3 = substratum3

4J> (Substance

properties) = state of aflairs2 = substratum2

(Substance

properties) = state of afiairsi = substratumi

Substratumo Figure 3.1

38

STATES OF AFFAIRS AND QUALITIES

become substratum 0 , and exists only potentially. However, I have difficulty in accepting that the Aristotelian view is unavoidable, even if I accept it as one possible option. It is quite possible to imagine the lowest level as consisting of substances with properties, but without any underlying substratum distinct from the substance; there is then no prima materia. This is how every atomic theory, from antiquity to the present day, has seen the matter. Whether one has to postulate an underlying substratum for the lowest level seems to have to do with whether the states of affairs on this level are assumed to be transformable into one another. In an atomic theory such transformations are excluded, since the atoms are immutable by definition. The different energy forms in physics can, on the other hand, be transformed into one another, and these transformations are most easily understood with the help of an underlying substratum. The alternative is the hypothesis that with a 'transformation' a certain quantity of energy of a particular sort disappears, at the same time as an equally large quantity of energy of another sort appears from nowhere. I have earlier pointed out that my category 'substance' corresponds to Aristotle's category 'second substance', and that his category 'first substance' corresponds to my concept 'state of affairs'. Aristotle's discussion of these notions is, however, far from unambiguous. C.J. O'Connor, in his article 'Substance and attribute' in the Encyclopedia of Philosophy writes: The various notions of substance as (1) the concrete individual, (2) a core of essential properties, (3) what is capable of independent existence, (4) a center of change, (5) a substratum, and (6) a logical subject are never thoroughly worked out and reconciled in Aristotle. He appears to emphasize now one and now another mark of substance as of paramount importance. 9 In order to further clarify my point of view, I shall relate it to O'Connor's six notions of substance. I shall take them in order. (1) ('substance' means 'the concrete individual'). Instantiated states of affairs on all levels exist as concrete individuals. Every state of affairs has however two aspects, substance and properties. The concept 'state of affairs' is consequently not an unanalysed concept. (2) ('a core of essential properties'). I have yet to distinguish essential from accidental properties. I see both substances and 39

ONTOLOGICAL INVESTIGATIONS s u b s t r a t a as p r o p e r t y bearers, a n d so as bearers of b o t h essential a n d accidental properties. Since a substance is an aspect of a state of afTairs, one can also call a state of affairs a property bearer. I n c h a p t e r 8 I shall briefly discuss the difference between essential a n d accidental properties, but in anticipation it m a y be said that properties can be essential only in relation to a particular substance. C e r t a i n sorts of substance cannot exist if they d o not sustain certain particular properties. Electrons, for example, cannot exist without having the property of being electrically charged, a n d commodities c a n p o t exist without having the property price. (3) ('what is capable of independent existence'). According to the presentation given thus far, only instances of states of afTairs on the lowest level c a n claim to have an i n d e p e n d e n t existence, i.e. h a v e an existence in and through themselves. Later, however, I shall argue that they are dependent on space for their existence, which is why I a m forced to maintain t h a t no state of affairs on a n y level exists in an absolute sense in a n d through itself. (4) ('a center of change'). A state of affairs on any level can be described as ' a center of change' given that the substance-instance of the state of afTairs can be assumed to e n d u r e through time. T h e concept 'state of affairs' (or 'substance') itself is completely compatible with, for example, a phenomenalistic ontology where the states of affairs are m o m e n t a r y perceptions. A s u b s t r a t u m , on the other h a n d , o u g h t not be described as 'a center of change' for the substances a n d properties it sustains. T h i s is because in m a n y cases the s u b s t r a t u m on the underlying level can be replaced without the state of affairs on the overlying level changing its identity. A firm can retain its identity even if all the owners, employees, a n d machinery are replaced. T h e s a m e applies to the b o d y ' s relation to its cells. A m a n ' s body does not ordinarily lose its identity w h e n the cells are successively replaced. (5) ('a s u b s t r a t u m ' ) . M y distinction between s u b s t r a t u m a n d s u b s t a n c e is u n u s u a l . Most often substance and s u b s t r a t u m are identified. Given the distinction, however, the situation becomes simple. A state of affairs can be a s u b s t r a t u m for a n overlying level, a n d nothing is s u b s t r a t u m in itself. (6) ('a logical subject'). As I noted earlier, logical subjecsts can refer not only to things b u t also to substances (as species) a n d properties (as determinates). Since the view I have p u t forward implies that things are states of affairs, it also implies that m a n y 40

STATES OF AFFAIRS AND QUALITIES

states of affairs - and especially fundamental states of affairs - can be referred to by means of a logical subject. In this respect my notion of fundamental state of affairs is similar to Aristotle's notion of first substance. But, of course, my position deviates very much from the popular view that only sentences can describe or refer to states of affairs.

41

4 EXCLUSIVE AND INCLUSIVE QUALITIES

An ordinary common-sense distinction is that between things which are blends or mixtures and things which are not, i.e. a distinction between heterogeneous and homogeneous things. T h e ancient philosophers worried a lot about this distinction, thus, for example, Anaxagoras distinguished between anomoeomerous and homoeomerous stuffs. Within contemporary analytic philosophy a similar problem crops up as a problem about the opposition between two kinds of terms, 'mass nouns' and 'count nouns', between, for instance, 'water' and 'apple'. 1 I shall, of course, discuss universals themselves, not terms, and I shall deal with a distinction which is only closely related to the aforementioned ones, a distinction within the category of quality. Qualities (i.e. both substances and properties) are, I shall say, either inclusive or exclusive. T h e distinction between the heterogeneous and the homogeneous has mainly been discussed with regard to stuffs or substances, but Armstrong has tried to revive the old Anaxagorian distinction with regard to properties. (As we saw in chapter 3, he does not even believe in the existence of substantival universals.) For Armstrong, a property is homoeomerous if and only if for all particulars, x, which have that property, then for all parts у of x, у also has that property. If a property is not homoeomerous, then it is anomoeomerous. 2 As I said above, the related distinction I want to introduce is applicable to both properties (see 4.1) and substances (see 4.4). M y concept of 'exclusive' corresponds to Armstrong's 'homoeomerous' (and, in turn, to the older 'homogeneous'), and my concept of 'inclusive' corresponds to his 'anomoeomerous' (or 'heterogeneous'). 42

EXCLUSIVE AND INCLUSIVE QUALITIES

The connection between the distinctions is that if something is homogeneous in the sense that it is made up of one kind of stuii only, then all other kinds of stuifs are excluded. 'Homogeneity' and 'exclusion' actually go together, and, in a similar way, 'heterogeneity' goes together with 'inclusion'. If something is heterogeneous it includes several stuffs. In the previous chapter I claimed that instances of the categories of substance and property are mutually dependent on each other, and that the category of state of affairs is a unity constituted by these two categories. This means that what I am going to say about exclusion and inclusion with regard to properties and substances will be applicable, in a derived sense, to things (states of affairs) as well. 4.1 EXCLUSIVE AND INCLUSIVE PROPERTIES Let us consider simple properties such as volume, shape, colour hue, and mass, i.e. properties which can inhere in a spatially limited thing. What is intended by 'volume' and 'shape' ought not to require further clarification, but the way I am going to use 'colour' and 'mass' needs some comment. When I speak of colour, I do not mean colour as an electromagnetic wave-motion, but colour as it is ordinarily understood - colour in its phenomenological sense. Colour in that sense inheres in things in just the way common sense supposes. By 'mass', I mean mass as understood in Newtonian mechanics, i.e. as a property which directly inheres in things. For those who are unaccustomed to the distinction made in Newtonian mechanics between mass (which exists in a thing) and weight (which is force on a thing), 'mass' may be replaced by the everyday notion 'weight'. Those who are perhaps unable to free themselves from the notions of relativity theory may replace 'mass' with 'rest mass'. Things exist in space and time and have the properties of extension in space and extension in time. Let us begin by comparing these properties with a property such as colour hue. Assume that we have a uniformly coloured thing. Take a line in space along which it has a certain extension. Assume further that we shorten the thing a certain amount along the given line, i.e. in thought we remove a certain piece, we do not compress it. That which remains is a part of the original thing. The 'shortened' thing 43

ONTOLOGICAL INVESTIGATIONS

instantiates the same specific colour hue as the original, but it necessarily instantiates a different, i.e. smaller, specific volume. If we remove another piece, the remaining thing still instantiates the same colour hue, but an even smaller volume, and this process can clearly be repeated an arbitrary number of times. This thought experiment shows that there exists a difference between determinate colour hues and determinate volumes. A colour-determinate can be instantiated in a thing without other colour-determinates being instantiated in the possible parts of the thing, but a volume-determinate cannot be instantiated if other volume-determinates are not instantiated in the thing's possible parts. There is a part—whole relation between volume-determinates which is lacking in the case of colour-determinates. A volumedeterminate includes other determinates (those which are of less volume) as parts, while a colour-determinate is, in a manner of speaking, only itself and excludes other colour-determinates. I shall thus call volume-determinates inclusive properties, when they are instantiated they always include other determinates, and I shall call colour-determinates exclusive properties. M y talk about parts and possible parts may seem a little confusing, but is a consequence of my immanent realism. Properties exist in re, and property-instances are, like thinginstances, actual. Such actual entities may have possible parts defined via the thought operation of shortening. But with regard to a property's universality aspect there can be no distinction between the actual and the potential. Therefore, each universal volumedeterminate just has other universal volume-determinates as parts; and the universal colour-hue-determinates just lack parts. In the former case the universals are inclusive, in the latter they are exclusive. The distinction between the inclusive and the exclusive concerns the universal not the instance - even if it is introduced by means of operations on instances. I n the examples given, the distinction between exclusive and inclusive properties relates to determinates, but it is equally applicable to determinables. In the latter case, however, it turns out that both volume and colour hue are exclusive properties. If one removes a segment from an extended thing, then both the original and the thing remaining instantiate the same determinable volume, in spite of the fact that they instantiate different volumedeterminates. The determinable volume is exclusive, while the 44

EXCLUSIVE AND INCLUSIVE QUALITIES corresponding determinates are all inclusive. With regard to colour hue, both the determinable and all its determinates are exclusive. It is therefore important to distinguish between determinables and determinates. 3 In what follows the examples employed are all determinates because most determinables seem to be exclusive, which means that on the level of determinables the exclusiveinclusive contrast is not apparent. Just as a thing has extension in space, so it has extension in time, its 'life time'. O n e can thus perform the same thought experiment with regard to time as we have just performed with regard to space. T h e last part of the time during which a thing exists can be removed in thought which leaves the thing 'untouched', but with a shorter life. Extension in time is, like extension in space, an inclusive property. If the distinction between exclusive and inclusive properties were such that the only existing inclusive properties were spatial and temporal extension, it would not be as important as it actually is. But there exist other inclusive properties. Let us take a look at the relation between the properties mass and density. T w o things can have different masses but nevertheless be exactly the same in all other respects, i.e. have the same volume, the same shape, the same colour hue, and so on. I f we imagine successively smaller parts of two such things (the corresponding parts however being equally large), then the mass diminishes successively in both cases. But for arbitrarily small parts there nevertheless remains the difference that one of the corresponding parts has a greater mass than the other. T h i s difference is due to the fact that, as we say, the things have different densities. O n e difference between the properties mass and density consists in the fact that mass necessarily diminishes when one takes smaller parts of a thing, while density can remain the same. It is the same difference as occurs between volume and colour hue. Mass is an inclusive property and density an exclusive property. A certain density can be instantiated in an arbitrarily large or small piece. Colour hues and densities must be carefully distinguished from colour patterns and density patterns. I f one removes the last part of a pattern one gets a different pattern, which means that the patterns are inclusive qualities. Densities and masses are not patterns; the former are exclusive and the latter inclusive. T h e r e are density patterns, however, and these are always inclusive. ( T h e concept of 45

ONTOLOGICAL INVESTIGATIONS

pattern will be discussed in chapter 6.) Were we to draw a coordinate axis through a thing which has a certain mass, then we could ascribe to every part of the line a certain density and a certain mass. If we imagine moving from the beginning to the end of the line, then we can determine the mass between each point at which we find ourselves and the beginning point. We can also determine the density pattern between respective points. If we go from one point to another lying further along the line, the new mass is the sum of the first and the intermediate masses. Densities as such, however, are not additive. They just exist side by side, part by part. O n the other hand, density patterns can in a sense be said to be additive. The pattern constituted of the intermediate parts can be added to the old density pattern, and a new density pattern can arise. A thing as a whole has one mass and one density pattern, but may have many densities. If the mass is 5 kg, then it also includes many masses 4.99, 4.98, 4.97, and so on, down to 0 kg. Here there is a part-whole relation between determinates. But if a homogeneous thing has a density of 0.005 kg/m 3 , it does not include the densities 0.00499, 0.00498, 0.00497, and so on. Here there is no part-whole relation between the determinates. And neither is there any such relation when the density varies, as, for example, when the thing has the density 0.005 kg/m 3 at one part, 0.0049 at the next, and so oil. These densities do not include one another. Quantity of an inclusive property means one thing, and quantity of an exclusive property something else (see below, page 52). Both mass and density can be quantified and linearly ordered in a relation of greater and less. But while an instance of a larger mass necessarily has potential spatial parts which have less mass, an instance of a larger density has no spatial parts at all. Each density is, like each colour hue, an indivisible unit which excludes other similar units. Between exclusive properties no part-whole relations hold, only grounded relations (see chapter 8). In certain cases the relations are of the type 'the same as' and 'more like . . . than . . .', such that the determinate in question can be quantified and placed in ä linear order. Such relations also hold between volume-determinates and between mass-determinates, but in the case of these two a part—whole relation exists as well. Furthermore, it is the existence of this part-whole relation which explains why 46

EXCLUSIVE AND INCLUSIVE QUALITIES

mass and extension, as distinct from density and colour, are additive entities. Every sum is the sum of something. A sum is a whole which must have parts. Properties which exclude parts by definition (exclusive properties) can never be added nor conceived as the sum of anything. 4 The distinction between inclusive and exclusive properties is absolute. It is not, as is the distinction made in analytical philosophy between extensive and intensive properties, arbitrarily dependent on the existence of a discovered physical operation fulfilling certain formal additivity conditions. 5 If a certain property is exclusive, it is impossible to find a physical operation which makes it additive. Temperature, as it is understood in everyday situations, and as it is understood in phenomenological thermodynamics, is an exclusive property. We can imagine smaller and smaller parts of a thing as long as we want, and in such a way that all these parts have the same temperature. So long as we have such an understanding of temperature, it is impossible to find an operation which makes it an additive property. Since both mass and density are quantifiable properties, the distinction between inclusive and exclusive properties cannot coincide with any distinction between quality and quantity. Quantifiable properties are a subclass of all properties. Exclusive properties have already been exemplified by cases of both quantifiable and non-quantifiable properties. By 'quantifiable property' I mean a property which can be ordered in either a ratio scale or an interval scale, which means that a 'non-quantifiable property' might be ordered in an ordinal scale. Phenomenological colour hue is a non-quantifiable property; but how do things stand with regard to inclusive properties? Are there non-quantifiable inclusive properties? Can there exist part-whole relations between determinates which are not quantifiable? My reply is in the affirmative, and an example is the property shape (i.e. the shapedeterminate). We obtain Table 4.1: Table 4.1

Exclusive properties Inclusive properties

Non-quantifiable properties

Quantifiable properties

Colour hue Shape

Density Volume, Mass 47

ONTOLOGICAL INVESTIGATIONS

Figure 4.1

Assume that we have a circle-shaped thing and that we cut it away in thought, by orthogonal cuts, bit by bit, along a given line (see Fig. 4.1). The parts which thereby result are well-defined, due to their relation to the line and the cuts. The shape of the parts is never the same as the shape of the original thing. The shape meets the criteria for being an inclusive property. The circle shape or indeed any particular group of shapes are not the only inclusive non-quantifiable properties, however. If one simply has a given partition line, which can be chosen in many ways, one obtains a well-defined part—whole relation for every shape-determinate. In order to see this, one has, however, to remember that we often lack concepts for the ultimate determinates. O u r ultimate determinate terms of both colours and shapes refer to a disjunction of ultimate properties without of course being identical with any such disjunction. 'Turquoise blue' refers to a range of blue nuances, and 'rectangular' to a range of shapes. Compare, for example, Figure 4.2 with the previous one, Figure 4.1. The shapes which parts 1, 2, 3, and 4 have are of course different shape-determinates. 48

EXCLUSIVE AND INCLUSIVE QUALITIES

Part 1

Port 2

Part 3

Part 4

Figure 4.2

Figure 4.3

Assume that we lay the partition line on the rectangular shape in a different way (see Figure 4.3). The subshapes which the rectangular shape then includes are completely different from those which appeared with the earlier partition line; this is so independently of where the cut is made and how small the parts are. Here we find something which distinguishes non-quantifiable inclusive properties from (continuous) quantifiable ones. In the case of quantifiable properties like mass and volume, every determinate which can be found along one partition line can also be found somewhere along every other partition line. The approach to the part—whole problems taken here, with its introduction of a 49

ONTOLOGICAL INVESTIGATIONS

conventionally determined partition line, allows shape to be considered as an additive property. If such a line is given, one shape can be conjoined to another, and these two can be added in a well-defined way to a third. They form one new shape, not a shape pattern, as would have been obtained if shape were an exclusive property. Thus one can define addition even for nonquantifiable properties, but only for inclusive ones. 4.2 INTENSIVE AND NON-INTENSIVE MAGNITUDES The distinction between inclusive and exclusive properties introduced here is manifestly similar to, but not identical with, Kant's distinction between extensive and intensive magnitudes. 6 The magnitudes Kant calls extensive are distinguished by their (1) being extended, (2) being additive, and (3) having numerous parts. These three criteria also delimit inclusive properties. This does not mean, however, that Kant's intensive magnitudes are captured by the notion 'exclusive property'. Exclusive properties are distinguished by their lacking these three features. Exclusive properties (1) can lack extension, (2) are not additive, and (3) lack parts. But Kant's intensive magnitudes are not the contradictory opposite of his extensive magnitudes. An intensive magnitude for Kant is primarily a magnitude which allows of gradations, more particularly of existence gradations. If ei is more intensive than then, in a manner of speaking, ej exists to a greater degree than does e 2 . Density is an intensive magnitude, or mass intensity. Two bodies with the same volume can have different masses. In the volume with the greater mass, mass exists more intensely. The distinction between inclusive and exclusive properties is introduced with the help of a thought operation. In thought, we shorten a thing in which the property we are interested in is conceived to be instantiated. If the shortening necessarily implies the disappearance of the property and the actualization of another similar property, then the property is inclusive, otherwise exclusive. As I see it, there is another, quite different thought operation hiding behind Kant's distinction between extensive and intensive properties, namely the stretching or compression of a thing in which the property under consideration is conceived to be instantiated. Different properties behave differently under such an operation. If we compress a thing that has mass, the same mass is 50

EXCLUSIVE AND INCLUSIVE QUALITIES

instantiated in spite of the compression; if we compress a thing with a spherical shape in such a way that the spherical shape is retained, then of course the same shape is instantiated after the act of compression. In this respect, mass and shape are similar. One and the same mass, and one and the same shape, can be instantiated in volumes of completely different size. But there is also a difference here between mass and shape. When the thing is compressed we obtain more mass per unit volume than before, but we do not obtain more spherical shape per unit volume. Contrariwise, if we stretch the thing we obtain less mass per unit volume than before, but not less spherical shape. A mass, as distinct from a shape, always has in this sense a certain intensity. Shape is an inclusive property which lacks intensity. Mass is an inclusive property which has intensity. Density is both an exclusive property and an intensive magnitude. Colour hue is, like density, an exclusive property; but it is not an intensive magnitude. On the other hand, every colour hue has an intensity. If we imagine compressing a thing with a single-coloured surface, we can imagine the quantity of the colour hue being retained in spite of the compression, but the quantity of colour hue per unit area is changed. We obtain a higher intensity of the hue. The distinction between hue and intensity manifests itself for us directly in perception, the thought experiment brings out that intensity is intensity of hue. 7 I hope that it is now clear that my distinction is not the same as Kant's. There are in fact two distinctions to be made, one based on the thought operation of shortening, and the other on that of compressing or extending. The former gives rise to the distinction between inclusive (= extensive) and exclusive properties; the latter to the distinction between intensive and non-intensive properties. In distinguishing between the above two distinctions, one obtains a rather surprising insight into 'the ontology of the photograph'. Shape is an inclusive property which lacks intensity. That is what makes it possible for one and the same shape to exist in exactly the same way in both a large area (reality) and a small area (the photo of reality). Mass is, on the other hand, an inclusive property with intensity, and thus cannot be diminished without its 'way of existing', i.e. its intensity, being changed. In this chapter three distinctions and their interconnections have been introduced: exclusive/inclusive, intensive/non-intensive, and 51

ONTOLOGICAL INVESTIGATIONS

quantifiable/non-quantißabU. They can be graphically represented as follows ('-q' means non-quantifiable): Intensive Non-intensive \(8)

\ l ) Exclusive

( 2 ) \

< 7 > \ q

q

\ \(6)

\ з ) Inclusive

v

' \ q - q \

( 5 )

\ q

Figure 4.4

Let us briefly categorize the examples we have discussed by reference to this table. Region (1) contains densities. I shall not discuss whether colour intensities are quantifiable properties and should be placed in the first box or whether they give rise merely to an ordinal scale and should be placed in box (2). But, as we see, this box need not be empty. In box (5) we have shapes, in (6) masses and volumes, and in (7) colour hues. Boxes (3), (4), and (8) are empty. The interesting thing about the table is that it now suggests the question whether they are necessarily empty. In one sense I think that boxes (3) and (8) are necessarily empty. The quantitative magnitudes of box (1) are connected with the thought operation of compressing/stretching, and those of box (6) with the operation of partitioning/adding. It may be argued that every quantitative magnitude has to be defined by means of such a thought operation. If this is true, and if we restrict ourselves to spatial considerations (notice that colour hues in physics bring in time since they are regarded as a kind of oscillation; oscillations are necessarily extended in time), the operations referred to may well be the only fundamental operations available. It then follows trivially that box (8) is empty since there is no fundamental operation left; it also follows that quantities placed in (3) must be fusions of quantities belonging to boxes (1) and (6). 52

EXCLUSIVE AND INCLUSIVE QUALITIES

4.3 SEMI-EXCLUSIVE AND SEMI-INCLUSIVE PROPERTIES The distinction between exclusive and inclusive properties rests on the thought operation of shortening or partitioning, and was introduced by reference to some definitions of Armstrong. So far, however, I have put forward no explicit definitions of my own for exclusive and inclusive properties, respectively. The time has come to provide such definitions. I shall hereby use the concept of 'part' as an abbreviation for the expression 'spatial parts defined by the thought operation of shortening', so that all parts spoken of will be parts along a partition line. The examples I have used have been concerned with properties which are determinates; all the same, there has been an implicit reference also to the corresponding determinables. This will now become explicit. T h e definitions which hold for the examples given above of volume, mass, shape, colour hue, and density are: D4.1: A determinate property is exclusive if and only if each possible part of all its instances instantiates this very same property (under the same determinable). D4.2: A determinate property is inclusive if and only if each possible part of all its instances instantiates some other property under the same determinable. The first definition amounts to the same as Armstrong's definition of homoeomerous properties except that it talks of determinables. But in definition 4.1 this is of no importance since if the parts have the same determinate they necessarily have the same determinable. The second definition differs from Armstrong's definition of anomoeomerous property in at least two ways. 8 First, determinables make here an essential difference since they restrict the clause 'instantiate another property'. Second, inclusive properties become contraries to exclusive properties, whereas anomoeomerous properties for Armstrong are the contradictory opposites of homoeomerous properties. The first difference has the notable consequence that categories as such can be neither exclusive nor inclusive since they cannot be determinates of any determinable. The second difference means that there may be interesting things in between what is exclusive and what is inclusive. And a mere look at definitions 4.1 53

ONTOLOGICAL INVESTIGATIONS

and 4.2 is needed in order to see the possibility of further definitions. If the word 'each' is exchanged for 'some' we obtain the following definitions: D4.3: A determinate property is semi-exclusive if and only if some possible parts of every such property-instance instantiate the same property (under the same determinable). D4.4: A determinate property is semi-inclusive if some possible parts of every such property-instance instantiate another property but under the same determinable. In section 4.1 above, I mentioned in passing that patterns must be considered inclusive properties. Armed with the distinction inclusive/ semi-inclusive we shall now look at two different patterns, a continuous colour spectrum and a line with a chequered pattern. In the former case all parts generated by shortening have a different colour pattern than the original spectrum, which means that the colour spectrum is a pattern which constitutes an inclusive property. If we think of the chequered line we may think away successive bits from one of the ends yet it retains the same chequered pattern. But we cannot retain the pattern if the operation is subsequently applied until the remaining parts are arbitrarily small. Somewhere on the way we meet that original unit which when repeated gives rise to the chequered pattern. And when part of this unit is taken away the chequered pattern disappears. Chequered patterns, we have to say, are both semiexclusive and semi-inclusive. Definitions 4.3 and 4.4 are not, like definitions 4.1 and 4.2, contrary opposites, and this means that both can truly apply to the same entity. When this is the case one can use the label which one thinks stresses the aspect one wants to stress. One conclusion which follows from what has been said can be stated thus: Some patterns are inclusive, some are semi-inclusive. The concepts of 'semi-inclusiveness' and 'semi-exclusiveness' are of importance for some of the points to be made in the rest of the book, but the fundamental distinction is the distinction between exclusive and inclusive properties. This fact will be underlined by the analysis of patterns which follows in chapter 6, where I introduce a distinction between first order and second order properties. I think that all first order properties are either exclusive 54

EXCLUSIVE AND INCLUSIVE QUALITIES

or inclusive. But if this belief of mine is false it makes no difference to the rest of the category system developed here. 4.4 EXCLUSIVE AND INCLUSIVE SUBSTANCES We shall now try to apply the operation of shortening, and the definitions of exclusiveness and inclusiveness based thereon, to substances instead of properties. Assume that we have a macrophysical thing composed of a certain stuff, for example, a certain lump of copper. If we think away a part of the thing (a small piece of copper), this does not affect the substance of the remaining thing. It consists of copper too. If we disregard all problems with microphysical things, the shortening operation can be repeated an arbitrary number of times and the remaining piece will still be copper. This means that copper, like the other elements, is, when conceived on the macrophysical level, an exclusive substance. It conforms to the following definition: D4.5: A substance is exclusive if and only if each possible part of all its instances instantiates the same substance (under the same determinable). The determinable here in question is 'element'. A semi-exclusive and a semi-inclusive substance is a substance where some of the possible parts of its instances are the same substance and some parts instantiate another substance-instance but the same determinable. With regard to properties we noted that some patterns were semi-exclusive and semi-inclusive. The obvious counterpart with regard to substances is a blend of elements. Such a blend forms a pattern of different elements in space, and is not itself an exclusive substance, i.e. not a 'kind of stuff. Let us now see whether we can also find examples of inclusive substances or at least inclusive substances that can be regarded as inclusive in the way copper was regarded exclusive. We shall not concern ourselves with microphysical considerations since our aim is merely to shed light on the distinction exclusive/inclusive. The definition we shall try to apply is the following: D4.6: A substance is inclusive if and only if each possible part of all its instances instantiates another substance under the same determinable. 55

ONTOLOGICAL INVESTIGATIONS

Assume that the determinable in question is element, and that we have a thing of a certain substance placed along a partition line. T h a t this substance is inclusive means that if we think away a bit of the thing then we should get another element, which in turn should include still another element, and so on. But this seems to conflict with our intuitions about substances. Substance-instances should be unities in a sense not required with regard to properties. If this intuition is correct all substances are exclusive. 4.5 INCLUSIVENESS IN T I M E In the thought experiments which we have carried out so fir, it has mostly been spatial parts which have been thought away. The actual conclusion of the preceding discussions is that mass and shape are inclusive in space, and that colour hue and density are exclusive in space. But the investigation of whether qualities are exclusive or inclusive with respect to time is just as straightforward. T h e thought experiments to be performed in that case would then involve thinking away parts of a thing's 'life time'. The conclusions are in general obtained immediately. In principle a thing can have the same colour hue throughout the whole of its 'life'; to think away a part of this period does not change the colour. Colour hues are exclusive in time. The same clearly holds not only for density, but also for the properties mass and shape, which are inclusive with respect to space, i.e. properties which are inclusive in space can be exclusive in time. Patterns in space can also be exclusive in time. If we imagine, on the other hand, a single-coloured thing which changes colour in such a way that during a certain period it goes through the spectrum, then we have a direct correspondence in time to the inclusive universal in space constituted by an ordinary spectrum. A thing whose density, mass, or form changes with time also has patterns which have a kind of second-order existence (cf. section 6.7), but the patterns are nevertheless properties inclusive in time. Apart from second-order properties, the fundamental ideas of modern physics, that is of Newtonian mechanics, thermodynamics, the theory of relativity, and quantum physics, seem to refer only to properties and substances which are exclusive in time. These theories, however, contain many notions referring to properties which are inclusive in space: besides mass, one can mention, for 56

EXCLUSIVE AND INCLUSIVE QUALITIES

example, charge, voltage, work, and energy. But all of these are exclusive in time. Without meaning to, one runs into the question of whether there exists or can exist a sort of universal which is both a first-order universal and inclusive in time. It is in this context that actions and functions become truly ontologically interesting.

57

5 ACTIONS AND FUNCTIONS

M y interest in actions and functions in what follows remains ontological. This needs to be stressed, because, if there is among philosophers a strong leaning towards nominalism with regard to qualities, this is as nothing compared with the dominance of the view that there are no 'actions in re' but only actions under certain descriptions. This is also the reason why I shall not tackle the ontology of actions and functions at once, and adopt instead an indirect approach. I shall take as my point of departure some remarks about explanations of action. M y aim is to show that there is a kind of universal exemplified by both actions and functions. This means that I am here at bottom interested in actions considered independently of any prior intentions. Intentions belong to the category of intentionality which will be treated later on (chapters 13 and 15). (I shall in fact maintain that there can be actions without intentions, and that there is an important truth to be noted about such actions.) Whatever the precise relation is between actions and intentions (mental states and events) there is an important feature of all actions that can be described independently of intentions and that I shall attempt to describe in this chapter. Actions are temporally extended universals.

5.1 APPROPRIATENESS AND R E D E S C R I P T I O N S In the discussion in Anglo-American philosophy of what it means to explain an action there are two analyses - still seemingly opposed - which provide us with our starting point. The one analysis has as a representative, for example, William Dray, and 58

ACTIONS AND FUNCTIONS

coincides with claims of the so-called methodological individualists. According to Dray, 1 an action is explained when one is informed why, in a given situation, it was appropriate; and an action's being appropriate is the same as its being rational in the given situation. An important aspect of the analysis in this context is that according to it every explanation of why an action was performed contains a reference to the situation in which it was performed. The other analysis is represented by, among others, Charles Taylor and A.I. Melden. 2 They have both made the observation that a why-question and a what-question in relation to a certain action can receive exactly the same answer. The questions 'What is he doing?' and 'Why did he open the window?' can both be answered by 'He's airing the room'. This leads them to say that an explanation of an action consists of a redescription. T h e action of opening the window is 'redescribed' in the above example as the action of airing the room. According to this analysis, the explanation does not contain a reference to the situation in which the action is performed. As a consequence it may be said that the two analyses do not coincide, but neither do they conflict, as will soon become apparent. In the case of actions there are actually two different kinds of why-questions. Taylor and Melden miss an important point with their 'redescriptions', namely that they can often be hierarchically ordered. 3 Compare the following three answers to the question 'What is Director X doing?': (1) 'He is replacing the machinery.' (2) 'He is reducing costs on the production side.' (3) 'He is maximizing profits'. Assume that after posing the question one receives answer (1). One can then ask, 'But why is he replacing the machinery?' A relevant answer is clearly that 'He is reducing costs on the production side,' i.e. answer (2) to the original what-question. One can now ask once again why he is doing that; and now 'He is maximizing profits' is relevant, i.e. that answer which corresponds to answer (3). It seems reasonable to draw the following conclusions from this example. The question ' What is X doing?' can sometimes be given many answers which stand in a particular relation to one another. They can be ordered according to the principle that, if upon 59

ONTOLOGICAL INVESTIGATIONS

receiving an answer one responds with the question 'But why?', then those of the original answers which can now function as an answer to this question can be said to be actions of a higher degree of abstraction. Assume once again that we pose the question 'What is Director X doing?', but we now receive answer (3), i.e. 'He is maximizing profits'. We do not now ask why but how he does that, i.e. in what way he does it. In this case the answer will lie not on a more abstract level in the hierarchy constituted by the original answers, but on a more concrete level, as is the case with answer (2): 'He is reducing costs on the production side.' We can now ask how he does this, and receive the answer 'He is replacing the machinery.' Depending on whether we pose a why-question or a how-question, we come to move upwards or downwards in the action hierarchy. A few things should be noted with regard to the generality of the last claim. Let me, for the sake of simplicity, speak metaphorically of 'higher' and 'lower' actions instead of more abstract and more concrete descriptions of an action. The hierarchy I have used as an example functions in such a way that a 'higher' action in space and time includes those which are 'lower', while a 'lower' action can be seen as the way in which the 'higher' action is performed. This does not mean, however, that the way in which an action is performed can always be considered an action which is smaller than the actual action in space and time. If a person runs quickly, he runs in a certain way, but quickness can never be seen as an included action, and an action's being quick must of course coincide in space and time with the action itself. The same holds for everything which falls under the concepts which Gilbert Ryle brings together under the term 'heed concepts', i.e. concepts such as 'observant', 'careful', and 'thoughtful'. 4 The questions of the why? and how? type which have been considered up to now have reference only to one level, and their respective answers have reference to a higher or lower level. However, there are also questions which have reference to several levels at the same time, for example, 'Why is X maximizing profits by reducing costs on the production side?'. In these cases it is inadequate to give answers which refer to a higher action, such as 'He is maximizing profits'. It seems rather that one is looking for an answer which refers to the situation in which X finds himself, and which will make it understandable that he acts as he does. 60

ACTIONS AND FUNCTIONS

Here we find the kind of question on which methodological individualism has focused. What is interesting is not that we run up against such questions but rather that we run up against another kind of question. It should now be clear that to explain actions one must distinguish among three different sorts of question: (a) 'Why does person Ρ perform action A?' (b) 'How does Ρ perform action A?' (c) 'Why does Ρ perform action A in way W?' The answer to question (a) refers to a higher action which encompasses A both spatially and temporally. The answer to question (b) can, but need not, refer to a lower action which is included in A. The answer to question (c) refers to the situation in which Ρ finds himself, and to its 'logic'.5 The analysis advanced by Taylor and Melden fits question (a), while that advanced by Dray and the methodological individualists fits question (c). Someone who poses question (a) is confused about action A: it appears to him to be somewhat unintelligible. And if he is satisfied with the answer he receives, then this is because A has been placed in a larger pattern, it has been located within an action which seems understandable. When question (c) is posed, action A is in itself intelligible, but the way in which it is performed is not. If one is satisfied with the answer, then one has received new information about the situation or person, which makes it understandable why A is performed in way W in just that situation. Both explanation patterns thus presuppose that there are actions which in themselves do not require explanation (relative to the explanation being given). Question (c) above ('Why does Ρ perform A in way W?') normally presupposes that A can be performed in a number of different ways. The question is why just one particular alternative among many has been chosen. Many who have concentrated their interest on this type of question have been inspired by theories in marginal utility economics (e.g. Weber and Popper). Here, as so often happens in the philosophy of science, some ideal is largely realized in a particular scientific discipline and then the philosopher of science tries to raise it to the status of a general norm. To make what I said about different types of explanation more concrete, I shall present and analyse the methodological individualists' 61

ONTOLOGICAL INVESTIGATIONS

paradigm case - the theory of how one calculates a person's or a household's demand for certain commodities. I shall simplify the situation in the way common in elementary accounts of the theory. A person with a certain sum of money finds himself in a commodity market which has the curious feature of containing only two sorts of commodity; let us call them food baskets and fun baskets. Each food basket is exactly similar to every other and contains diverse foodstuffs, and each fun basket is also exactly similar to every other and contains various possibilities for amusement. The person in question has a particular preference system. He wants in principle to have as many food and fun baskets as possible, but he is restricted by the fact that the baskets cost money, of which he has only a limited quantity. His options are not, however, completely predetermined. He can, within certain limits, choose to buy either food baskets or fun baskets or a mixture of both, and here there enters another aspect of his preference system. Presented with a choice between on the one hand 23 food baskets and 25 fun baskets, and on the other hand 25 food baskets and 23 fun baskets, he perhaps finds that he is indifferent. It makes no difference to him which alternative he chooses, which does not mean that it would also make no difference to him were he to receive only 23 baskets of both sorts. If one constructs a coordinate system where the axes represent the number of respective baskets, one can draw so-called indifference curves, i.e. lines joining the points in the coordinate system representing alternatives to which the person in question is indifferent. Given certain assumptions about preferences, these curves are always convex with respect to the origin; and a part of the person's preference system is that, of all the curves he might possibly find himself on, he prefers that which is furthest from the origin (see Figure 5.1). The question now is how a reasonable or rational person who wants to maximize his well-being or utility should behave in this situation. The answer is as follows. The number of food baskets the person could have bought if he had spent all his money on food is registered on the vertical axis, while the maximum number of fun baskets is registered on the horizontal axis. These two points are joined, thereby giving a line which represents those cases where the person truly uses all his money. If he then buys the number of food and fun baskets determined by the point at which an indifference 62

ACTIONS AND FUNCTIONS Food baskets Indifftrtnct curves it where well-being or utility

/

—I

1

Budget line: delimits the / m a x i m u m quantity which can be bought

1 тг I

Fun baskets

Figure 5.1

curve touches this line (since the curves are convex and do not cross one another there is only one such indifference line), then he has clearly ended up on that indifference line which lies furthest from the origin. By doing so he has also maximized his utility in the given situation. If one assumes that all the persons in a two-commodity market are at least approximately utility maximizers and thus rational in that specific sense, and if one knows the preference systems and incomes of all the persons, then one can in principle calculate the total demand for the respective commodities. It is in this way that what has been said fits in to certain theories in economics. Let us now consider this situation in relation to the already distinguished types of questions, i.e. to: (a) 'Why does person Ρ perform action A?' (b) 'How does Ρ perform A?' (c) 'Why does Ρ perform A in way W?' If A means maximizing well-being or utility and W means buying 10 food baskets and 20 fun baskets (per month), then the following three questions arise: (ai) 'Why does Ρ maximize his pleasure or utility?' (bi) 'How does Ρ maximize his pleasure or utility?' (ci) 'Why does Ρ maximize his pleasure or utility by buying 10 food baskets and 20 fun baskets?' 63

ONTOLOGICAL INVESTIGATIONS

The answer to question (ci) follows the schema of methodological individualism and can be explained with the help of 'the logic of the situation'. Such an explanation, if it is seen as definitive, presupposes that the answer to question (aj) is understood as selfexplanatory; no explanation by redescription is thus necessary. The answer to question (bi) is of course that Ρ maximizes his wellbeing or utility by buying 10 food baskets and 20 fun baskets. If A is instead the action of buying 10 food baskets and 20 fun baskets, and W is the buying of both the food baskets and the fun baskets at the general store (instead of, say, at the special food and fun stores), the following three questions arise: (a 2 ) 'Why does Ρ buy 10 food baskets and 20 fun baskets?' (b 2 ) 'How does Ρ buy his 10 food baskets and 20 fun baskets?' (c 2 ) 'Why does Ρ buy his 10 food baskets and 20 fun baskets at the general store?' The answer to question (b2) is of course that Ρ obtains the 'baskets' by buying them in the general store. Question (c2) is answered by referring to 'the logic of the situation': perhaps it is quicker at the general store and Ρ is always in a hurry. The answer presupposes that the aim of buying 10 food baskets and 20 fun baskets is seen as given. If one accepts this, question (a 2 ) does not have to be answered. If one does not accept this, then one demands an explanation by redescription, for example, that Ρ is maximizing his well-being or utility, in which case we find ourselves once again faced with the previous set of three questions. This discussion may, for all its oversimplification, show that even in the methodological individualists' paradigm case we must distinguish between the two types of explanation mentioned ('appropriateness' and 'redescription'), and see them as complementary. 5.2 FUNCTIONS The apparently self-evident distinction between teleological and functional explanations that is often made today has, as is well known, not always been regarded as self-evident. The two kinds of explanation have sometimes even been seen as different names for the same thing. We shall now investigate how great the similarity is between explanations of actions and explanations of functions. 64

ACTIONS AND FUNCTIONS

Figure 5.2

We shall begin by looking at the following three questions: (a) 'Why does thing Τ have function F?' (b) 'How does Τ fulfil F?' ( = 'In which way does Τ fulfil F?') (c) 'Why does Τ fulfil F in way W?' We shall take a look at the functions of a compressor in a device such as a refrigerator. Consider Figure 5.2. The description of the refrigerating device is approximately as follows. When a liquid is vaporized its surroundings are cooled since heat is required in order to change the liquid to steam. This vaporization takes place in the refrigerating element. With the help of a system of valves and a piston in a cylinder, the steam is compressed and at the same time its temperature is raised. The compressed hot steam is forced from the compressor to a condenser, where it is condensed to a liquid by cooling. The condenser itself is cooled with air or water. From there, the liquid goes to a fine nozzle, the regulating valve. By diminishing the pressure in the nozzle, a part of the liquid is vaporized, thus leading to a lowering of its temperature. The liquid then continues on to the refrigerating element, where it is vaporized. In this way it takes up heat from the space that is to be chilled. The temperature inside the cooling device decreases. The 65

ONTOLOGICAL INVESTIGATIONS

steam which has been formed is sucked to the compressor, where the process begins anew. There are at least three answers to the question: ' What does the compressor do?' (cf. page 59 above, 'What is Director X doing?'). (1) 'With the help of a system of valves, it sucks steam in from the refrigerating element, compresses it, and forces it toward the condenser.' (2) 'It creates vapour pressure and pumps steam around.' (3) 'It pumps heat from the refrigerating element to the condenser.' With respect to the first answer we can ask 'Why does the compressor suck in steam, compress it, and force it toward the condenser?' An answer is obtained from (2): 'It thereby creates vapour pressure and pumps the steam around.' We have received a redescription. But then we can ask 'Why does the compressor create vapour pressure and pump the steam around?' And now we find the answer in (3): 'It thereby pumps heat from the refrigerating element to the condenser.' The answer gives us yet another redescription. If we begin instead with answer (3), we can pose the question: 'How does the compressor pump heat from the refrigerating element to the condenser?' An answer is received via the specification which answer (2) provides: 'Through creating vapour pressure and pumping steam around.' But we can once again respond to this with a How?: 'How does the compressor create vapour pressure and pump steam around?' An answer and yet another specification is given by (1): 'By sucking steam in from the refrigerating element with the help of a system of valves, compressing it, and forcing it toward the condenser.' The same kind of redescriptions and specifications which are present in explanations of actions are also present in the explanation of functions. This means, among other things, that functions can also be given a 'higher' and 'lower' hierarchical ordering. In the example given, answer (3) presents a function of the compressor which is of a higher degree of abstraction than the functions described in answers (2) and (1). This is equally true of functions within an organism - for example, the heart in a body - as of functions within a machine like the cooling machine. But let us now take a look at a question corresponding to that which has been discussed by the methodo66

ACTIONS AND FUNCTIONS

logical individualists in the case of actions: 'Why does the compressor pump heat by creating vapour pressure and pumping steam around?', i.e. a question of the form 'Why does Τ fulfil F in way W?'. The question is difficult to answer because it is difficult to imagine another way in which the compressor would be able to fulfil the function of pumping heat. As I pointed out earlier, a question of the form 'Why does Τ fulfil F in way W?' presupposes that there are alternative ways to W by which Τ can fulfil F. Let us nevertheless assume that this is the case. What kind of answer would we then receive? That which lies nearest to hand is that W is more effective than the other ways. Such an answer places the 'choice mechanism' outside T . It is not the machine that chooses the way in which it performs the function, but the human beings who constructed the machine and who are externally related to it. The corresponding question in the case of actions, 'Why does Ρ perform A in the way W?', does not lead to a reference to something totally outside P. This is because P, as distinct from T, is assumed to be a conscious being which has the capacity to choose, which in its turn means that we attribute to Ρ the capability of being the choice mechanism. Here the difference between teleological and functional explanations reveals itself. But this difference should not be used to hide the similarities described above with regard to explanations via redescriptions and specifications. In the case of functions it is often reasonable to pose a completely different kind of question, namely, 'Why is F fulfilled by T?', for example, 'Why is heat pumped by a compressor?'. In this question interest is not focused on a given thing but on a given function. One does not ask why, how, or in which way a given thing performs a certain function; one asks rather why a certain function is performed by a certain thing. The question presupposes that other things can perform the same function, and it, too, leads to a reference outside the machine: the machine's constructor felt that Τ fulfilled the function better than U. Refrigerators, for example, do not need compressors to pump heat. There are other devices, such as those used in the old absorption refrigerators. The question 'Why is F fulfilled by T?' has, in the case of actions, its counterpart in the question 'Why is A performed by P?'. Interest can be focused on the action instead of on the person. 67

ONTOLOGICAL INVESTIGATIONS

In fact a reasonable answer here often includes reference to something outside of P. The person Ρ is part of a group or community, and the answer explains why just Ρ was chosen from the group to perform A. This should also be kept in mind in discussing the relation between teleological and functional explanations. 5.3 BASIC A C T I O N S AND BASIC F U N C T I O N S When someone is said to have performed a particular action it is often implicit that the person could have done otherwise.6 An implicit reference is made to a possible choice. This aspect of actions is clearly manifest in the kind of explanation in which the methodological individualists are interested. This feature of actions is also connected to the thought that the person acting in some way himself occasions or brings about the action. The person is conceived of as the cause of his action, i.e. the acting person is a causa sui. A thing which performs functions is conceived on the other hand neither as an entity which chooses, nor as a causa sui. Methodological individualism's explanations are teleological, but functionalist explanations are not. In the other kind of explanation, redescriptions, however, there is no allusion to efficient causes and choices in the relation between the explanandum and the explanans. This kind of explanation is equally applicable to functions and to actions, and thus there is more to be said with regard to the similarities between actions and functions. 7 As we have seen, both actions and functions can be hierarchically ordered: one can in a given context speak of actions and functions which are above or below one another. This structure has an intimate connectioin with what has come to be called 'the accordion effect' in discussions regarding the occurrence of 'basic actions'. T h e term 'basic action' derives from an essay of the same name by Arthur C. Danto. He holds that if we accept the notion of an action at all, then we must also accept that there are actions which are basic or simple 'in the same sense in which the old "simple ideas" were said to be: they were not compounded out of anything more elementary than themselves, but were instead the ultimately simple elements out of which other ideas were compounded'. 8 68

ACTIONS AND FUNCTIONS

Danto takes as his starting point the claim that an action can cause or have as a consequence a great deal of things, as well as the claim that such consequences cannot be seen as parts of the action in question. That a person throws a stone can have as a consequence that a window breaks. Here the action (to throw a stone) can be clearly distinguished from this consequence (that a window broke). In such a situation, however, we often simply say 'He broke a window', as though it were only a question of an action, not of an action and its consequence. The action 'He threw the stone' becomes, so to speak, included in the action 'He broke the window'. It is this relation which constitutes 'the accordion effect'. Danto maintains that if one accepts the concept of action, then one accepts that not everything can be reduced to causes or consequences, which is the same as saying that somewhere in the accordion there must exist a 'first' or 'lowest' action, a basic action. John R. Searle has given the following example; the main role is played by Gavrilo Princip, who shot the Archduke of Austria in Sarajevo in 1914: He produced neuron firings in his brain contracted certain muscles in his arm and hand pulled the trigger fired the gun shot the Archduke moved a lot of air molecules killed the Archduke struck a blow against Austria avenged Serbia ruined Lord Grey's summer season convinced the Emperor Franz Josef that God was punishing his family angered Wilhelm II started the first World War 9 I believe that most people are inclined to agree with Searle when he says that that which lies below the lower line and above the upper line, as well as that which stands on the side, are not actions at all; while everything between the lines can be an action. Danto would presumably analyse the actions between the lines in such a 69

ONTOLOGICAL INVESTIGATIONS

way that 'avenged Serbia' is a consequence of 'struck a blow against Austria', which is a consequence of 'killed the Archduke', which is in its turn caused by 'shot the Archduke', which is caused by 'fired the gun', which is caused by 'pulled the trigger', which is not caused by anything at all but is a basic action. For Danto, all composite actions are basic actions taken together with certain of the causal consequences of the basic action. If we compare Searle's example with our earlier discussions in this chapter, a simple observation can be made. The two lines in the schema correspond to the two endpoints of the regresses which arise from a given action when one poses how-questions and whyquestions. If we were to begin, for example, with the action shooting the Archduke and ask how?, and repeat the question with every answer we receive, then we would move upwards and stop at 'pulled the trigger'. Why-questions lead us in an analogous way to 'avenged Serbia'. This observation immediately raises the question whether there does not exist an 'accordion effect' also in the case of functions, and whether there do not exist basic functions in analogy to basic actions. Drawing on the discussion above of the cooling device, we can set up the following schema: The compressor has a crankshaft which goes round has a piston which moves up and down sucks in steam, compresses it and forces it further along creates vapour pressure and pumps the steam around pumps heat vibrates makes noise disturbs the neighbours What lies between the two lines are the compressor's functions, and what is above or below the lines are things other than functions. With the help of how-questions and why-questions we can move up or down in the function hierarchy. The function just below the upper line in the schema ('sucks in steam, compresses it, and forces it further along') seems to consist of three basic functions. The function 'pumps heat' can be understood as a consequence of the functions 'creates vapour pressure and pumps 70

ACTIONS AND FUNCTIONS

steam around', which can in its turn be understood as a consequence of the basic functions. The schema seems in all respects to be analogous to Searle's. The discussion of basic actions has only taken into account one of the lines - the upper one - in the schema in which the accordion effect was presented. But if there are two limits, there must be two kinds of basic actions. In what follows, I shall call the traditional basic actions simply 'basic actions', and call those actions lying nearest the lower line 'ultimate actions'. In a corresponding way, I shall also speak of basic functions and ultimate functions. This has as a consequence that what I previously called 'lower' actions (functions) are those which are near to the basic action (function); and those I called 'higher' actions (functions) are those near to the ultimate action (function). The question which now arises concerns the nature of the relations among actions, basic actions, and ultimate actions, and among the corresponding functions. As I understand Dan to, he feels that an action is always either a basic action or such an action taken together with its consequences. The difficulty with this sort of conception is that it allows no natural limit to an action. In particular, it does not allow the delimitation of an ultimate action from the intended causal consequences to which it gives rise. One can have two types of intentions, those which one considers realized in the performing of the action itself, and those which are realized only if a causal process initiated by the action transpires in a certain way. An example of the latter is the intention of obtaining flowers, which leads one to the action of sowing seeds. In the causal process it is people who sow the seeds, not the seeds themselves. On the other hand it is the plants themselves which grow. The intention in performing the action is that something will occur as a result of one of the causal processes initiated by the action. Not all the consequences of an action add up to a new action.10 Searle says that the relations between actions in the case of the accordion effect is a 'by-means-oF relation, which, without discussing the matter, he says can be either causal or of some other sort. But if the relation is causal, then Searle's view is open to the same objection as was levelled at Dan to; and if the relation is of some other sort, then Searle must clarify what this sort is, which he has not tried to d o . " 71

ONTOLOGICAL INVESTIGATIONS

Searle does define the notion of a basic action, however: A is a basic action type for an agent S iff S is able to perform acts of type A and S can intend to do an act of type A without intending to do any other action by means of which he intends to do A.12 This definition is ambiguous, for Searle distinguishes between two sorts of intentions, 'intentions in action' and 'prior intentions'. Sometimes the intending of an action appears only in the action itself, and sometimes it also appears in the form of a preceding conscious intention. This is the same distinction which Alasdair Maclntyre has marked in terms of the 'coherence model' and the 'planning model' of action. 13 Searle does not himself elucidate the ambiguity in this definition; I shall therefore discuss Maclntyre's distinction. The coherence model is applicable to routine actions, but remarkably enough it also suits many spontaneous actions, i.e. actions where one acts first and thinks later. Actions attributed to small children and animals also belong here. The planning model is suited for actions which are preceded by some form of mental construction or planning ('psychic work'). One has a model of the action before the action is performed (whence the names 'planning model' and 'prior intention'). The actions of deceiving and lying i.e. actions where one tries to get others to think there is a different intention in the action than is actually there - are special cases of the planning model. Such a double structure is impossible in actions which follow the coherence model. The basic idea in the coherence model is that phenomena sometimes reveal a special sort of agreement (coherence); a special sort of pattern arises which cannot be reduced to 'ordinary' causal relations and properties or relations of the kind which belong to inert nature. In this model a basic action is supposed to be the same as that which in a given context is the smallest conceivable pattern of that type. Danto's argument for the existence of basic actions, i.e. for 'atomic actions', also has a striking similarity to the classical argument for traditional atomism. If there do not exist smallest indivisible things - atoms - then 'thing' is not an irreducible category, since things are then constituted of parts which are not themselves things. If there do not exist smallest actions, basic actions, then 'action' cannot be an irreducible 72

ACTIONS AND FUNCTIONS

category, for actions are then constituted of parts which are not themselves actions. If the concept of a basic action is delimited with the help of a notion of a 'smallest possible pattern', 1 4 one immediately sees that it is possible in principle to delimit the concept 'basic function' in a similar way. In the planning model, an action is seen as the realization of an underlying intention, which functions as a sort of efficient cause. In this model a basic action is simply an action which is the first consequence of the intention. Actions which are not basic actions are consequences of actions which are consequences of actions which are consequences of a basic action which is a consequence of an intention. In this respect one cannot speak of basic functions, since things and machines do not have intentions. The two definitions of basic action coincide if it can be shown that 'the smallest pattern' is always the pattern which is the first consequence of a conscious intention. This would mean that in the above schemata the conscious intention would always include only the 'lowest' action in the hierarchy. But this is not the case. One can have a conscious intention which includes the whole hierarchy. In fact, this is essential to training. The better one can do something, the 'larger' is the action which one can directly intend. If Gavrilo Princip had not been familiar with shooting, he would have been restricted to the intention of squeezing the trigger. But if he had been used to shooting, then he could simply have had a direct intention to shoot as such. To be able to do something is the same as to have learned unconsciously to integrate a whole set of subactions — a point to which Michael Polanyi, 15 among others, has drawn attention. The distinction between prior intentions and intentions in action also has implications for the notion of basic action, implications which have not yet been cleared up. The difference between prior intentions and intentions in action is an essential one, and it connects up with the two kinds of explanations of actions discussed above. Prior intentions are necessary in order to speak of choices lying behind the actions, and they connect up with the methodological individualist type of explanation. Intentions in action can occur without there being prior intentions, and are necessary for redescription explanations. I shall for now leave prior intentions aside - they will later reappear in the discussion of the category of intentionality - and concentrate instead on the analysis of 73

ONTOLOGICAL INVESTIGATIONS

intentions in action and of related phenomena such as redescriptions and the accordion effect. 5.4 INCLUSIVENESS IN T I M E The accordion effect understood as applying to both actions and functions means that certain actions and functions include other actions and functions in time. The action of killing the Archduke is completed a moment after the end of the action of shooting the Archduke; and that in its turn is completed only after the action of firing the gun. But firing the gun presupposes pulling the trigger. If one begins with the action of killing the Archduke, and in a thought experiment takes away a bit of time at the end of the action, one obtains another action. The original action necessarily disappears. The same thing happens with the other actions mentioned. The actions in question behave, then, like inclusive universals. But here it is a question of inclusiveness in time}6 Mass, which is inclusive in space, is: (1) a property which is always extended in space (i.e. it can only exist between two points in space); (2) something which always has parts; (3) always additive. If we investigate the action of killing the Archduke we shall find the same peculiarities, but now in time. The action in question can never be considered point-formed in time; it must be executed between two points in time. It must also contain subactions: this is the real content of the accordion effect. But the action is also additive, even though not in a quantitative sense, as we shall now see. The idea of a 'line of partition' was introduced earlier to make it clear that shape determinates are inclusive properties. If a shape is, in a well-defined way, to include other shapes, these other shapes must be shapes on some given line of partition. We shall now look at a new such example, a cone along a given χ axis. The cone presented in Figure 5.3 is divided into four subshapes, one between Xo and one between X) and x 2 , and so on. If one imagines a motion from the left to right, one can say that the cone shape is obtained by adding the four subshapes. But form χ 0 t o *4 also includes form x 0 to x3, which includes * 0 to * 2 and x0 to Х]. These 74

ACTIONS AND FUNCTIONS

Figure 53

killed the Archduke shot the *Archduke fired the gun t

A

pulled the trigger

>

Figure 5.4

shapes are analogous to the different subactions that make up the killing of the Archduke as a result of the accordion effect. In order that actions ίο to /4 in Figure 5.4 can be a sum of actions, it is not only necessary that ίο to tu to to i2, and ii to t3 be actions, but

/

/

/

/

/

/

/

/

/

Figure 11.4

force) but another velocity, since the acceleration has now had time to retard the velocity. If one follows the ball's path moment by moment, one sees that the result is a parabola-shaped curve. Let us now instead see how the situation would look if the gravitational force directly affected the velocity instead of the acceleration. Then we obtain the Figure 11.4. Here, vg is the velocity to which the gravitational force gives rise. It can be (vector-) added to va, which means that velocity is here explicitly a tendency. We have real partial velocities. The line of trajectory becomes straight, directed at an angle downwards when the partial velocities are those given in the figure. The velocity va is not changed via an acceleration as in the previous case, it is quite simply added to another velocity. Vectorial magnitudes always allow in principle something which counteracts. But a theoretical system, like Newton's mechanics, can exclude that this categorial possibility can be realized for certain vectors. Such vectors are, however, nevertheless tendencies. If, moreover, they are not caused by anything outside of the thing they belong to, then they instantiate the category causa sui. (I shall say a few words in parenthesis on the force relation. In my presentation of Newton's second law there were two superposition principles, one for forces and one for accelerations. I have spoken of the latter, while physicists normally speak about the former. If one has an instrumentalist view of Newton's second law it makes no difference which of the two one speaks about. Law (3) can be derived from (1), (2), and (4); and (4) can be derived from (1), (2), and (3). But ontologically speaking such a conventionalist view is out of the question. Accelerations are a special kind of property, i.e. tendencies, while force is a relation. The superposition principle for forces gives rise to the question whether forces need not also be some sort of tendency because it allows one to speak 167

ONTOLOGICAL INVESTIGATIONS a b o u t co-operative a n d counteracting forces. M y reply to this question is negative. Consider Figure 11.2. W h e n we a d d partial accelerations, w e add properties which refer to one a n d the same thing, and it is not difficult to give real content to the result of the addition. If on the other hand, we add forces, we a d d relations between one thing (the piece of iron) a n d four different things. And to give a real content to the relation which constitutes the addition's sum seems impossible. W h a t does the relation relate? I take as given that a s u m of relations must itself constitute a relation. As far as I can see, the addition in question m u s t be understood instrumentally. T h e forces are added as though they were properties a n d not relations. F r o m a n ontological point of view it is better to speak of the superposition principle for accelerations t h a n of the superposition principle for forces. T h e other criterion for tendencies which I have b r o u g h t forward is that they lack contrary opposites. T h e force relation does not fulfil this criterion either, even if one sometimes obtains the opposite impression. T h e latter is the case w h e n one has two bodies which, d u e to the force of gravity, attract one another b u t which, due to their electrical charge, repel one another. T h e contrary opposites 'attract one a n o t h e r ' and 'repel one another' exist simultaneously as relations between two things, and therefore c a n n o t in fact be contrary opposites; they involve, however, gravitational attraction a n d electrical repulsion. In the case of acceleration this sort of distinction does not exist.) Velocity a n d acceleration are, like all tendencies, universale which are temporally exclusive. T h e y can thus be thought to exist at a m o m e n t a r y point of time. T h a t they point toward other points of time does not invalidate this claim. I n the examples we have j u s t looked at (Figures 11.2-4) one can imagine that it is a pointshaped particle which moves, i.e. a thing whose properties are exclusive in all of the spatial dimensions. T h i s shows very clearly that the vector addition which represents co-operative action a n d counteraction of tendencies is an addition of exclusive universals. Scalar addition of exclusive universals is, as has been pointed out, impossible; scalar addition requires universals which are inclusive in the dimension or direction in which the addition takes place. Tendencies require no spatial or temporal line along which to be added. T h e y can be a d d e d at a point. T h u s there is good reason to keep tendencies and genuine properties separate. T h e examples worked through in this chapter have had to do with quantified tendencies, which m e a n t that it was easy to 168

TENDENCY

represent them by mathematical vectors. This does not mean however, that all tendencies are quantifiable. In section 7.3 it was shown that we usually conceive of many actions as a form of internally produced changes, i.e. as outflows of feelings, emotional states, character traits, or intentions. These latter are, in categorial terms, purely qualitative temporal vectors, i.e. fundamentally the same as Aristotelian substantial forms. We can now see that normally they are best described as a sort of resultant tendency, i.e. their expressions cannot be cancelled. But it is also easy to find examples of qualitative partial tendencies, namely wishes. Different wishes can co-operate and counteract with one another and give rise to an intention ( = resultant tendency). Another example of qualitative co-operative action and counteraction are drives. I would like, once again, to point out that the tendency category exists both in our common-sense metaphysics of action and in modern physics. The tendency category cannot be dismissed as an anthropomorphization of physical reality. 11.3 T E N D E N C Y , P O T E N T I A L I T Y , AND D I S P O S I T I O N The distinction between tendencies and genuine properties also has a bearing on the concept of potential properties. (I discussed potential parts earlier, see in particular page 137; what was then called a potential part is a type of potential property in the sense now to be discussed.) If there only existed genuine properties the contrast with actual properties would be easy. A thing's potential properties should quite simply be those properties which the thing does not have but can have. T h e colour and shape a thing has at one particular moment inhere actually, while all of the colours and shapes which it could have without losing its identity inhere in it potentially. O n e can here note that a thing can potentially and simultaneously have both the colour black and the colour white, which means that potential properties, like tendencies, lack contrary opposites. This similarity does not, however, imply that tendencies should be identified with potential properties. When a tendency is actual there exists - in our terminology above - a 'pointing' toward certain genuine properties. And such a pointing does not exist for 'ordinary' potential properties. Those properties which the tendency in question points to must indeed also be potential properties of the thing, but they are at the same time, so to speak, more than potential properties, without for that 169

ONTOLOGICAL INVESTIGATIONS reason existing actually. A thing with a certain velocity has as a potential property in the ordinary sense all other possible locations, b u t at the present place and velocity it also has as a potentiality, in a markedly stronger sense, those places which the velocity vector points most closely to. H a v i n g m a d e these remarks, I should like to reserve the term 'potential property' for de re possibilities which a r e not in themselves tendencies. T h e term 'tendency' will merely refer to the category of tendency as just explicated. 'Potential properties' are closely connected with the concept of 'dispositional properties'. T h e big flaw in analytical philosophical discussions of dispositions is of course that tendencies are not taken into account. 4 But even a m o n g philosophers w h o consciously advocate the concept of 'tendency' or some related concept w h e n analysing dispositions, there is an insight which is missing or at least not clearly spelt out: 5 A thing's potential properties can be of two different kinds, either de re possibilities to have as well as c h a n g e genuine qualities, or de re possibilities to have tendencies. T o have a tendency is one thing, to have a tendency potentially is another. I shall not devote space here to a detailed analysis of dispositions. M y aim h a s been to show the i m p o r t a n c e of the tendency category; therefore I shall restrict myself to pointing out the usefulness of the concept of the de re possibility of having a tendency. In traditional analyses of the dispositional predicate 'magnetic', usually two situations are envisaged. In the one we have a piece of iron outside of any magnetic field, a n d in the other the piece is imagined as moving towards a m a g n e t . In the first situation there is a potential property, the iron m a y move towards a magnet; in the second situation this potentiality is actualized. But there is also a third kind of situation, the one described when I introduced the concept of tendency (cf. Figure 11.2). A magnetic piece of iron m a y be placed near a m a g n e t without ever moving towards it because of counteracting forces. T h e piece of iron cannot, however, be placed near the magnet without becoming subject to a tendency to move towards the magnet. T o say that something is magnetic is not to say merely that it can or will move in a magnetic field. I t is to say that it can acquire a tendency to move in such a field, a n d that because of this, it sometimes will move. T h i s is in outline, how I think dispositions should be analysed. 6 170

12

EFFICIENT CAUSALITY

The category now to be presented — efficient causality — is related to the category of tendency, but is distinct from it. One difference between them is that while tendencies are a sort of property, i.e. inhere in things, efficient causality is a relation between things. The things (states of a thing, or changes of a thing) between which the relation holds I shall call efficient cause and effect. We shall primarily be concerned not with the cause or the effect, but with the causal relation — efficient causality. Is 'efficient causality' simply the classical concept of efficient cause? I do not simply want to rehabilitate the latter, for in certain respects my category of efficient causality diverges radically from both classical and everyday conceptions. It is similar to them, however, in that it implies a total rejection of the empiricist reduction of efficient causality to constant conjunction. I shall not repeat the refutation of the empiricist conception of causality, but merely refer to the arguments of Ushenko, Bunge, Harre, and Armstrong. 1 If one rejects the identification of causality with constant conjunction a set of problems reappear which would otherwise not have been noticed. The classicial question whether cause and effect must be temporally and spatially contiguous, to take one example, lacks interest vis-a-vis constant conjunctions. Of course constant conjunctions can in principle exist between phenomena lying both next to and not next to each other. But if one takes seriously assertions like 'The cause always brings forth the effect' and 'The same cause always produces the same effect' then the 'contiguity problem' and the other traditional problems of causality intrude: Are cause and effect conceptually distinct? Is the causal relation a

171

ONTOLOGICAL INVESTIGATIONS

relation of necessity? What does the cause's production of the effect actually consist in? We have already discussed questions concerning whether one phenomenon can produce another. The tendency concept subsumes the concept of internally produced change, i.e. change causa sui. The paradigm example was the way a velocity (i.e. an inertial property) produces a change of place. This example does not directly help us here, however, since efficient causality is a relation between things, while internally produced changes exist in things. In the latter case, that which does the producing (the velocity) and that which is produced (the change of place), exist in exactly the same interval in time and space. In the former case, that which produced and that which is produced seem to belong to different intervals in time and/or space. This chapter is divided into five sections. In the first section, non-causal laws are analysed so that in what follows there will be no confusion of causal with non-causal laws. I shall argue, among other things, that numerical causal laws express a relational invariance between grounded relations; the invariance in question is consequently a universal of third order. The second section contains the thesis that efficient causality cannot be understood in terms of classical internal relations, but is nevertheless a very special sort of existential-dependence relation. In the third section, it is shown that since efficient causes can co-operate with and counteract each other, efficient causality always involves the tendency category. Section four discusses the problem of contiguity, i.e. action at a distance versus action by contact, and leads to the apparently amazing view that both of these alternatives must be rejected in favour of a new concept: 'action by mixture'. In the final section it is shown how the concept 'action by mixture' opens up new possible ways of understanding the role of space as a causal factor.

12.1 NON-CAUSAL LAWS It is impossible to understand the category of causality unless one accepts that there are both causal and non-causal laws. Those who try to force all laws into one and the same compartment, either causal or non-causal, are led to unreasonable conclusions. 2 I shall begin by briefly discussing a non-causal law, the Boyle-Charles' or 172

EFFICIENT CAUSALITY

gas law. It says that for gases, pressure (p), volume (V) and absolute temperature (7) satisfy the equation pV = RT, where R is the so-called gas constant. Here it is impossible to say that changes in either temperature, pressure, or volume of the gas bring about or cause changes in one of the others. The relation among the three variables is fulfilled at each moment and no variable has existential priority in relation to the others. It is impossible to speak here of unilateral production; the concept 'mutual' is more adequate. Pressure, volume, and temperature determine one another mutually. What can be misleading is the fact that, when we want to change the variables, we most often see the situation as one where we manipulate only one variable but get changes in the other variables as well. If we set a burner under a container of a certain quantity of gas, we seem only to cause a change in temperature; and if we alter the volume of the container, we seem only to cause a change of volume. Actually, in both cases we cause temperature, pressure, and volume changes. It is we, who are external in relation to the gas, who cause all of the changes. It is not that a change in one of the variables causes a change in one of the others. The variables determine one another mutually. T o show that the gas law does not contain a causal relation I have merely relied on an intuitive understanding of mutuality as well as of the contrast between mutuality and bringing about or causing. In chapter 8 I introduced a distinction amongst three categorially distinct types of relations: external, grounded, and internal relations (= mutual existential dependence). I shall now analyse the gas law more thoroughly by making use of this distinction. Every law has, independently of what it explicitly says, an ontological context. The gas law presupposes the existence of the determinable properties pressure, volume, and temperature, as well as the existence of corresponding determinate properties. Between the determinates of every determinable there are grounded relations of the type 'has a greater pressure than', 'has twice as large a volume as', and 'has the same temperature as'. These relations exist just as much as the determinates they connect and they also belong to the ontological context of the law in question. Between the determinables there are mutual existential-dependence relations; in gases it is logically impossible for pressure to exist without volume and temperature, and vice versa. But this relation 173

ONTOLOGICAL INVESTIGATIONS

is not something expressed in the gas law; it is a silent presupposition hiding in the background. The gas law takes up explicitly either the determinates or the grounded relations. But which? This is a question which must be answered before we can proceed. The determinate properties have as counterparts on the conceptual level classificatory concepts, i.e. concepts whose extension either has or lacks the properties in question. The grounded relations also have classificatory concepts as their counterparts. Either an ordered pair falls under concepts like 'has the same pressure as' and 'has greater pressure than' or it does not. (External relations, for instance 'lies beside', also correspond to classificatory concepts. Classificatory concepts can refer to both properties and relations.) But from grounded relations and corresponding concepts so-called comparative concepts can be constructed. A comparative concept is really a scale. 3 To have a comparative concept such as 'pressure' is the same as being able to order all determinates of the determinable pressure in such a way that of any two instances of determinates either both have the same pressure or one has a higher pressure than the other. One can profitably divide up the different determinate concepts in a suitable way with indices such as pi, p2, . . . If a certain gas falls under the concept '/>35' then it has a higher pressure than do all gases with lower indices, and it has lower pressure than all gases with indices higher than thirty-five. Two gases both of which fall under ^35 have the same pressure. The grounded relations 'has the same pressure as' and 'has a higher pressure than' constitute the bases for the comparative concept of pressure. If the grounded relation 'has the same difference of pressure as' exists, and we can obtain knowledge about it by means of some operation, then we can also construct a quantitative concept of pressure. If a certain quantity of volume gas falls under the concept '2 pascals' then it has the same difference of pressure with respect to a gas which falls under Ί pascal' as it has to one which falls under '3 pascals'; in addition it has a greater pressure than the former and a lesser pressure than the latter. A quantitative concept always presupposes the possibility of a comparative concept and the latter presupposes the possibility of classificatory concepts. These conceptual relations have their ontological counterparts. Grounded relations always presuppose 174

EFFICIENT CAUSALITY

properties, and certain determinate properties are such that their existence also implies the existence of grounded relations; certain determinate properties are so constituted as to imply the existence of some form of 'distance relation', for example: 'has the same difference of pressure as', 'has the same volume difference as'. The variables in the gas law, 'p\ 'V and 'T, refer to the quantitative concepts of pressure, volume and (absolute) temperature. But what do these quantitative concepts refer to? Do they refer to determinate properties or grounded relations? The answer is that, taken one at a time, the quantitative concepts refer to both determinate properties and grounded relations. When, for example, a particular gas is said to have the pressure 5.7 pascals, it is asserted that the gas has a particular determinate of the determinable pressure, but also that the gas has certain particular grounded relations to other gases; between the gas in question and the gas with unit pressure, for example, there holds the relation: 'has 5.7 times higher pressure than'. To the extent that all quantitative concepts have this type of double reference it is not obvious what the gas law is about. If one holds volume in the gas law constant, the law is reduced to 'p/T = constant', and we can carry on our discussion as though we had a law with only two variables. This law implies a whole set of'sub-laws': 'pi if and only if T{, '/»2 if and only if T2', and so on. In every such sub-law the concepts '7У, '/>2\ '7Y, . . . function only as names of the respective determinates not as numerical magnitudes. As regards the relation connecting the determinates that which is expressed above as 'iff — different views are possible. On one view (I shall keep to the first law in what follows) p\ and Τι are only connected by the (external) relation of simultaneity (or by the external relation of temporal contiguity), and the lawlikeness involved only means that this simultaneity always holds between px and T\. This is Hume's analysis; laws are constant conjunctions. If one is not so stingy with universale as is the nominalist Hume, more possibilities present themselves. With Armstrong and others one can maintain that 'laws of nature are dyadic relations of necessitation (or probabilification) holding between universale'. 4 According to Armstrong, there is a special (second-order) universal 'nomic necessitation' which can relate other universals. One universal, F-ness, can necessitate another, G-ness. And this relation, he argues, is not a relation of logical necessity. 175

ONTOLOGICAL INVESTIGATIONS

In the gas law example we then obtain the claim that 'pi-ntss necessitates Ti-ness'. A relation of nomic necessitation holds between the d e t e r m i n a t e s ^ and T\. This relation exists of course in the individual cases between the instances of the properties, but it exists between the properties as such, i.e. as universals. It is this which makes the problem of induction disappear here in a sense: if one knows that nomic necessitation is present in the individual case, then one also knows that it is always present. 5 That the relation is primarily a relation between properties makes it perhaps possible to think of it as a grounded relation. But this is not so. Grounded relations are derivable from the properties between which the relations hold, but this is not so in the case of nomic necessitation. From and '7y no relation of nomic necessitation can be derived. The relation is an external relation. Whether or not one analyses i p i if 7У like Hume or like Armstrong, one can with some justice maintain that the law 'p/T = constant' is nothing other than a clever linguistic construction which at one and the same time describes a range of the sub-laws. The law lp/T = constant' is then in itself no law, but is actually an infinite conjunction of laws: 'p\ iff T\ and 'p2 iff 7У, and so on. Another way of looking at the law 'p/T = constant' is to maintain that it describes a relation of nomic (or factual) connection between grounded relations. The fraction p/T itself lacks meaning so long as the actual reference is assumed only to be the determinate properties. These have as such no numerical values and can just as little be divided by one another as one can divide light blue by dark blue. But certain grounded relations do have numerical values, and the quotient the law expresses receives a real content when it refers to them. The determinate properties are universals of first order, relations between these universals (e.g. all grounded relations) are universals of second order. The best possible interpretation of the law we are discussing, as well as of all numerical laws, is that it refers to a universal of third order, i.e. to a relation between universals of second order (grounded relations). Universals of third order seem a little fanciful, and if all laws had only two variables they would perhaps not be required. Laws could then be identified with infinite conjunctions of sub-laws; and the latter would be the real laws. But if one investigates laws with three or more variables, this apparently self-evident view must be rejected. 176

EFFICIENT CAUSALITY

Let us look at the original gas law: it consists of three variables plus the gas constant, 'p-V = R T\ To interpret it as a relation between grounded relations does not lead ίο any difficulties which are greater than the one we found above, i.e. the necessity of postulating universals of third order. But to interpret the law as a concentrated expression of an (infinite) set of constant conjunctions or necessity relations between determinates appears impossible. Since there are three variables there do not exist (as in the case of two variables) any one-one relations. For every value of one variable there exists an infinite set of ordered pairs of values of the other two which satisfy the equation. Where, then, is the constant conjunction? And what happens to Armstrong's necessity relation when one determinate 'nomically necessitates' an infinite set of ordered pairs? (He considers only laws with two variables.) 6 The most that can be said is that the law describes frameworks of possibility for the determinates. From the three-dimensional relational space spanned by the three variables the law cuts out a surface which can be called the possibility-space. This space is undeniably a consequence of the law, but we ought not to reduce the law to this space if we think there is at least something to the idea of natural laws. Rather, there exist relations (nomic or factual) between grounded relations, and numerical laws describe such third-order universals.

12.2 E F F I C I E N T CAUSALITY AND EXISTENTIAL DEPENDENCE The analysis above was an analysis of a non-causal law and the kind of connection which such a law expresses. The relations we found were universals of third order. I shall now analyse a causal law by peeling away everything in such a law which can also be found in non-causal laws. The efficient causality will be what is left over. The law we shall look at is Newton's second law: force is equal to mass times acceleration. 7 Here the natural interpretation is that force necessarily brings about an acceleration, i.e. force is the cause, or the causal relation, and acceleration the effect. But this law, too, contains quantitative concepts; and if one puts it in the form: F = ma, the impression of causality easily disappears. The similarity to the non-causal gas law is striking. In both cases we have three universals which are momentarily connected in a 177

ONTOLOGICAL INVESTIGATIONS

mutuality relation. T h e gas law expresses explicitly a relation between grounded relations, but what does Newton's law explicitly say? T h e variable 'от' stands for the quantitative concept 'mass'. This in turn involves grounded relations which have their ultimate basis in the specific determinates of the determinable 'mass'. These determinates are, like the determinates making up the gas law, genuine properties not tendencies. T h e variable 'β' refers to the quantitative concept 'acceleration' which involves grounded relations which have their ultimate basis in the specific determinates of the determinable acceleration. These determinates are (in contrast to those contained in the gas law) tendencies, not genuine properties - but this difference does not affect the possibility of quantification. T h e variable 'F' is ultimately based neither on determinates of genuine properties nor on tendencies. I n contrast to the gas law, Newton's second law does not only have to do with properties which inhere in things, but also with relations between things. Mass and acceleration are properties which inhere in a thing. Force must therefore be a relation if the law is to be concerned with how one thing affects another. Here I must make a short digression to point out a complication in my ontological system. J u s t as the existence of certain properties implies the existence of grounded relations, so too the existence of some external relations implies the existence of grounded relations. If, in the definition of grounded relations (cf. D8.3 on page 120), one exchanges the two-place relation for a four-place one and makes the relevant modifications, then one obtains a definition of grounded relations which are grounded in relations. T h e relation 'x and у are at a certain distance from one another' is a two-place external relation; the relation 'the distance between χ and у is greater than that between ζ and v' is a grounded four-place relation. If the distance between the things a and b is greater than the distance between the things с and d, and if the things e and f have the same distance between them as do a and b, while the things g and h have the same distance between them as с and d, then the distance between e and f must necessarily be greater than the distance between g and h. This conforms exactly to the definition of grounded relation. T h e relation — ' and — have a greater distance between them than - and - ' is a grounded relation. T h e example above also makes it clear that one can construe quantitative concepts based on grounded relations which are 178

EFFICIENT CAUSALITY

grounded on external relations. T h e quantitative distance concept is of course the best example — it figures in a number of laws. This means that we also can say the following: T h e variable 'F refers to the quantitative concept 'force' which involves grounded four-place relations which have their ultimate bases in the specific determinates of the relation/determinable force. It seems, then, that all essential differences between the ostensibly causal Newton's second law and the non-causal gas law have disappeared. It seems as if those who wanted to do away with the concept of cause and replace it with the concept of a functional relation were right. The fact that Newton's second law contains grounded relations which are grounded in relations instead of properties makes it just as little a causal law as the presence of the grounded relations 'distance fallen' and 'time of fall' in Galileo's law of falling bodies make this law causal. Galileo's law says that the distance fallen s = v0-t + V2gt2, where vD and g are constants and t the duration of fall. It is impossible for the duration of fall to bring about or cause the distance fallen. Galileo's law of falling bodies is, like the gas law, a non-causal law. T h e conclusion to be drawn from these remarks, is that the numerical law 'F = ma', like the gas law, describes a relation between grounded relations, i.e. it describes a third-order universal. And in this universal there is no trace of anything which resembles an efficient cause. If the category of efficient causality is in some way involved in Newton's second law, then it must have its place somewhere else than in the explicit numerical formulation. In analysing the gas law I said that every law, independently of what it explicitly says, has an ontological context, I maintained that the context of the gas law is primarily related to the determinables: (the substance) gas and (the properties) pressure-volume-temperature, and that it implicitly says that they stand in the relation of mutual existential dependence. We ought therefore to look at the determinai/« force, mass, and acceleration and their relations to one another. T h e original intuition was that force in Newton's second law brings about acceleration, and we shall now return to this thesis, understood as a thesis about certain determinables. If a category of efficient causality is at work here, then acceleration is the effect and force either the cause or the causal relation. Mass is neither effect (the mass can be assumed to exist before both the force and the 179

ONTOLOGICAL INVESTIGATIONS

acceleration occurred) nor cause (the mass can exist without any acceleration being brought about); mass is a property and it is therefore impossible for it to be the causal relation. O n the other hand, mass is a necessary condition for force, but a necessary condition is not a cause. Neither can force be a cause, even if one seems to be talking causally when one says that 'the force brings about the acceleration'; the associations these words have led to the expression 'the cause brings about the effect'. However, the force (determinable) must, as previously pointed out, be a relation if Newton's second law is at all to treat of how a thing can affect another thing. In order to avoid ambiguity, I shall in what follows often speak of 'the force relation' instead of 'force'. T h e first question we have to ask ourselves with regard to the force relation is what kind of relation it is. Is it a grounded or an external relation, or perhaps some kind of existential dependence? If the force relation is a grounded relation it ought (in a loose sense) to be derivable from properties of the cause and the effect (cf. page 120). In the case of a gravitational force the relation ought to be derivable from the masses of the things that affect one another; in electrical attraction (think here of Coulomb's law), the relation ought to be derivable from the electric charges of the things involved. T h e force-determinable ought to be derivable from the determinable mass in the first case and from the determinable electric charge in the other, in the same sense as the relation 'is larger t h a n ' is derivable from two volume-determinates. Obviously this is not possible, which means that the force relation (and by implication the causal relation) is not a grounded relation. T h e second possibility is that the causal relation is an external relation. If that is the case then it ought to be logically possible that, to take a concrete example, there are two electrically charged things between which there is a force relation, and also another pair of electrically charged things which are qualitatively identical with the first pair and are in similar conditions, but between which there is no force relation. I do not think this is possible and do not know of any proposed law within physics which has such a character. So, in spite of what philosophers have written about 'singular causality', I maintain that the force relation and by implication the causal relation cannot be external relations. W h e n it comes to the third and last possibility, that the force relation is a relation of existential dependence, things become more 180

EFFICIENT CAUSALITY

complicated. At the moment our aim is to describe the force relation itself, as an example of the causal relation. The causal relation relates moments of different things, and so does the force relation. But in what follows it will look as though I am not discussing a relation between a cause and its effect, but a relation between the causal relation and its effect·, i.e. the relation between the force relation and acceleration. At first, therefore, I have to explain why we are tempted to think in these terms and why it is of no importance for the analysis of efficient causality. Let us for a moment once again return to the gas law and its ontological context. The relation between the determinables gas pressure and gas temperature is a relation of mutual existential dependence. This fact, however, does not entail that a unity of pressure and temperature is an independent unity. Their relation of mutual dependence can in turn be dependent upon something else, and this is in fact the way things are. It is dependent upon the determinable gas volume. (The best way of phrasing this fact is to say that there is a relation of three-sided dependence; D9.3 can easily be extended to deal with many-sided dependence.) In a similar way the relation between an acceleration and that which causes the acceleration is dependent on (at least) a 'third something'. Let us look at mass and the force relation. Mass inheres in things, and it is logically possible for every mass to exist even if no force relation exists between things. There is here a relation of neither dependence, nor concrete dependence, nor individual dependence. Mass is totally independent (D9.10) of force. The force relation can, on the other hand, exist only between things with mass, which means that the force relation is existentially dependent upon the property mass. Together, this means that force is one-sidedly dependent upon mass and, since mass is not a part of force, that force is founded upon mass. This point can be generalized. Just as the force relation is founded upon mass, so every causal relation is founded upon properties of the things between which it holds. This fact explains why we have to talk of the relation between the causal relation (the force) and its effect (the acceleration), when actually we are investigating the causal relation itself. It is a way of taking for granted that the conditions upon which the relation is founded are fulfilled. We can now turn to the relation between force and acceleration. 181

ONTOLOGICAL INVESTIGATIONS

First, it is logically impossible for the force relation of Newtonian mechanics to exist if accelerations do not exist. Force is existentially dependent upon acceleration. But the opposite does not hold. It is logically possible for accelerations to exist without any force relations existing. The determinable concept of acceleration (which does not need the infinitesimal calculus) was also worked out before Newton and the Newtonian concept of force. Acceleration is not dependent upon force, and neither is it concretely dependent. There is no 'variation' in the determinable acceleration which makes acceleration dependent upon force in the way a determinable is dependent upon its determinates. Like mass, acceleration is generically independent (D9.9) of force. So far so good. But we have still not captured the obvious intuitive difference which, in Newtonian mechanics, exists between the force-mass relation and the force-acceleration relation. Forces are said to cause accelerations but they are not said to cause masses. The difference, I suggest, is the difference between 'generic independence' and 'total independence' defined at the end of chapter 9 (D9.9 and D9.10 respectively). Masses are totally independent of forces whereas accelerations are merely generically independent of forces. In positive terms, acceleration is individually dependent upon force. In chapter 9 it turned out that it is at least formally possible to define a very curious kind of existential dependence, a dependence which concerns what is particular rather than what is universal. I shall repeat the definition: D9.8: a is individually dependent upon В if and only if it is logically impossible for this instance of Л to exist if some instance(s) of В do(es) not also exist. The important thing here is that the modal operator 'logically impossible' governs 'this'. Ordinary generic dependence of Л upon В implies in contradistinction that 'for this instance it is logically impossible . . .'. In generic dependence the dependence concerns the universal aspect of states of affairs only. 'This' does not fall within the scope of the modal operator. What is formally possible may not be materially possible. Can a modal operator really work on something non-universal? It seems almost to be a common-sense assumption in many different philosophical traditions that necessities have to be necessities 182

EFFICIENT CAUSALITY

between what is universal. 8 In spite of this, however, I claim that individual dependence is materially possible. In order to understand this kind of dependence it is necessary to keep two aspects of my theory of categories in mind: firstly, all universale are universale in re, secondly, some relations lack mutuality, i.e. A can be related to В without В being related to A (cf. pages 130-1). I shall bring out the importance of the latter fact first. That an instance of acceleration is, as I maintain, individually dependent upon force does not imply anything at all about how force is related to acceleration. As a matter of fact force is generically dependent upon acceleration. This means that in one direction there is individual dependence and generic independence, and in the other direction there is generic dependence. Compare the determinable-determinate relationship analysed above (pages 142-3): in one direction the relation between determinate and determinable is generic dependence, in the other direction the relation is concrete dependence. Such juxtapositions seem natural enough once it is realized that relations of existential dependence can lack mutuality. And our results afford us a long-awaited tool with which to harmonize two conflicting requirements on any theory of efficient causality. On the one hand, we have the Humean intuition that cause and effect are generically independent and conceptually distinct, on the other hand we have the Spinozist and Hegelian intuition that causality has to be some kind of logical or internal relation. As long as all relations are assumed to be mutual, these intuitions are contradictory. But my analysis allows me to regard the Humean intuition as the right one when we consider the relation between acceleration (effect) and cause, and the other intuition as the right one when we go from force (the causal relation) to the effect. Individual dependence and generic one-way independence is consistent with generic dependence in the other direction, but what is individual dependence? When i4's generic dependence upon В is defined by saying that it is logically impossible for A to exist if В does not also exist, the definition seems to be Platonic, although the intention is Aristotelian. All generic dependence is dependence in re. The dependence exists in the relevant particulars although, so to say, the dependence does not depend on their particularity. When it comes to individual dependence, the first thing to be noted is that though the modal operator governs the 'this-operator' it 183

ONTOLOGICAL INVESTIGATIONS

does not work on 'this' alone. It works on 'this instance of A', which means that the logical impossibility at hand is not concerned with mere particularity, but with a universal (A) in its particularity. Within immanent realism mere particularity is an impossibility. All particularity is particularities of universale. Mere particularity belongs to nominalist or 'particularist' ontologies. T o speak of mere particularity within immanent realism is to turn particularity into a universal, which cannot be done. T h e upshot of this is that individual dependence is not to be conflated with the impossible, what might be called 'particular dependence'. Within both nominalism and Platonic realism there are only two possibilities to be discussed, the existence of generic necessities and of particular necessities; and the latter are usually and rightly declared to be impossible. Within immanent realism particular necessities are likewise a n impossibility, and generic necessities are retained with the modification that the necessities exist in re. But immanent realism makes a third alternative possible: necessities which relate to the particularity of a universal. When a is individually dependent upon B, it is the particularity of A which is dependent upon B. Let us now, armed with the concept of 'individual dependence', take a new look at our intuitions about causality. Let us think of a piece of iron which is caused to accelerate by a magnetic field. T h e field produces the acceleration. Is not the idea here, that it is primarily the force-field that produces a new particularity, a new particular acceleration, rather than a universal (acceleration) which, as a second requirement, must also be instantiated? A lot of everyday examples can be adduced in favour of such an idea. T h e phone stands on my table. T h e table makes my phone stand where it stands; the table causes the phone to retain its place. This cannot be a matter of place-of-phoneness depending on tableness. It is this phone depending on this table; it is only the particular phone which depends on the particular table. T h e 'generic phone' is independent. Epistemologically, (onto-) logical impossibility in the sense here used often goes with inconceivability. Mutual generic dependence between colour and extension goes together with the inconceivability of a colourless extended surface. In the case of the accelerating piece of iron something similar occurs. It is possible to conceive of instances of acceleration without conceiving of any forces, but it is impossible to conceive that this instance of acceleration in the 184

EFFICIENT CAUSALITY

magnetic field would remain instantiated if the force-field disappeared. The concept of 'individual dependence' is a coherent notion. I have taken the relation between force and acceleration to exemplify the causal relation. The analysis of efficient causality given so far can be summarized as follows. For every relation of efficient causality: (1) the cause is generically dependent upon the effect; (2) the effect is generically independent but individually dependent upon the cause, (3) the relation constituted by the dependence relations above is founded upon qualities of the things which stand in the causal relation. 12.3 CAUSALITY AND UNIVERSALITY We have now seen what the causal relation in itself consists of; it is a juxtaposition of different kinds of existential dependence. But there are more problems to be solved. What are the relata of the causal relation? The analysis given above conforms to the requirement that causality shall involve universality or 'the invariability of the connection'. If a causal relation exists, then its effect has to exist. This requirement is in my analysis entailed by the generic dependence of a cause on its effect. So, this view, which has a long past, is not confined to empiricist philosophers. But also the opposed view, i.e. that efficient causes do not involve universality, has been historically popular. Both Aristotle and John Stuart Mill, to take two very different philosophers, presuppose that causes can work against each other. One cause can cancel out the effect of another, which means that a cause can exist without its effect. Universality of connection has disappeared. In the debate concerning Karl Popper's falsificationism, views have also been advanced which imply that causality cannot unproblematically imply universality. Lakatos writes: Similarly, 'all swans are white', if true, would be a mere curiosity unless it asserted that swanness causes whiteness. But then a black swan would not refute this proposition, since it may only indicate other causes operating simultaneously. Thus 'all swans 185

ONTOLOGICAL INVESTIGATIONS

are white' is either an oddity and easily disprovable or a scientific proposition with a ceteris paribus clause and therefore undisprovable. 9 At first glance, it seems that Lakatos must deny that causal laws imply universality. Lakatos did not himself try to clarify this issue. If the tendency category is taken into account, it is, however, easy to understand both why many conceive of causality and 'invariability' as being thoroughly connected, and why others want to deny such a connection. One finds the solution when one notes that causal laws always have an explicit or implicit reference to a superposition principle. Newton's second law can be analysed in the following way: (1) (2) (3) (4)

FT = m ar F p = m-ap FT = the vector sum of Fp a r = the vector sum of ap

All four of the above law relations describe universal connections. T h a t which is not universal, and which is therefore not included as a law, is that every acceleration (a) becomes realized in itself. Some become so, and some not: at always becomes realized as such, ap never (cf. pages 163-5). It is this fact which makes accelerations belong to the tendency category. If one is clear about this and accepts that causal effects are tendencies, then one understands how the concept of causality includes both universality and nonuniversality. A cause always produces its effect, which is a tendency, but this tendency as such is not always realized. 10 In what follows I shall assume that the category of efficient causality can be identified with the help of Newton's mechanics, especially Newton's second law. Therefore, I shall round this section off by giving some more reasons for relying so heavily on the law mentioned. Most physical laws have a phenomenological or instrumental character, but their context gives them an implicit reference to deeper-lying real laws. Look at the law 'heating causes metals to increase in length'. In its mathematical formulation, 'L(t) = L(0) · (1/ + xt)\ it is a non-causal law which momentarily connects temperature (/) and length (L) of one and the same body. When one hears the causal formulation, one presumably thinks partly of a thing (e.g. a burner) or a process (e.g. a rise of temperature in the air), which is distinct from the piece of metal and which creaces the heating, and partly of the piece of metal. In this case one has 186

EFFICIENT CAUSALITY

both a relation between two things and a tendency (to increase in length) in one of them. The problem is only that it is difficult for us to understand this relation as a relation which is not in need of further clarification. The advantage of Newton's mechanics is that it allows one, so to speak, to take relations at their face value as they occur in the theory. When two bodies affect one another according to Newton's second law, we can in a simple way distinguish that body which affects (the cause), that body which is affected (the effect being acceleration), and the efficient causality relation itself (the force relation). 12.4 A C T I O N BY CONTACT, A C T I O N A T A DISTANCE, AND ACTION BY M I X T U R E If the analysis of efficient causality were to stop here, the problem of contiguity would not have been treated at all. This problem is in fact two problems: 'Do cause and effect necessarily stand in the relation of temporal contiguity?', and 'Are cause and effect necessarily in spatial contact (spatial contiguity)?'. If one relies on Newton's second law one gets an answer to the first problem but not to the second. The force relation holds momentarily between two things, i.e. it is a question of neither contiguity in time nor action at time distance, but one of simultaneity. However, the law says nothing about whether forces relate spatially contiguous things (action by contact), nor whether it is a question of action at a spatial distance. This silence on the part of Newtonian mechanics has constituted the background for the centuries-long debate on whether action at a distance is possible or not. If one allows causality to be reduced to constant conjunction the problem becomes simple. Nothing prevents there being constant conjunctions not only between phenomena which are at a distance from one another, but also between phenomena next to one another. And this reduction has allowed many philosophers to remove the problem from the philosophical agenda. But here, where causality is not constant conjunction, we must take up the problem: Can one thing at a spatial distance from another thing produce a change in the latter? In order to understand 'production' we have to go back to the idea of individual dependence involved in 'efficient causality'. Generic dependence is both timeless and spaceless, as it is a relation between universale as universale, and universals as such are timeless and spaceless. Therefore, generic dependence makes no 187

ONTOLOGICAL INVESTIGATIONS difference between action by contact a n d action at a distance. But individual dependence concerns universale in their particularity, a n d particulars are situated in time a n d space. I n the example I used when introducing the concept of 'individual dependence', there was no question of action a t a distance. A n d I can merely confess that I a m u n a b l e to see how there could be individual dependence between particulars at a distance f r o m each other. I regard action at a distance as an impossibility, a n d , like Newton himself, I regard his law of gravitation as a phenomenological law. T h e next step, then, is to analyse situations w h e r e things exist which are spatially contiguous a n d between which a force relation holds. W e shall proceed via the analyses of Descartes a n d Boscovich. First, Descartes. 1 1 If two things lie next to one another, how can one of them affect the other? T h e characteristic property of a thing is to occupy space, i.e. to have volume. If one thing moves away f r o m another, no effect takes place, since the other can remain in its place. But if the one thing moves toward the other, something m u s t happen since in t h a t case the two things come to compete for possession of the same spatial volume. T h e first thing will collide with the other a n d thereby produce a change of motion. Material things can have only one form of action by contact, impact. Everything which is usually described in terms of 'pull', 'lift', or 'suck' must actually be a form of impact or push. W h e n we pull a wagon, part of us (our h a n d ) is behind a part of the wagon (the handle), which m e a n s t h a t the h a n d actually collides with or pushes the wagon forward; to push is the same as to collide continuously. W h e n we lift a bucket a p a r t of us (our h a n d ) is under a part of the bucket (the handle), which m e a n s that the h a n d pushes the bucket in an upward direction. W h e n a p u m p sucks u p water, the air actually pushes the water up. W i t h the help of the p u m p we move things which otherwise resist this pushing. M u c h is to be said for the view that i m p a c t is the only kind of efficient causality which makes it comprehensible how one thing can affect another. I m p a c t is a consequence of a thing's property of occupying a space, while other kinds of force (e.g. some form of indefinable cohesion capacity as an explanation of how one thing can pull another) a p p e a r s completely arbitrary and without basis in those things between which it operates. T h e latter kinds of force thereby come to a p p e a r to be as incomprehensible as action at a 188

EFFICIENT CAUSALITY

distance. The impact relation can be subsumed under Newton's second law, and it seems almost certain that efficient causality should be defined with the help of spatial contiguity. Roger Joseph Boscovich, however, showed that appearance here are deceptive. Boscovich's argument against the possibility of impact — now more than 200 years old 12 — has as far as I know not been refuted. I shall give a summary of it. Assume that we have two equally large bodies of the same mass. They move in the same direction on the same line. The first body has the velocity 6 m/s, the other the velocity 12 m/s. When they collide, the velocity of the first body increases, while the latter's decreases. Because of reasons of symmetry, as well as from Newton's mechanics, it follows that both bodies after the collision have the velocity 9 m/s. The question is what happens in the impact itself. What happens before and after is unproblematic. At the moment the bodies meet they have velocities of 6 and 12 m/s, respectively. There is thus a temporal interval during which the first body increases its velocity from 6 to 7 m/s, while the other decreases its velocity from 12 to 11 m/s. In this temporal interval the first has the average velocity 6.5 m/s and the other 11.5 m/s. But this means that the second body moves into and partly takes up the same space as the first, which goes against the premise that material things always take up different spatial volumes. The concept of impact contains a contradiction. It is thus impossible for impacts to exist! Boscovich's argument does not presuppose, as one all too easily assumes, that bodies are rigid, but it needs the following two premises: (1) change of motion is always a continuous change (if the things could momentarily change velocity from 6 to 9 and 12 to 9 m/s respectively, of course, no problem would occur), and (2) that things cannot penetrate each other. Things which cannot penetrate one another but are not rigid must be elastic. One can understand elasticity either as a primitive property or as an 'atomistically defined' property. In the latter case, an elastic thing is actually an aggregate of rigid things with space between them. It is these spaces which disappear when the thing in question is compressed. Such an elasticity concept does not, however, solve the difficulty pointed out by Boscovich. Somewhere two rigid bodies - atoms must meet; and then the problem reappears. If elasticity is a primitive property the ultimate thing in itself can be compressed. 189

ONTOLOGICAL INVESTIGATIONS

This can solve the problem with regard to those sides of the things which do not make contact in the collision, but not with regard to the sides which actually collide. They are two material surfaces which collide, and for them the problem remains unchanged. I should like to call this problem 'Boscovich's paradox'. The similarity to Zeno's paradoxes is striking. The arguments are, from a logical point of view, both simple and clear. Moreover, they seem, on first presentation, to be cogent, but the conclusion is in each case astonishing. In the case of Zeno, that motion is impossible - in Boscovich's case, that collisions are impossible. I have commented on Zeno's paradoxes in the introductory chapter, and have taken them as examples of the fact that one sometimes has to 'teach oneself to think what one sees'. One can be deceived by logical argument. In Boscovich's case I would also have thought this to be the case, were it not for the development of modern physics. I shall maintain that Boscovich's argument is valid, but not that his own proposed solution is the right one. Boscovich assumed that there was either action by direct contact via collision or that there was action at a distance; and since the impact relation appeared in itself self-contradictory, action at a distance had to exist. He did not think himself able to question the premise that motions are continuous, and I do not believe I can either. But there is yet another premise to be discussed, one which Boscovich himself never considered, namely the tacit presupposition that two things cannot occupy the same place at the same time. By 'things' I mean things in the sense of the category thing; i.e. not only bodies or corpuscles. Penetrable things did not exist according to the physics of the eighteenth century, but they exist according to today's physics. The different sorts of fields which today's physics treats fall under my category of thing as a state of affairs (fields consist of both a substance and related properties) but there can exist many such things (fields) simultaneously at one and the same place. Certain fields can even exist at the same place as a traditional corpuscle. This line of thought can be exemplified with the help of the bestknown field, the electromagnetic one, and with a very everyday phenomenon. We have all learnt that where we see empty space there is not only air, but also a lot of different radio and T V programmes each sustained by its own electromagnetic field. All fields exist in our antennae. We do not have to change the antenna 190

EFFICIENT CAUSALITY

to change the programme. Fields do not melt together in spite of their being at one and the same place at one and the same time. Things (and other kinds of states of affairs) which belong to different ontological levels can of course exist at one and the same place simultaneously. This is a direct consequence of the concept of irreductive materialism. But now we are interested in whether things on one and the same level can occupy the same place at the same time. The category of thing, given a certain level, breaks up into two subcategories: things which cannot be at the same place at the same time, and things which can. I shall call the former 'corpuscles' and the latter 'penetrables1. In physics this distinction appears as a distinction between phenomena with and without rest mass. Most elementary particles have, but fields and photons lack, rest mass. Penetrables do not, according to the definition, collide; and neither do light rays or electromagnetic radiation in general. One of Huygens' reasons for proposing the wave theory of light was that waves do not collide whereas corpuscles do. 13 Huygens, however, thought that light was vibrations or pulsations in the ether. Now, since we do not believe in the ether we must say that since light rays do not collide they have to consist of penetrables. Two corpuscles cannot occupy the same place at the same time, but two penetrables can. Corpuscles can thus be defined by reference to their ability to repel one another away and to preserve their location from occupation. Identity may be given by position in space. Penetrables, however, behave differently. Their identity cannot be directly grounded in space. Not only can penetrables exist at the same place as other penetrables, they can exist at the same place as corpuscles. Their identity can thus reveal itself via their effects on corpuscles. These effects are tendencies, and various tendencies can exist at the same time and place. Thus they can be added according to the superposition principle. Without the tendency category, it would seem to be impossible to capture the identity of penetrables. The solution to Boscovich's paradox can now be formulated in two different ways: (1) In the case of efficient causality the cause and its effect must exist at the same place in both space and time; (2) all efficient causality presupposes penetrables. Efficient causality involves neither action at a distance nor action by direct contact, but rather 'action by mixture*. When the force 191

ONTOLOGICAL INVESTIGATIONS

relation holds between two things at the same place, one of the things can actually produce an effect. Here one can also note that the requirement that cause and effect must exist at the same place is naturally fulfilled by the category causa sui. Wherever we have causa sui there is real production of an effect. The cause and its effect need only be at the same place when the causal relation between them exists. If by 'cause' one means the actual thing in itself which in a certain situation is the cause, then of course that thing can exist independently of its effect both before the effect and at another place than the effect. In this sense causes also precede their effects in the case of efficient causality. But when a causal relation exists, the cause is both at the same time and same place as the effect. The cause and the effect make up a mixture}* Action by mixture, not action by contact or action at a distance, is what we find in the real world. It is now time to summarize the analysis of efficient causality. In order for a relation of efficient causality to obtain, five conditions must be met: (1) The cause is generically dependent upon the effect. (2) The effect is generically independent but individually dependent upon the cause. (3) The (mutual) relation constituted by (i) and (ii) together, is founded upon qualities of the things which are related. (4) The effect is a tendency. (5) The cause and the effect coincide in space and time. From the point of view of common sense, the last requirement is astonishing. The most common example of efficient causality has, ever since ancient times, always been that of a person using an instrument to chop, cut, or hit a material object, i.e. action by contact via some kind of impact. On the other side, it seems today equally undeniable that our ability to see is so defective that we cannot see in any detail what actually happens in an impact. In the final chapter I shall discuss the relation of this theory of categories to empirical data and to the necessities of thought. Here I shall simply point out that it is Boscovich's argument plus the development of modern physics plus what has just been said about our ability to see, which leads to our strange fifth requirement. It takes a little time to accept the idea that impact is not a primitive example of efficient causality. But as we shall now see this opens 192

EFFICIENT CAUSALITY

up new perspectives on space. It becomes possible to consider space as something truly generative. 12.5 SPACE AS E F F I C I E N T CAUSE If one holds that paradigmatic cases of generative causality must be cases where one corpuscle affects another corpuscle, one is almost automatically led to the view that space must be a passive container. Efficient causality then comes ultimately to base itself on the impossibility of two things being in the same place at the same time. Even if one equates container space with things in the categorial sense (both things and container space are 'particulars'), space is unable to figure in a relation of efficient causality, since other things, of course, can exist at the same place as space. Space cannot possibly have the ability to keep corpuscles away from itself since it is defined via the contrary ability, that of being able to contain things. In Newtonian mechanics space is a passive container, but not in Aristotelian physics nor in the general theory of relativity. This fact alone creates problems for an ontological system which does not allow container space to generate changes. But there are stronger reasons for allowing container space to act as a cause in a relation of efficient causality. Using an example from Neriich, 15 I shall make it clear that space in Newtonian mechanics can be seen as causally passive only if one presupposes that the geometry of space is Euclidean. If, on the other hand, one assumes that it is nonEuclidean, one is forced to ascribe a causal ability to space, even if this conflicts with the corpuscularian idea that such a capacity always is rooted in the more basic ability to fll a part of space. It is only by bringing in the concept of action by mixture that we can see how space can be an efficient cause. When Newton presented his mechanics, and for a long time afterwards, the law of inertia appeared unequivocal. It says that a thing which is not affected by a force remains in rest or continues in a straight line with unchanged velocity. With the discovery of the non-Euclidean geometries, however, an ambiguity appeared in the concept 'straight line'. If the straight line between two points is defined as the shortest line between them, one cannot decide what 'in a straight line' is without first having chosen a geometry. Let us imagine a two-dimensional container space - the extrapolation to 193

ONTOLOGICAL INVESTIGATIONS

three dimensions we will leave to the reader. We shall compare the geometry of a flat surface (i.e. a two-dimensional Euclidean space) with the geometry of the surface of a sphere (i.e. a two-dimensional non-Euclidean space with constant curvature), and see what consequences the law of inertia has in the respective geometries. Two point-formed things which move in a straight line in the same direction in Euclidean space, and which are not affected by any forces, will never be able to meet. They move along two parallel lines. But if the things move in a straight line in the same direction in the spherical space, the law of inertia gives different consequences. The things come to move along two great circles on the sphere, and as a consequence sometimes approach one another and sometimes move away from one another, depending on where they are located on the great circles. They can even meet at the 'north' and 'south' poles without any forces operating on them. In a spherical geometry there are no parallel lines, and the shortest, i.e. straightest, line between two points goes along the great circle connecting the points. If instead of two point-formed things, we imagine one thing whose endpoints are located where we have just imagined the two things to be, the following happens. In the Euclidean space the thing continues straight ahead with unchanged velocity and nothing happens within the thing. But in the spherical space the space comes sometimes to create a tendency toward contraction of the thing, sometimes toward expansion. The tendency to contract is strongest nearest to the 'poles', and the tendency to expand is strongest nearest the 'equator'. It is a question of tendencies, since there can exist counteracting factors. Spaces in both Einstein's and Newton's physics are homogeneous in the sense that they 'treat' all things in the same way. But this is not a necessary feature of a space. Let us look a little at space in Aristotle's physics. Aristotle's ontology is a level-ontology of which we shall now only consider the lowest level. It appears as follows. Space is a finite three-dimensional sphere with a Euclidean geometry. There are (beneath the moon) four different kinds of elements: earth, water, air, and fire. Every element has a natural place as a tendency — when it is not at its place it tends toward it; when it is at its place it is at rest. An element which is not affected by external forces moves toward its natural place. Earth has its natural place at the centre of the sphere, water on a sphere concentric with the centre, air on an even larger concentric sphere, 194

EFFICIENT CAUSALITY

and fire has its natural place on the inside of the sphere of the sublunar space itself. Aristotelian container space explains tendencies in approximately the same way as each three-dimensional container space explains the enantiomorphism of 'right-' and 'left-handedness' (see section 10.3). It is not space in itself which provides the explanation, but space together with something belonging to the thing in question. It is this which creates the possibility that a space be heterogeneous like Aristotelian space, i.e. that it 'treats' different kinds of things differently. In the two examples reproduced above, container space generates changes in corpuscles. This efficient causality comes about, as does all such causality, via properties — container space also has properties. In the case of efficient causality between things, it is mainly a question both of effect and countereffect. When a thing affects another with a certain force, the other affects the first with a certain force. Nothing in this exposition excludes the case where such an interaction also occurs between space and things. And it would be unfortunate were this case to be excluded, for the general theory of relativity considers just such causal relations. If that theory is correct, space has a structure which is dependent on the mass-energy distribution in it, and vice versa. It should perhaps be pointed out that such interaction in no way affects the view presented earlier, that things are one-sidedly existentially dependent on container space, i.e. things cannot exist without space, though space can exist without things. The equations which constitute the general theory of relativity can also specify which space-time structure is possible when space is empty, i.e. when there is no mass (no things) in space. 16

195

13

INTENTIONALITY

Intentionality is the category without which people and society would have to be regarded as mere machines and natural processes. Intentional phenomena have a direction toward something outside of themselves. A statement is intentional, for it refers to something outside itself. A wish is a non-linguistic intentional phenomenon, for a wish points forward to a possible fulfilment lying outside itself. In the present chapter the category of intentionality will be presented and related to the other categories. Peculiarities in the whole-part structures which are exhibited by systems with intentionality will be dealt with in chapter 15. The ontological system I am presenting is Aristotelian in its orientation, but for Aristotle himself the two categories tendency and intentionality are merged in an unfortunate way in teleological explanations. Both the concept of'tendency' and that of'intentionality' have had something of a renaissance in Anglo-American philosophy during the last decade — unfortunately, however, each on its own. 1 No one has shown the connection between the two concepts. But this is necessary if one wants to rehabilitate the two categories without confusing tendencies and teleological explanations. Both tendencies and intentionality are forms of 'directedness', but there are teleological explanations only where there is intentionality.

13.1 I N T R O D U C T I O N T O I N T E N T I O N A L I T Y The modern philosophical concept of'intentionality' was introduced by Franz Brentano. 2 He employed the concept, among other things, in order to delimit the psychical; he saw intentionality as a 196

INTENTIONALITY

necessary and sufficient condition for something's being psychological. In the tradition in which the concept of intentionality has been most prominent - phenomenology - intentionality has generally been regarded not as a necessary, but as a sufficient condition for something's being psychological. There exist, however, those who have tried to introduce a non-psychological variety of intentionality, most notably Maurice Merleau-Ponty. In this, Merleau-Ponty represents a kind of return to Aristotle. In Aristotle there was no clear dichotomy between the material and the psychical. Examples of intentional phenomena are the following: linguistic phenomena (statements, questions, exhortations, etc), seen pictures, perceptions, memories, beliefs (as well as fantasies, thoughts, wishes, etc.), certain sensory states (like being angry at, happy about, sad about, and so on), and intentions. Real intentionality is bipolar. The 'directedness' which characterizes it is a directedness from something to something. Objects or things lack this kind of directedness. They never 'get beyond' the space-time they occupy. But subjects can, thanks to their intentionality, 'point' beyond themselves in space and time. There are two kinds of fundamental states of affairs in this theory of categories: subjects and objects (the terms 'object' and 'thing' are synonymous terms in this book). Subjects are states of affairs which have (normally, intermittently) intentionality. Objects (or things) are states of affairs totally lacking intentionality. Intentionality is a special form of connection between a subject and objects or between subjects and subjects. As can easily be understood, and as will later be shown in greater detail, this peculiar kind of connection is the basis for completely different types of part-whole relations than those exhibited by the other categories. (The intentional subject-object connection is described in sections 13.5-7; the intentional subject-subject connection is described in chapter 15: nested intentionality.) In the phenomenological tradition, that which an intentional act is directed at is sometimes called the 'intentional object' sometimes the 'intentional correlate'. For many reasons the former phrase is a bad one, and I shall only speak of the intentional correlate. Developments within the phenomenological tradition itself have made the 'object' terminology I am opposing obsolete. But if we 197

ONTOLOGICAL INVESTIGATIONS

stick to the type of intentionality which thinking and language exemplify, then it may be justified to speak of intentional objects. W e often think and speak precisely of objects in the sense of spatially isolated things, or moments of these. But if we turn to that type of intentionality which perceptions constitute, this terminology is clearly misleading. We perceive, as so many phenomenologists have now told us, not only things and other states of afTairs, b u t things and states of affairs in a world. Things, and states of affairs in general, do not only appear against a background, they have definite relations to this background and constitute with it a quite different sort of structured whole than that formed by things and states of affairs alone. T h e things and the other states of affairs we perceive always appear as parts of a situation, and it is a situation extended in both space and time. T h e intentionality which occurs in the normal perception of an everyday situation is not only the pointing of a subject toward a simultaneously existing state of affairs in a momentary situation, but is also a pointing to both a past and a future. When we perceive an ordinary material object, we see it as something that has existed for some time and which will continue to exist at least a few moments longer. There is a kind of pointing both toward the thing's past and towards its future. This feature manifests itself even more clearly in the perception of objects such as tools and furniture. These are seen as things which have been used in a certain way and which in the future can be used in the same way. And this pointing toward the past and the future makes the term 'intentional object' unsuitable as an umbrella term. Thus I shall use the term 'intentional correlate'. W e can now pose the general qustion: W h a t types of entities can figure as correlates of intentional acts? Part of the answer may of course be found in the examples already given. Things and relations in container space have been mentioned, as well as their moments, but it should be pointed out that not all intentional correlates need be situated in space and time. T h e intentional correlate can be a universal. When, for example, we think of the category of property in itself, or some specific property subsumed under the category, we can think of it without thinking of it (as) being positioned in space and time. T h a t an intentional correlate can be independent of space and time does not of course mean that the subject which has intentionality can exist outside space and 198

INTENTIONALITY

time. This is impossible. Every intentional act must exist in space and time, but not every intentional correlate. If we keep to the intentional phenomena whose correlates can have a spatio-temporal position, we can make an observation which is seldom accorded the importance it deserves: intentional pointing is normally a pointing over or across a spatial and/or temporal distance. The subject is spatially and/or temporally separated from the intentional correlate. Intentionality is normally 'intentionality at a distance'. When we think of something other than ourselves, remember something or are angry at someone, as subjects we are clearly spatially and/or temporally distinguished from the intentional correlate. 3 In the case of perception the relations are more complicated, due to unclarity regarding whether the intentional correlate is a spatial part of the actual intentional act or not. I shall later return both to this problem and to the meaning of the expression 'intentionality at a distance'. But for present purposes it is enough to point out that in every perception there is a pointing by a subject over a spatial distance toward that which is perceived, whether or not the latter should be considered, in some deeper sense, as a part of the perceiving subject.

13.2 REAL AND F I C T I O N A L I N T E N T I O N A L I T Y Intentional phenomena point at something, their intentional correlates, but they do not always reach what they point at. When a statement reaches or 'hits' its correlate, it is true. When it does not reach its correlate it is false; in this case there is no correlate to reach. Just as one can distinguish between true and false statements, one can also distinguish between realized and unrealized intentions, correct and mistaken memories, obeyed and unobeyed commands, correct and incorrect pictures, fulfilled and unfulfilled wishes, correct and incorrect perceptions. In these cases of intentionality one can ask oneself whether that which is pointed at is hit or not, whether the acts are satisfied or not. Such intentional acts have what I shall call a modality of satisfaction.' 4 The satisfaction modality, however, does not only have the two values satisfied and unsatisfied. Wishes, for example, are often neither completely fulfilled nor unfulfilled, but are rather partly fulfilled. In the case of wishes the satisfaction modality can assume the value partially satisfied. This possibility also occurs, as can be 199

ONTOLOGICAL INVESTIGATIONS

easily seen, in many other types of intentionality. Most misperceptions are, for example, not completely mistaken, but rather reach parts of their goals. T h e situation — much discussed in the empiricist tradition — in which a straight oar appears as a broken oar in water, is an example of a partially satisfied perception. Part of the intentional correlate is reached (the facts that there exists an oar and that there exists water) while part of the intentional correlate is not reached (the fact that the oar is broken). T h e same possibility of partial satisfaction also occurs in the case of statements — as has been shown by, among other things, the discussion of 'verisimilitude'. T h e old bipolar concept of truth should be given up in favour of a truth concept which allows grading; I shall return to just this case of partial satisfaction in section 16.1. I n order fully to understand the category of intentionality it is important not only to discern the satisfaction modality, but also to realize that some intentional phenomena lack this modality. Stories about invented characters are obvious examples. When one reads such a story one has intentional acts, but there can be no question about truth and falsity in the ordinary sense. Intentional acts which do have a modality of satisfaction, on the other hand, are directed at entities 'presumably existing somewhere in space and time. This world-directed kind of intentionality I shall call real intentionality. Its opposite I shall label fictional intentionality. T h e latter term, of course, reflects the fact that fictional discourse is a kind of fictional intentionality. But not the only kind. Lively fantasies which are understood as fantasies exemplify fictional intentionality of a perceptual kind. Both stories and fantasies are characterized by the fact that the question whether they reach any correlate is a meaningless question. Fictional intentional phenomena can neither hit nor miss any correlate since they lack the satisfaction modality. In the case of a real intentional phenomenon, like an ordinary descriptive statement, it is often impossible to decide whether it is satisfied or not, i.e. whether the statement is true or false. But this is completely in keeping with these statements having a satisfaction modality. Fictional intentionality, on the other hand, as just pointed out, lacks this whole modality. Fictional intentionality points without pointing at anything. There are many problems to solve with regard to fictional 200

INTENTIONALITY

intentionality. Some concern the nature of fictional intentionality's relation to classical phenomenological categories like 'mere presentations' and 'non-positing acts' and some concern the by now very large discussion about fictional discourse. 5 All those problems, however, will, except for some remarks in the next section, be ignored in this book. For the purpose at hand, the category of fictional intentionality is necessary only as a foil for real intentionality. 13.3 PRESENTATIONAL AND REPRESENTATIONAL INTENTIONALITY All types of non-fictional intentional acts have a satisfaction modality which can assume different values. The two values satisfied and unsatisfied must be carefully distinguished from two of the subcategories of real intentionality: presentational and representational intentionality. 6 The distinction between the two values of the modality of satisfaction is based on the parallel between statements which are true and false, perceptions which are correct and incorrect, and intentions which are realized and unrealized, while the distinction between these two new subcategories is based on the special position which perceptions have as compared to such things as thoughts, uses of language, pictures, and memories. Let us, in order to understand these subcategories of real intentionality, compare a statement and a perception. A statement and a perception can have the same intentional correlate. Compare, for example, the situation where a person A stands and watches another person С occupied in digging a trench with the simultaneous situation in which person В at another place states that С is occupied in digging a trench. B's intentionality has - at a certain level of generality - the same intentional correlate as •4's; and the satisfaction modality has in both cases the value: satisfied. The radical difference between the two intentional acts is that in the one case the intentional correlate is obviously located in space outside the act (B's utterance or statement), while in the other case it seems in some way to be a part of the intentional act (Л'б perception). The difference can also be expressed by saying that an utterance or statement is in no way itself part of the evidence for the possible fact that it is satisfied, while in the case of normal perceptions the perception itself constitutes part of the evidence for 201

ONTOLOGICAL INVESTIGATIONS the claim that it is satisfied. I n a n utterance the intentional correlate is represented·, in a perception it is presented. I n the above example involving С s digging, the satisfaction modality was given the value satisfaction in both cases. Let us now see how the e x a m p l e works if we let the modality a s s u m e the value non-satisfaction; t h a t is, we assume there is nothing corresponding to the utterance or perception of С s digging. A has a hallucination or experiences an illusion, and В makes a false statement. As regards the utterance, it is still obvious that the statement contained in the utterance points to a state of affairs outside itself; even though in this case the state of affairs in question does not obtain. Every s t a t e m e n t represents a s t a t e of affairs outside itself as obtaining or as not obtaining, whether or not this state of affairs does obtain or not. As regards the misperception t h a t С is digging, it m a y first be pointed out that the perception is qualitatively identical with the corresponding correct perception. Also, misperception, like correct perception, claims to have an intentional correlate as a part of itself. It is this that characterizes presentational intentionality. T h e c h a r a c t e r of presentational intentionality is, like representational intentionality, an inner property of the intentional act itself; and presentational intentionality can, like t h e representational variety have the values satisfied and unsatisfied. Presentational versus representational intentionality is not a distinction between intentionality which does and intentionality which does not reach its correlate. Presentational intentionality is characterized by its making a claim to directly present its intentional correlate. Representational intentionality is characterized by its making a claim not to present its intentional correlate directly; it claims to point to a correlate which is outside the intentional p h e n o m e n o n itself. If one thinks only of utterances as examples of representational intentionality, it can seem unnecessary to say that representational intentionality makes a claim not to directly present its correlate. W h y not j u s t say that representational intentionality does not directly present its correlate? N o chance of making a mistake seems to present itself. But representational intentionality also includes such things as different types of pictorial representation, including d o c u m e n t a r y films; a n d here the possibility of mistakes is sometimes present. A w a x doll of a person is a sort of picture of the person. Assume, now, that you 202

INTENTIONALITY

believe yourself to see a wax doll representing a certain person in spite of the fact that what you actually see is the real person. This is an example of representational not of presentational intentionality, because the intentional act is such that it claims to have its intentional correlate outside itself, in spite of this not being the case. Both presentational and representational intentionality have a satisfaction modality, but the way we determine the modal value is not the same in each case. In representational intentionality the modal value is determined by acts of presentational intentionality. Whether it is true that С is digging (utterance = representational intentionality) is determined by someone's going and seeing whether this is the case (perception = presentational intentionality). In the case of presentational intentionality the modal value is determined with the help of other presentational acts. This is possible because presentational intentional acts also point beyond themselves. As remarked above, a normal perception contains not only a pointing from a subject to a simultaneously present state of affairs; it also points to both the past and the future. It is of course especially the latter which allows us to determine whether the modal value is 'satisfied' or 'unsatisfied'. 7 If we take into account fictional intentionality and disregard partial satisfaction, we get Table 13.1. It should be stressed that to be a modality is quite different from being a subcategory; that is the reason why the concepts of 'satisfied' and 'not satisfied' each appear in two rows. That an intentional act lacks the modality of satisfaction is not the same as non-satisfaction. Nevertheless, there is a similarity between unsatisfied presentational, unsatisfied representational, and fictional intentionality. They all lack correlates. False stories and fiction are in this respect similar, although different in kind. Only satisfied Table 13.1 Category

Subcategory

Sub-subcategory

Real intentionality

' Presentational intentionality Representational . intentionality

Intentionality Fictional intentionality

203

Modality-value , satisfied \ not satisfied r satisfied not satisfied

ONTOLOGICAL INVESTIGATIONS presentational and satisfied representational intentional acts d o h a v e correlates, a n d only in these cases m a y intentionality be relational. (This topic will be discussed in section 13.5). I n those cases where there is no correlate, the elusive peculiarity of the category of intentionality is most a p p a r e n t . T h a t which exists is simply pure directedness. A n d that this type of p h e n o m e n o n truly exists should actually be equally as obvious as that p h e n o m e n a belonging to the category of property exist. Both are self-evident from a phenomenological point of view.

13.4 M I X E D I N T E N T I O N A L I T Y T h e fact that there are three basic kinds of intentionality, presentational, representational, and fictional, does not m e a n that a specific intentional act in its totality instantiates only one of these categories. An intentional act can be a unity of different kinds of intentionality. I n such a case we have mixed intentionality. L e t us once more use the example discussed in the former section, the perception of a digging m a n , C. W h e n we perceive С s digging we see that we see only one side, the outside of С and the shovel. T h e perception makes the claim that parts of these things are presented, b u t at the same time it makes the claim that other p a r t s are represented. I t contains a pointing to facts of the type t h a t b o t h С a n d the shovel have back-sides, insides, a n d so on. C s digging is thus not completely presented even if we disregard its temporal extension. Such a perception contains a mixture of presentational and representational intentionality. T h e presentation of the thing is only partial, part of the thing is represented by m e a n s of its presentation of other parts. Most perceptions seem to exemplify mixed intentionality, b u t there are perceptions which are purely presentational. Pains a n d after-images, as well as everything Husserl calls a d e q u a t e perception, 8 are purely presentational. Both pains a n d after-images lack back-sides a n d insides. They have nothing which is spatially hidden, a n d are therefore completely prfesented. Most novels seem also to exemplify mixed intentionality. But in this case there is a mix of representational a n d fictional intentionality. Actually, fictional discourse is not purely fictional. A story with fictional characters is usually situated in a real environ-

204

INTENTIONALITY

ment, known places in a known country. Even a genre like science fiction is partially fictional. These stories are situated in our real universe with its stars and its planets. Hypothetical deliberation as when one thinks: If such and such a thing would happen to me, what would I then do? - is a mixture of representational and fictional intentionality, too. However, when, in what follows, I discuss cases which actually are partly presentational (perceptions) or partly representational, I shall for the sake of simplicity regard them as purely presentational and purely representational respectively. 13.5 I N T E N T I O N A L I T Y AND T H E O T H E R CATEGORIES I shall now bring out in more detail the differentia sperifica of real intentionality, its subcategories, and the satisfaction modality by comparing intentionality with some other categories with which it is easily confused. It is all too easy to think that intentionality is some kind of relation (i.e. external, grounded, or existential dependence) or that it in fact belongs to the category of tendency. I shall start by discussing the question whether representational intentionality can be a relation. In chapter 9 I made the important point that some relations (of existential dependence) lack mutuality. They relate one of the relata to the other without thereby necessarily relating the other to the first. Still, both relata have to exist in order for the relation to be instantiated. This is a necessary condition for something to be called a relation. If intentionality is going to be a relation the intentional act and the intentional correlate must be the two relata. Consider the representational intentional act which is my judging that it is going to rain in my home town tomorrow. The presumed intentional correlate, rain-tomorrow-in-my-home-town is both spatially and temporally distinct from the actual intentional act, my judgement. Now, if there is no rain tomorrow there is no correlate; the act is unsatisfied and the intentional act cannot possibly be a relation. The same conclusion, however, is reached even if it rains tomorrow. Both the existence and the identity of my judgement, the representational act, are independent of the rain which will fall tomorrow. The act is instantiated now. It can exist without the presumed relata, the correlate, because this is transcendent. This means that representational intentionality, whatever the value of the 205

ONTO LOGICAL INVESTIGATIONS satisfaction modality c a n n o t be a relation, neither a n external nor a g r o u n d e d relation nor a relation of existential dependence. T h e fact that representational intentionality (or intentionality in general) is not a relation of existential dependence is compatible with t h e facts that it can b o t h contain existential dependence (see below, pages 210-12) a n d itself in turn be existentially d e p e n d e n t u p o n other categories. In m y view, intentional acts a r e founded u p o n a material s u b s t r a t u m (some sort of body, brain, a n d nervous system). W h e n I speak of an intentional act I am speaking of something which is a m o m e n t o f a relatively independent state of affairs constituted by the intentional act a n d its material s u b s t r a t u m . Intentional acts can only exist on upperlying ontological levels, a n d as such are dependent. A n o t h e r point which has contributed to the view that intentional acts are some kind of relation is, I think, a confusion of the intentional act itself with its modality of satisfaction. T h e latter is really a grounded relation, although of a special kind. It is evident f r o m what has been said that, w h e t h e r a representational act is satisfied or not is a fact which does not inhere in the act. W e have so far met two kinds of grounded relations, i.e. grounded relations w i t h two different kinds of relata. First, relations like 'taller t h a n ' which are grounded in ordinary properties, then relations like ' h a v e greater distance between them t h a n ' where the relata are external relations. When a representational act is satisfied there is a grounded relation between the act and the correlate.9 T h e act a n d the correlate exist independently of each other, b u t w h e n they both exist then it follows immediately that satisfaction obtains. Sometimes, but only sometimes, such a grounded relation is a relation of similarity. T h i s is the case in pictures a n d iconic languages b u t not in ordinary languages. I n the last case there is for the most p a r t nothing at all in the acts which is similar to the intentional correlates. Intentionality is a n irreducible kind of universal. But this does not m e a n that intentionality is unanalysable. Irreducibility does not entail unanalysability. This was stressed earlier when the category of state of affairs was introduced. A state of affairs is a complex universal constituted by a substance a n d some properties, a n d so it can be analysed into substance a n d property. T h e irreducibility of states of affairs follows f r o m the fact that this category is a relatively independent universal whereas both 206

INTENTIONALITY

substance and property are dependent categories. What is independent cannot be reduced to what is dependent. Every ontological reduction has to go the other way. I would like to repeat what I said in chapter 3 (see page 34), namely that it is my impression that many philosophers have a strong temptation to connect what is simple with what has independent existence, and vice versa, but that the category of state of affairs breaks with such associations. Now I shall add that the category of intentionality also makes such a break. I do not intend to try to lay bare the parts (dependent and/or independent) of intentional phenomena — a kind of analysis of which Edmund Husserl is the great founding father. In the Logical Investigations the main parts of an act are 'sensory content', 'act matter', and 'act quality'. 10 Such analyses are of course ontologically very important, but for the purposes of this theory of categories it suffices to point out that there is a category of intentionality; and that it is a complex universal. The addition is necessary in order to avoid the mistake of looking for an intentional act in one of its simple parts. With regard to Husserl's tripartition, I maintain like Husserl, that intentionality is not to be found in any of these parts, only in their unity." Now to the question of representational intentionality's relation to tendencies. Here the superficial similarities are considerable. 12 Tendencies point at something and point forward in time. That a thing has a certain velocity is the same as that the thing has a tendency to be, in the future, in other places in the direction of the velocity. The thing points from itself towards these places. The similarity with intentional phenomena can be made even greater if we bear in mind that the tendency category is so wide that it covers Aristotelian physics. The tendencies in Newtonian physics, velocity and acceleration, can perhaps with some justice be said to point more away from a place than toward a place. But the Aristotelian tendencies are definitely tendencies towards something. Earth, the stuff, has a tendency toward the midpoint of the universe, water to the surface of a sphere with the universe as midpoint, etc. As a consequence, each bit of earth points toward the midpoint of the universe. The tendency towards the middle of the universe postulated by Aristotle, has as a presupposition that that midpoint exists. And in the same way Newtonian velocity cannot point toward any places if 207

ONTOLOGICAL INVESTIGATIONS

those places do not exist. This peculiarity is not, however, necessarily connected with the tendency category, which means that the similarities with presentational intentionality can be made even larger. Even if physics up to now has not needed to assume the existence of the tendency speed of colour change, there is nothing in principle which prevents such tendencies existing (cf. pages 102f.). A body which at a certain moment has a certain colour and a certain speed of colour change has a tendency to have another colour in the next moment. This latter colour instance does not exist in the first moment; the tendency points towards the nonexisting colour instance. What distinguishes this tendency from a belief that the object will come to have this colour in the next instant? Let us assume that speed of colour change is a resultant tendency. Resultant tendencies are characterized by their necessarily being realized if the temporal interval in which they exist is sufficiently long. This means that resultant tendencies can never point at something which is in principle impossible to realize. It need not, like the above-mentioned Aristotelian tendencies, point toward something which actually exists, but it must point at something which in principle can exist. Since the difference between resultant tendencies and partial tendencies is not a difference internal to these, it can be seen as yet another characteristic of tendencies in general, and a characteristic which is a distinguishing mark with respect to representational intentional phenomena. The latter can point at something which is in principle impossible to realize. An intentional phenomenon, but not a tendency, can point at a colour which for physical reasons cannot exist. This may appear to be a reductio ad absurdum of the view that tendency and intentionality are different categories. Did I not use wishes and intentions as examples of tendencies when introducing the tendency category? And of course one can both wish for what is in principle impossible, and have an intention to make it happen! The question can be asked whether I have not directly contradicted myself. In chapter 1 1 1 used intentions as examples of tendencies but in the present chapter I have used them as examples of intentionality. Moreover, I have maintained that tendency and intentionality are categorially distinct. The problem, however, is not so great. One must remember that certain categories can be exemplified at exactly the same place 208

INTENTIONALITY

simultaneously. The best examples of this are properties and substances, which always coincide in things. In a corresponding way there exist at one and the same time in every (prior) intention instantiations of both tendency and intentionality. Often each category points in its own way to the same thing; and then of course it is not easy to see the difference. But it is there. One can find examples where intentionality is directed toward one thing and tendencies toward another. What we ordinarily call an intention is a state of affairs which contains instances of both the category of intentionality and the category of tendency, i.e. both an intentional act and inclinations to perform certain actions (cf. section 7.3). This holds true even when the intentional correlate in question is a perpetuum mobile. The man who tries to do the impossible nevertheless performs actions. Intentions point forward in time, but memories point backward. The difference between presentational intentionality and tendency can also be clarified with the help of a discussion of memory as an intentional phenomenon. Memories should thus be compared with so-called hysteresis phenomena. 13 When certain elastic materials are exposed to pressure a mechanical tension occurs in the material; but when one then reduces the pressure the corresponding tension does not disappear but persists. The same phenomenon occurs when one magnetizes a ferromagnetic material with the help of a coil. When the current increases the piece of iron is magnetized, but when the current is then reduced the magnetization does not disappear to a corresponding degree, but persists. It seems as though the materials in question remember what they have experienced, as though they remember that they have earlier been exposed to external pressure or a magnetizing current. The mechanical tension and the magnetic flux density is at each point of time a function both of an external force operating at the moment in question and an outer force operating at an earlier moment. Hysteresis phenomena can perhaps be better understood in the following way. Assume that we have two identical elastic things. The one we subject to a certain pressure for a short while, while the other remains untouched. A little later we subject both things to the same pressure. Due to the hysteresis phenomenon, the thing which had previously been subject to pressure comes to have a higher inner tension than the other apparently identical thing. 209

ONTOLOGICAL INVESTIGATIONS

Since we do not believe that the past can, without mediation, affect the present, we assume that the earlier pressure changed the inner constitution of the thing in question in a lasting way. It is thus not as though two identical things have been subjected to the same pressure. The changed state of the first thing is a result of earlier occurrences, but it is not a memory in the intentional sense. The state does not point backward in time towards an earlier effect. With these remarks I end my discussion of representational intentionality, the conclusion being that it can be neither a relation nor a tendency. The next topic is presentational intentionality; first the question whether it can be a tendency, then the question whether it can be some kind of relation. The characteristic feature of presentational intentionality is that it lays claim to be in direct contact with its correlate or, in other words, claims that the correlate is immanent in the act. It claims to include in itself that which it points at. Tendencies make quite the opposite claim. They point away from themselves. If a tendency exists now, it points to something in the future. Consequently, presentational intentionality cannot be a sort of tendency. Similarities between relations and presentational intentionality occur only when the intentional acts are satisfied. In the case of non-satisfaction, the argument for presentational intentionality not being a relation is exactly the same as for representational intentionality. If there is no correlate, one relatum is missing and so there is no relation. When presentational intentional phenomena are satisfied, there is both an act and a correlate, but intentionality can nevertheless not be a relation. The reason is that a satisfied and an unsatisfied presentational act can be qualitatively identical. If intentionality cannot be a relation in the one case, then it cannot because of this identity be a relation in the other case either. Representational and presentational intentionality are neither relations nor tendencies. That presentational intentionality is not a relation is compatible with the fact that there are relations at work in presentational acts. Take for instance a veridical perception. In such an intentional phenomenon both (part of) the subject and the intentional correlate must, prima fade, be considered to be spatial parts of the act. (In such cases I shall not talk of the subject but of the subject pole of the act.) This observation implies that there is an external relation, the distance between the subject and the correlate, within the act. 210

INTENTIONALITY

Intentionality is however not identical with this spatial distance. Spatial distances lack directedness. As in the case of representational intentionality, the mistaken view that presentational intentionality is a relation is the result of a failure to keep the intentional act in itself and its modality of satisfaction clearly distinct. We need to look more closely at satisfaction in the case of presentational intentionality. It is advisable to analyse a veridical perception where the intentional correlate is an ordinary material object rather than the case where it is something psychic like a pain. In the former case, but not the latter, the correlate is totally independent (D9.10) of the act, which means that the only dependence relations we have to investigate are those relating the act to the correlate. I shall focus the discussion around the perception of a birch tree. What kind of act a certain act is, is determined by its inner qualities. A veridical perception of a birch tree and a corresponding mirage of a birch tree are generically the same act; they instantiate the same universal but have different values on the satisfaction modality. An immediate consequence of this fact is that even a veridical perception of a birch tree is existentially independent (in the sense of D9.2) of the birch tree. Let us therefore consider concrete dependence. The claim that the act of seeing a birch tree is concretely dependent upon real birch trees, means that it is logically impossible for all, but not some, instances of birch perceptions to exist if some instance(s) of birch trees do(es) not also exist (D9.7). A logical impossibility here would be a neat solution but I find it impossible to detect any such impossibility. The conclusion this far is that a veridical perception of a birch tree can be neither dependent nor concretely dependent upon the birch tree. The act is generically independent (D9.9) of the tree. Is then the curious relation of individual dependence (D9.8) exemplified by perception? The perception of the birch tree would be individually dependent upon the birch tree if and only if it is logically impossible for this instance of a birch perception to exist if the birch tree did not also exist. In my view the same intuitions are relevent here as in the case of efficient causality (see above, pages 184-5f.). Birch-tree perceptions in general can exist without birch trees, but this specific perception is logically impossible without the birch tree. The epistemological 211

ONTOLOGICAL INVESTIGATIONS

test also holds. It is impossible to conceive of this instance of birch perception were the birch tree to disappear. Individual dependence, like some of the other kinds of dependence defined, lacks mutuality. The claim that the birch perception is individually dependent upon the birch tree is consistent with the assumption made that the tree is totally independent of the perception. When a representational act is satisfied there is a grounded relation between the act and the correlate. The conclusion now is that when a presentational act is satisfied the act is individually dependent upon the correlate. Such an act is both directed towards the correlate and dependent for its existence upon the same correlate. But this does not turn the intentionality in question into a dependence relation. Intentionality is in general founded upon a subject substratum, but some satisfied presentational intentional acts are in addition to this (individually) founded upon the intentional correlate. In neither case can the dependence relation bear the burden of an ontological reduction; a reduction means in the first case that intentional acts are brain states, and in the latter that intentionality if a dependence relation. At the end of section 13.3 I said that only satisfied intentional acts have correlates and may be relational. The arguments put forward in this section yield the result that intentionality never is a relation, but that (a) a satisfied representational intentional act is relational in the sense that there is a grounded relation between the act and its correlate, and that (b) a satisfied presentational act is relational in the sense that there is a relation of individual dependence between the act and its correlate. 14 This, in turn, means that the former kind of acts has transcendent correlates and that the latter kind of acts has immanent correlates. All other kinds of acts lack correlates; see Table 13.2. The analysis of veridical perception given above according to which such an act is individually dependent upon an immanent correlate, brings out both the similarities and dissimilarities Table

13.2

Kind of intentionality

Intenlional correlate

Presentational intentionality, satisfied Representational intentionality, satisfied Fictional intentionality and unsatisfied real intentionality

Immanent Transcendent None

212

INTENTIONALITY

between such a relation and efficient causality as analysed in the last chapter. It should be remembered that we are not talking about the substratum level, i.e. the flow of energy from the tree to the subject. We are staying on the upper level where the intentional act as such exists. But even on this level it is tempting to say that the intentional correlate causes the specific act. The temptation should be resisted, but the kernel of truth is that as a veridical perception is individually dependent upon its intentional correlate so an effect in efficient causality is individually dependent upon its cause. But then there are also differences. The effect belongs to the category of tendency, the act to the category of intentionality. Cause and effect are moments of different things, but the act contains the correlate in cases of presentational intentionality.15 Another way to approach the specificity of presentational intentionality is to try once again to use the thought operation with whose help the important distinction between inclusive and exclusive qualities was introduced. If one tries to use this operation on ordinary perceptions one discovers that this is impossible, in spite of the fact that the perception of a definite object is a spatially well-defined phenomenon. Think of the perception of a birch. The intentional phenomenon extends itself in space from the subject to the birch (see Figure 13.1). If one now performs the thought operation of cutting off the intentional phenomenon just in front of the birch, neither of the two things which ought to happen were intentionality a quality does in fact happen. Either (a) a new welldefined intentional act with an intentional goal lying in the 'crosssection' ought to come into being, or (b) the entire intentional act

φ

Figure

213

13.1

ONTOLOGICAL INVESTIGATIONS

disappears. But neither is the case. The cross-section remains empty and is not the goal of any act and the act does not disappear; instead it rather continues to point to the empty crosssection. The result of the cutting-ofF operation is a presentational intentionality without an intentional correlate; it is not even an example of fictional intentionality. The conclusion of this thought experiment is that the cutting-ofF operation is quite simply not applicable to presentational intentionality. There exists, however, a similar yet different operation which can be applied. One can in thought exchange one intentional correlate for another, or move the intentional correlate in space. This operation captures well the peculiarity of presentational intentionality we called 'intentionality at a distance'. If we imagine the birch moved further away, without another object being inserted, then intentionality so to speak 'follows along'. Metaphorically speaking, presentational intentionality extends itself till it hits something, which means that if that something is moved then the intentionality's extension in space is also changed. Presentational intentionality has spatial limits, but they cannot be the same sort of limits that ordinary things and their qualities have. 13.6 INTENTIONALITY AND CONSCIOUSNESS Intentional phenomena are thus neither relations, nor tendencies, nor genuine properties. Now I shall discuss the question whether intentional phenomena are always phenomena belonging to consciousness. Can there exist unconscious (or α-conscious) intentionality? The intentional phenomena which are best known are conscious. To believe something, to perceive something, and to remember something are phenomena directly given in our conscious experience. The evidence for unconscious intentionality is indirect. Phenomena like post-hypnotic suggestion and subliminal perception are difficult to reduce to tendencies and hysteresis phenomena. One can imagine a case of post-hypnotic suggestion where a person under hypnosis is ordered to do something which is in principle impossible, e.g. to build a perpetuum mobile. The hypnotized person's actions must then be understood against the background of both an 'intentional intention' and a 'tendential intention', which do not coincide. That subliminal perceptions are intentional phenomena means that they can in principle be 214

INTENTIONALITY

mistaken, i.e. that one can literally speak of subliminal mwperceptions. It is not self-evident that unconscious misperceptions are impossible. Whether one believes in unconscious intentionality or not will depend to a large extent on one's attitude to psychoanalysis. Psychoanalysts refer to unconscious intentions, memories, and beliefs. If there is unconscious intentionality, psychoanalysis must be both true (or truthlike) and impossible to reduce to neurophysiological theories which speak only of tendencies and hysteresis phenomena. Many modern machines are described as if they contained intentionality. Some typewriters, tooling machines, and of course computers, are said to have 'memories'. But these 'memories' are not taken to be memories in the literal sense. A typewriter remembers just as little as a stone containing a fossil remembers prehistorical eras. Here 'memory' is a metaphorical characterization of a complicated setup of ordinary properties, relations, tendencies, and hysteresis phenomena. If we bear in mind developments within electronics we may wonder whether claims made by psychoanalysts about the irreducibility of psychological phenomena are really justified. But on the other hand, advanced computers do not prove that unconscious intentionality cannot exist. They prove rather the existence of extremely complicated hysteresis phenomena. The question posed, that of the relation between intentionality and consciousness, should not be conflated with those of the relations between intelligence and intentionality or between intelligence and consciousness. Modern research in artificial intelligence shows that our concept of intelligence has been much too closely linked to intentionality and consciousness. But, as Searle stresses, 16 this in no way implies the non-existence of intentionality or consciousness, nor that there now exists artificial intentionality or artificial consciousness. Whether the lower left hand square in Table 13.3 is necessarily empty or merely, as a matter of fact, empty today, is something for the future to decide. It might well be the case that intelligence without intentionality and consciousness will turn out to be much more intelligent than intelligence with intentionality and consciousness. The category of intentionality has been introduced without the help of the concepts of intelligence and consciousness and is, as a 215

ONTOLOGICAL INVESTIGATIONS Table

Natural intelligence Artificial intelligence

13.3

Intelligence with intentionality

Intelligence without intentionality

higher animals

lower animals

?

computers

category, neutral with respect to the question whether artificial or unconscious intentionality can exist. And I see no scientific reasons for limiting the category in one way or the other. W e may also say that it is intentionality which is the ontologically important category, not consciousness, nor intelligence. 13.7 I N D E F E N C E O F N A I V E R E A L I S M W h e t h e r one does or does not accept the existence of unconscious intentionality, a classical philosophical problem remains. With the distinctions already made in this chapter, the problem can be formulated in the following way: Can an intentional correlate to which someone has access via conscious presentational intentionality be material? Or, in other words: C a n something psychic be in (intentional) contact with something material? This last question raises the question of naive realism, i.e. the view that, most of the time, we are in some kind of direct contact with the objects we are looking at, objects which exist independently of the onlooker. Naive realism is enshrined in common-sense attitudes, and, very possibly in those of all past cultures. An ontology which conforms to naive realism is on the side of the majority. It also, as we shall see, has strong arguments on its side. Hitherto in the present chapter, my strategy has been that generally accepted among phenomenologists. O n e silently assumes a naive realism in one's investigations; perceptions and our 'Lebenswelt' are analysed as though one did not have to worry at all about whether the state of aiTairs in question actually exists. Husserl, as is well known, was not so rash. H e introduced a conscious methodical approach, which he called 'epoche'. Phenomenological investigations were to be performed after the world had been 'put between parentheses': phenomenological analyses should remain neutral with respect to the question whether the perceived 216

INTENTIONALITY

world actually exists or not. As a natural consequence of this, the natural sciences were also bracketed. The idea was that phenomenology need not worry itself about results within natural science. In this way phenomenology was to free itself from the difficulty which science actually presents for naive realism. It seems to me as though surprisingly few philosophers and scientists are today aware of the extent to which science (physics, physiology, and perceptual psychology) implies a world view of an almost Liebnizian type — a monadology, even if it has a materialistic basis. Within contemporary science one assumes that there are material things which emit or reflect some form of energy which moves towards other material things, some of which are so constituted (the higher animals) that when the aforementioned energy hits them, a mental entity appears on the scene, a perception. But this perception is presumed to be completely spatially distinct from the material object which ultimately caused the perception (Figure 13.2). The perception is connected via a body to a certain place in space and time, but is a whole completely closed within itself, which is mental and does not even have a spatial connection with other people's perceptions, even though they often have the same causes. Every person is a monad, on this view, though a monad with a material foundation. We are very far from the ' L e b e n s w e l f which both naive realism and today's phenomenology take as given. The whole problem can be presented very briefly in graphical form as a conflict between the Figures 13.2 and 13.3. Husserl's epoche involves what is, from a phenomenological point of view, an internal difficulty which phenomenologists tend to skip over. Phenomenology takes its task to be that of giving a correct description of our perceptions, and in order to do this the

causal relation (energy flow)

Figure 13.2

Perception according to science 217

ONTOLOGICAL INVESTIGATIONS

Figure 133

Perception according to naive realism

world is bracketed. Phenomenologists thereby avoid taking a position on naive realism. But if, as is actually the case, the assumptions of naive realism form a n integral p a r t of most perceptions, then the epoche m e t h o d will prevent phenomenology from reaching its goal. If perceptions are not neutral in relation to naive realism, phenomenology o u g h t not to try to be so either. T h e result has been two tendencies within phenomenology. O n e is to be found in works by philosophers like M e r l e a u - P o n t y a n d Alfred Schutz, w h e r e the lasting impression, no m a t t e r w h a t they explicitly say, is their belief in a form of naive realism. T h e other tendency is represented by the later Husserl, who was d r a w n more a n d more towards a classical idealistic position w h e r e the world in the end is constituted by a 'transcendental ego'. 1 7 Q u i t e a p a r t from its inconsistency with the goals of phenomenology, the m e t h o d of epoche is unacceptable within the framework of the present category system. T h e task of a m o d e r n theory of categories m u s t be, as I said at the beginning, to reunite m a n , n a t u r e , a n d society. A n d such a union is not possible if one puts the world a n d n a t u r a l science in parentheses. W h a t I a m going to say a b o u t naive realism will be r a t h e r brief. But I p u t it forward as a first step a w a y from the intellectually unsatisfactory situation in which we live today, where science implies a monadology which n o b o d y really takes seriously, or which is not even noticed due to the specialization within the sciences a n d the l a n g u a g e - m o n o m a n i a in philosophy. According to naive realism we can be in direct contact with objects located a t a distance from us; w h e n we see a tree we are in direct contact with it in the same sense as w h e n we hold on to it. T h e difficulties in fitting this view into a materialistic framework a r e mainly two. O n e difficulty (Л) has to do with w h a t is specific to the category of intentionality itself, the fact that it is normally, even in the case of presentational intentionality, 'intentionality at a 218

INTENTIONALITY

distance'. The other difficulty (B) has its basis in the point the scientist makes, according to which all perception of outer objects requires the transportation of energy from the object in question to the subject, from which it follows that the intentional correlate must be temporally and so also numercially distinct from the actual material object. Let us consider these difficulties. Difficulty A. Traditional materialist world views, like the present ontology, are based on a container conception of space and time. All material things exist in space, and the things are characterized not only by their extension but also by their compactness in space and time. The latter means that if two thing-bits lack spatial connection, then they cannot be parts of the same material thing; 'thing' is taken in the fundamental sense, i.e. not in the sense of aggregate things. The same 'compactness condition' holds for the temporal dimension. If there is a temporal gap between two thingbits, then those two bits cannot be parts of the same thing. It is true that some of the categories presented in earlier chapters have diverged from traditional materialistic categories but they have not diverged in such a way that their instances do not fulfil the compactness condition. The divergencies from 'ordinary' materialism I am thinking of are mainly (1) the introduction of universale which are temporally inclusive, (2) the tendency category, and (3) the substitution of 'action by mixture' for 'action by contact'. (1) Universals which are inclusive in time are universals which are necessarily extended in time, i.e. which do not like material atoms allow themselves to be thought of as existing at momentary points of time. Instances of such universals must extend through the whole of the time in which they exist. They cannot exist at one time, and then cease to exist only to exist again later. Inclusive universals, like exclusive ones, do not allow their instances to have temporal gaps. They thus satisfy the compactness condition, even if this now concerns compactness in time. (2) Tendencies diverge from categories usual in modern materialist philosophies in that they represent a form of pointing which is not usually explicitly included in materialistic ontologies; I have in fact maintained that they must always be implicitly included. But it is a pointing which does not come into conflict with the compactness 219

ONTOLOGICAL INVESTIGATIONS

condition. An instantiated tendency is just sis compact in space and time as instances of every other described universal. (3) The division of the category ofthing into 'corpuscles' and 'penetrables', which the discussion of the category of efficient causality led to, conflicts radically with traditional materialism. On this view the category of thing is usually defined in terms of occupying a spatial place and hindering other things from occupying the same place. Nevertheless, even the subcategory 'penetrables' fulfils the requirement of compactness in space. Presentational intentionality, however, unlike inclusive universals, tendencies and penetrables, does not fulfil the compactness condition. When a subject perceives a situation, the subject is at a particular place in space and the situation at another. The intentional phenomenon connects the subject and the situation without, so to speak, 'filling out' the space between them. It is this spatial 'hop' which keeps the compactness condition from being fulfilled, and does not permit the applicability of the cutting-off operation. The actual experience of this type of phenomenon is undeniable, it reoccurs in every everyday perception. Nor has anyone directly denied it; but most philosophers have been content to say that it must be a mental phenomenon, as though a simple classification would solve the ontological difficulty. Perhaps there has been a confusion of ontological problems with problems of the special sciences. Ontologically speaking, the bewildering phenomenon 'intentionality at a distance' exists whether or not one calls it material or mental. But if one calls it mental the possibility arises for the special sciences of conceiving it as an epiphenomenon. The point of the distinction between primary and secondary qualities used in laying the foundations of modern Galileian-Newtonian physics was to place all phenomena which did not fit into the category of physical entity into another sphere, and thereby to place them outside science. All 'mystical' qualities are dispatched into the mental sphere. It is time to ask ourselves whether there actually are any reasons for taking such a course. One of the truly great difficulties for a different (and more correct) understanding of perceptions and presentational intentionality has been the common failure to grasp the peculiarities of 220

INTENTIONALITY

the type of level-ontology which this category system contains. If intentionality exists directly in space, it must be thought of as existing at the same place as material things. The subject-pole (see above, page 210) in intentionality is (when it concerns people) onesidedly existentially dependent on a thing with a nervous system and brain, but exists at exactly the same place as they do. Intentionality must exist on a higher level. Irreductive materialism here cuts through a tangled knot. Where do our dreams, for example, exist? Answer: As an overlying level in the same place as our bodies! With the help of the category of one-sided foundation, the mental allows itself to be placed in the same space as the material. We are used to thinking that everything which a person perceives in a certain place at a certain time, other people can also perceive in the same place at the same time; blindness and similar phenomena constitute exceptions which seem to prove the rule. If, however, dreams exist in space in the way that I maintain, then ordinary conception must be revised in the name of consistency. That a dream exists at a certain place in space in a dreamer implies that another normal person cannot experience that dream in spite of his being able to perceive the actual place where that dream exists. This is an unusual way of thinking, but is actually no stranger than the fact that we can perceive a place where there is air, but not perceive the air; not to speak of the fact that we can have a lot of radio and TV programmes in our bodies without being aware of them (cf. the discussion of 'penetrables' in section 12.5). If we now accept that the subject-pole of the intentional phenomenon can exist in the same space as material things, then the next question is whether such a subject-pole can be in direct contact with material states of affairs which are spatially separated from the subject-pole. The point of departure must be, as I pointed out earlier, that every complete ontology must somewhere have a place for 'intentionality at a distance'. The problem is thus not to understand and accept the phenomenon in itself - that has to be taken as given; the problem is rather the phenomenon's relations to the other categories. In the case of representational intentionality (as in the case of presentational intentionality which does not reach its correlate) intentional phenomena obtain their place in space only through 221

ONTOLOGICAL INVESTIGATIONS their one-sided existential dependence on that which is the s u b s t r a t u m of the subject. But in the case of presentational intentionality which reaches its goal, intentionality is one-sidedly d e p e n d e n t both on the subject-substratum (generically) and the s t a t e of affairs which is perceived (individually). Since intentionality can b e conscious a n d the perceived state of affairs can be material, this existential dependence m a y hold between something m e n t a l a n d something material. A n d n o t h i n g more need actually be said if one accepts that the subject-pole can be placed in ordinary space. Irreductive materialism itself explains the principal p r o b l e m of the connection between mental subjectivity a n d material objectivity. Difficulty B. W i t h this I leave the general problem of 'intentionality a t a distance' for the spetific problem, i.e. the problem to which today's science gives rise. In the discussion so far I have silently let 'intentionality at a distance' be equivalent with 'intentionality at a spatial distance'. T h e intentionality of a state of affairs a n d that s t a t e of affairs itself have been understood as simultaneous b u t spatially incongruent phenomena. If, however, one takes seriously, as I believe one must, the view of today's science that perceptions of external states of affairs are always based on transport of energy f r o m the perceived state of affairs to the perceiving subject, then the discussion m u s t free itself from the presuppositions hitherto made. Given the 'energy transport thesis', 'intentionality at a spatial distance' implies 'intentionality at a temporal distance'. Energy t r a n s p o r t must be thought of as existing on that level which is the s u b s t r a t u m of the intentional p h e n o m e n o n itself. W h e n we see or h e a r something, we have an intentional act directed towards that something, b u t we have no intentional act directed towards the light or sound waves which make up the energy transport which is a necessary existence condition for the perceptions in question. I n o r d e r for the perception itself to occur, the energy transport must h a v e reached the subject substratum, which m e a n s that when perceptions of the state of afTairs in question occur that state of affairs n o longer exists — the transport of energy takes time. If one w a n t s to m a i n t a i n that, in our o r d i n a r y perceptions, we can be in contact with spatially distant material states of affairs, then one also has to m a i n t a i n that we perceive not only across space b u t also across time. Presentational intentionality becomes 'intentionality at a spatial and temporal distance'. 222

INTENTIONALITY

I have chosen the expression 'intentionality at a distance* in order to give associations to the classical discussion of 'action at a distance' — otherwise the expression 'intentionality across a distance' would be better. When an intentional correlate is reached across space, the space often appears as empty, but not always. When one perceives in the fog one does not only see that distant state of affairs, but the fog itself is also included in the perception. One perceives a state of affairs through the fog, i.e. also through space. And this description is adequate even when the space itself appears empty. The same description also applies to time. Presentational intentionality is 'intentionality across time'. The connecting together of distance in space and time means that if one, for example, perceives an object through a lot of panes of glass placed after one another, then the perception of the different panes is extended in time in such a way that the nearer a pane is located in space the closer it is also located in time, and vice versa. The perception extends through both space and time. The difficulty with 'intentionality through time' in comparison with 'intentionality through space' is that in the first case, but not in the second, one is in contact with something which no longer exists. In both cases it of course holds that one is in contact with something which is located at a distance. Space and time are not equivalent dimensions. The 'hop' in space which presentational intentionality involves can be rather difficult to accept, but it is nevertheless a relation between subjects and objects which are simultaneous. A 'hop' in time is a relation which connects the present subject with an earlier existing object. This means moreover, if we assume that a subject can be affected by its intentional correlates, that we are forced to accept that the past, without any hysteresis phenomena, can affect the present. 'Intentionality at a distance' thus leads in the end to a (mediated) 'action at a distance' — in the sense of both 'spatial distance' and 'temporal distance'. This presents an almost insurmountable problem. But let us take a broader perspective. What point of view are we forced to if we do not accept the views just presented? We are nowadays often taught to make choices in terms of alternative costs. The question is not what a thing in itself costs, rather how much it costs in relation to the other possible alternatives. One is sometimes also forced to argue in a similar way in ontological contexts. Let us see what the 'costs' of the different ontological alternatives are. 223

ONTOLOGICAL INVESTIGATIONS O n e alternative is to accept that 'spatial hops' can only occur in a completely m e n t a l sphere, and t h a t the ' h o p ' c a n n o t be temporal. If w e dismiss p u r e idealism, then this alternative, together with traditional perceptual psychology, implies a monadology. Every person is confined within his own m e n t a l world. T h e only connection which exists with other people are causal relations on the material level. If we are not to completely a b a n d o n the belief t h a t we live in a c o m m o n world, we are forced, as was Leibniz, to postulate a predetermined harmony a m o n g all of the different separated m e n t a l spheres. Fundamentally, we each live in our own m e n t a l world, b u t the worlds have great similarities with one another. Ontologically seen, we are as h u m a n s , in contradistinction to clumps of p u r e m a t t e r , completely a n d helplessly isolated from one another. T h e other alternative is to accept an irreductive materialism a n d 'intentionality at a distance' as a special form of connection across b o t h space a n d time. A connection which, without necessarily being mental, allows 'hops' over b o t h spatial a n d temporal distance. T h i s type of connection is the only type which allows us to retain our conception that, in a literal sense, we live in a c o m m o n world. It implies, given the view t h a t perceptions require energy t r a n s p o r t f r o m the object to the subject, t h a t we normally perceive backwards in time. 1 8 T h e subject- a n d object-poles in a perception are not simultaneous. W e perceive through time (as t h r o u g h space), b u t only backwards. Such a conception does not u p s e t our everyday conception very m u c h , as does the view that we can perceive forwards in time. But there are neither ontological nor scientific reasons for the latter view. T h e two alternatives described here are the m a i n two possible alternatives today. T h e choice consists — to p u t it more sharply - in either accepting a monadology or in accepting that we can be in direct contact both with distant states of affairs a n d with the past. I t seems obvious to m e that the costs of the first alternative are too high. W e must begin to accustom ourselves philosophically to the t h o u g h t we all daily take as implicitly given, namely that we are in at least partially direct contact with b o t h n a t u r e a n d other people. W h a t is most difficult to accept in this form of naive realism which I a m advocating, is of course t h a t part which is not directly in keeping with genuine naive realism, namely the view that one perceives across or through time instead of at a particular m o m e n t . 224

INTENTIONALITY

Here one ought to remember that 'through time' is always connected with 'through space', otherwise it is easy to become confused. Intuitions which presuppose that spatial perceptions are conceived as momentaiy, might easily be transferred to the temporal dimension. Think of a perception where a door is suddenly opened out towards nature. If one conceives this perception as momentary, one must regard both the perceived door and all the perceived further distant states of affairs as being simultaneous. The spatial dimension so to speak opens itself out towards more distant but simultaneously existing states of affairs. It is tempting to conceive of intentionality through time via this type of picture, i.e. suddenly time opens itself towards the past; one perceives new, more temporally distant existing states of affairs. These are supposed to be in some way both past and present. The temporal dimension is supposed to function precisely as does the spatial dimension in traditional naive realism. From my point of view both of the pictures described are mistaken. Space cannot, in the present sense, open itself towards more distant states of affairs without time's also opening itself — and vice versa. It is an illusion that ordinary temporal perceptions are momentary in time. If one accepts this, then it is not difficult to imagine perceptions through time. I have pointed out in this chapter that intentionality towards states of affairs which exist in space and time is normally 'intentionality at a distance and through a distance'. I shall conclude with a few comments which indicate that this is not only normal, but is a defining characteristic of intentionality. In the case of representational intentionality and perceptions of outer objects and states of affairs, this 'hop in space', as I also call it, is obvious. But the 'hop' is also there in the case of inner perceptions, i.e. perceptions of such things as pains. Here, perhaps, it seems as though there were no spatial distance between subject and intentional correlate. One must however introduce a distinction here between subject and subject-pole, which was not necessary in the earlier examples. A subject is a definite spatio-temporal entity which can have intentionality, normally a body with a brain and nervous system. Such a subject can have presentational intentionality towards parts of this same spatio-temporal entity, which is what happens not only in the case of pains but also when one performs actions such as cutting one's hair or pinching oneself. In 225

ONTOLOGICAL INVESTIGATIONS

these cases the subject is divided into a subject-pole and an objectpart. The intentional correlate thereby becomes, in this case also, distinct from the subject-pole. Another possible exception to the rule that intentionality is 'intentionality at a distance' is provided by acts of self-consciousness. Here it can really seem as though the subject-pole must coincide with the object-part. But then one has forgotten the temporal features of such acts. I believe that self-consciousness is always consciousness of the moment which is just past, i.e. it is 'intentionality at a temporal distance'. I believe it impossible to notice one's own noticing, other than in the way one notices an earlier noticing. The subject pole is in this case at a temporal distance from the intentional correlate. I have earlier defended the possibility in principle of such intentionality.

226

14

NATURE: PARTS AND WHOLES WITHOUT INTENTIONALITY

We arc now in a position to make use of all the categories introduced so far in order to arrive at our final conclusions about nature, man, and society. Everything that exists, exists as we have seen, in container space. In this chapter I shall look at the part-whole structure of the non-intentional denizens of container space and in the next chapter at the structure of man and society in container space. Since our topic here is the part-whole structure of nonintentional phenomena in container space we shall of course not repeat what we have already said about space. But it must be borne in mind in what follows that we assume that functions, machines, and systems are in space. We have already made the same assumption about, for example, existential dependence. The concept 'logically impossible to exist', which appears in the definitions of the different kinds of existential dependence, concerns existence in time and space. This chapter is divided into six sections. In the first section I make some remarks about aggregates, and in the second I bring out some consequences of the distinction between lower- and equidimensional parts which was introduced in section 6.2. Both sections are concerned with universals which are exclusive in time. The third section, on the other hand, is concerned with a universal which is inclusive in time, motion. The important insight of this section is the irreducibility of the concept of 'mechanism'; a mechanism, we shall argue, is not the same as a system with efficient causality. Section 14.4 is not concerned with ontology. It is an interlude about language which is necessary in order not to misunderstand the claims made in 14.5. 227

ONTOLOGICAL INVESTIGATIONS

T h e key section is the fifth, which analyses yet another type of temporally inclusive universals, functions, and functional wholes. Unlike actions (which also are inclusive in time), functions are not permeated by the intentionality category. I argue that functions are another type of irreducible mechanism. This leads to a description of the ontological insights of systems theory. W h a t is said about mechanisms in this fifth section is meant to apply to both organisms and natural processes in general as well as machines, although only a machine example is dealt with in the text. The extrapolation ought not, however, to create any difficulties. T h e difference between machines and organisms is in principle only a genetic difference. M e n make machines but discover organisms (as well as other natural processes). I n both cases, however, the same type of universals is instantiated. And this is the important insight from an ontological point of view. T h e last section- takes up the idea that machines and organisms must be understood with the help of normative principles or 'rules of Tightness' (Polanyi). This idea is criticized and rejected. 14.1 P A R T A N D W H O L E I N A G G R E G A T E S I n nature there are not only things, there are also heaps of things, i.e. aggregates. Aggregates have a second-order existence (see section 6.1) and are one-sidedly dependent upon their parts (see section 9.3), but they do exist. Some comments on this kind of whole are necessary in order to understand what is distinctive about functional wholes in nature (14.5). Properties can be added and things can be aggregated, but when things are aggregated some properties are always added, even if these are only the spatial volumes of two things. Consider an aggregation of atoms each of which has a certain volume, shape, mass, and colour. T h e n to aggregate the atoms is the same as to place them next to one another in space. By means of this operation a spatial whole is obtained, which (because volume, shape, and mass are spatially inclusive) has a specific volume, a specific shape, and a specific mass. T h e property colour is however spatially exclusive and cannot be added (cf. section 6.1). T h e aggregate has, as an aggregate, no colour in the sense that it has a volume, a shape, and a mass; it has only a colour pattern. If colour were a quantifiable property (like that other exclusive property, 228

PARTS AND WHOLES WITHOUT INTENTIONALITY

density), the aggregate could of course be assigned a colour average, but this is something quite different from a true aggregate property. Inclusive properties are necessary conditions for aggregation, but they are not always sufficient conditions. If the properties in question are subject to or mentioned by natural laws, the requirement that the laws will be preserved in the aggregation is often added. Assume for example, that for every atom it holds that m = k-v, where m is mass, υ volume, and к a constant. If the aggregation takes place in such a way that m and ν are added as in ordinary scalar addition, the aggregate also turns out to satisfy the law. But if, instead, for every atom it is the case that m = k-v2 the situation is different. If m and ν are now added by ordinary addition the aggregate will not satisfy the law. In order for this to be the case, the aggregation rule for m, for example, would have to be changed so that the aggregate's mass Μ = У/Σηι2. The problem of aggregation also appears in a new light when tendencies are taken into account. Because tendencies can cancel each other out, an aggregate of corpuscles can have the resultant tendency zero although all the constituent corpuscles have nonzero tendencies. But I shall not go into the connections between aggregation, laws, and tendencies here. 14.2 LOWER-DIMENSIONAL AND E Q U I - D I M E N S I O N A L PARTS In order to know what the relations are between the properties of an aggregate and those of its parts we must know how the determinates in question are formally related to each other. We must know what the relevant inclusion relations are. Consider the properties mass and volume. Some theoretical work is required in order to show that the two kinds of determinates can be ordered on ratio scales and added like the real numbers, something which is not possible in the case of shapes. This difference between mass and volume on the one hand, and shape on the other hand, reflects the fact mentioned earlier (see above, pages 85-6) that a certain mass-determinate or volume-determinate includes a specific set of other determinates independently of any partition line; whereas in the case of a shape-determinate one usually gets one set of included determinates along one partition line and another set along another line. 229

ONTOLOGICAL INVESTIGATIONS

Given a partition line, however, it is possible to investigate how determinates can be added and give rise to other determinates even in the case of shapes. For mass, volume, and shape such an investigation is an investigation of equi-dimensional parts (the distinction between such parts and lower-dimensional parts was introduced in section 6.2). We add three-dimensional masses to three-dimensional masses, volumes to volumes, and threedimensional shapes to three-dimensional shapes. We do not add a two-dimensional mass (a property not talked about in physics but which can be defined as mass/cm along a line) to a threedimensional mass, nor an area to a volume, nor a two-dimensional shape (for example, a circular shape) to a three-dimensional shape (example: spherical shape). Addition is addition of equi-dimensional parts, but this fact should not hide the fact that in all three-dimensional spatial wholes there are also relations between the whole and its lowerdimensional parts; relations which can be theoretically investigated. I have earlier mentioned in passing (section 7.2) the case of the relation between density and mass. Mass is three-dimensional and density can be regarded as zero-dimensional (i.e. punctual in space) because density is exclusive in all spatial dimensions. If the density of the whole is constant, its mass equals the density times volume, if the density varies the mass is given by integrating the density over the volume (cf. pages 88-9, especially note 7). In a corresponding but more complicated way, there are relations between the three-dimensional shape of a spatial whole and its lower-dimensional shapes (cf. pages 88-9). These relations cannot be captured by the mathematical operation of integration since shapes cannot be ordered on a ratio scale, but ontologically they are relations of the same kind as that between mass and density or that between a volume and (actually and/or potentially) enclosed areas and lines. Neither of the two different kinds of investigations of inclusive properties which I have distinguished above has any place within classical geometry. Let us look at the famous theorem that there exist only five regular polyhedra (the tetrahedron, the cube, the octahedron, the dodecahedron, and the icosahedron); there are only five iArw-dimensional shapes which are bounded by regular too-dimensional part-shapes. The three-dimensional shapes in question are not viewed as a sum (along a partition or additivity 230

PARTS AND WHOLES WITHOUT INTENTIONALITY

Figure 14.1

line) of three-dimensional shapes, nor are they viewed as the result of an (integration) operation over enclosed two-dimensional shapes. Still, the theorem relates three-dimensional wholes to lower-dimensional parts. All the important points made in what follows are concerned with the addition of equi-dimensional determinates. All such additions of spatially inclusive three-dimensional qualities are very simple, but the results are nevertheless true. They can very easily be graphically illustrated if we 'go down' one dimension and look at two-dimensional shapes (see Figure 14.1). Along line of partition 1 and the rightangled cuts indicated, the square shape consists of four different rectangles. Along line of partition 2 with its cut, it consists of two triangles. Partitioning can be varied in many ways. The line of partition may be curved, crooked, etc. This means that one can investigate whether different lines of partition can have the same parts, and so on. The same type of 'geometry' can be developed for patterns. As was shown in chapter 6, patterns are a quite specific type of property which arises because shapes can not only limit things but also limit properties. A thing limited by a certain shape can have different colours that are limited by shapes included in the thing's shape. Some of the included shape-determinates thereby assume a more manifest Gestalt; they stand out more clearly. We see shapes in shapes, or patterns which are spatially inclusive universals. 231

ONTOLOGICAL INVESTIGATIONS

They are Gestalten in the Gestalt-psychological sense of the word. From an epistemological point of view such Gestalten or patterns are often primary; what we see are the patterns. But from an ontological point of view it is the parts that are primary. A pattern is constituted by, is one-sidedly dependent on its 'pattern-parts' just as every other inclusive universal is one-sidedly existentially dependent upon its parts. 14.3 MECHANISMS AND T H E I R PARTS In this section we shall begin to see how spatial part-whole relations connect up with part-whole relations in time. Universals which are inclusive in time can in a sense be added in time. But let us first repeat some of the main features of the exclusive-inclusive distinction. Inclusive universals are inclusive in at least one space-time direction; exclusive universals are exclusive in at least one spacetime direction. A universal can be inclusive in one dimension and exclusive in another. Those universals which are inclusive or exclusive in all three spatial dimensions, I have simply called spatially inclusive or exclusive respectively; otherwise I have indicated in which dimension the inclusiveness obtains. We can now classify some important universals (Table 14.1). If one wants to understand inclusiveness in time a brief analysis of point-formed motion is necessary. Point-formed motion is related to time in the same way as mass is related to space. Given a particular motion, 'the whole motion', we know that it must contain other shorter motions. The point must have passed all the spatial points that exist along the (direction of) the whole motion. But if we divide up motion in, e.g., three part-motions, e> Figure 14Л

is three-dimensional. If a three-dimensional shape can be said to be an accomplishment along a partition line in space, then the confining two- or one-dimensional boundary surface is an achievement. T h e inclusion relation between shapes and their part-shapes is undetermined if a line of partition is not fixed, but is univocally determined if such a line has been chosen. If shapes and actions are inclusive in the same way, the inclusion relation between equidimensional part-actions and actions ought to be univocal if time provides the partition line, since there is only one temporal dimension. But this seems not to b e true, at least not superficially. Pulling the trigger can, but need not, be part offiring the gun, and firing the gun can, but need not, have pulling the trigger as a part - one can fire a gun in other ways. T h e inclusion relation appears, in spite of everything, to be undetermined. But on closer analysis this difference between the inclusion relations in the case of shapes and actions disappears, and actions and functions also turn out to be constituted by their parts. Figure 14.5 shows three things, each of which is pointed. When we say that they are pointed, we indicate that they have closed three-dimensional shapes, but we give an unfinished description of these shapes. We leave it open in the description in which way the shapes are closed. When we say that Gavrilo Princip 'killed the Archduke' we indicate in a corresponding way a n action completed in space and time, but we describe only the final phase of the action. T h e description of the whole action is unfinished. It is this which explains why the answer to the question how or in which way the Archduke was killed cannot be derived from this description. If we know a shape to be conical, then we also know which part-shapes the shape can be divided into. But if we only know that a shape is pointed, then we do not know which part-shapes it can be divided into. Actions behave in the same way. O n e does not see this 238

PARTS AND WHOLES WITHOUT INTENTIONALITY

immediately because most accomplishment terms resemble the concept 'pointed' in that they give unfinished descriptions of actions. Even when we have received an answer to the question in which way the Archduke was killed - 'He was shot' — we can continue our question and ask 'How or in which way was he shot?'. The answer to this question is obtained by a new unfinished description - 'With a rifle' - and the accordion effect is generated. The inclusion relation between actions and part-actions is, like the inclusion relation between shapes and part-shapes, univocally determined; but it appears undetermined because actions are most often described by giving unfinished descriptions. The descriptions usually focus one's interest on the final phase of the action; this is because we are not normally interested in the action itself but in its result. Journalists and politicians in 1914 were more interested in the fact that the Archduke was shot than in the fact that he was shot with a rifle. O u r conclusions about accomplishments and achievements in the domain of actions carries over directly to functions. Functions can also be described by means of both accomplishment terms and achievement terms. I also spoke earlier (see page 76) of an 'accordion effect' in the case of functions. The function pumping heat can include the function creating steam pressure and pumping the steam round, and this function can in its turn include the function sucking in steam, compressing it, and pushing it on. These three functions are here described with three accomplishment terms. An achievement term would be, for example, the expression 'to reach the pressure at which the liquid changes to steam'. The upshot of this is that most functions, like actions, are described by means of unfinished descriptions. This fact has to be borne in mind when we turn to an investigation of functional wholes in general. Function universals are very easily misunderstood because the danger of confusing concepts and universals is here much greater than elsewhere. 14.5 F U N C T I O N A L W H O L E S AND T H E I R PARTS We shall now discuss functional wholes, and our example of such a whole will be the refrigerator pictured in chapter 5 (Figure 5.2). It can be represented as in Figure 14.6 with the help of systemtheoretical boxes: 239

ONTOLOGICAL INVESTIGATIONS compressor

r e g u l a t i n g valve

Figure 14.6

The function of the refrigerator is to cool down the cooling compartment, but since this cannot be done once and for all it is better to say that the function of the refrigerator is to pump heat from the cooling compartment to the condensor. According to the latter description the refrigerator performs a series of continually recurring accomplishments. But let us begin by disregarding this, and imagine an apparatus consisting only of the cooling element, the surrounding cooling compartment, the compressor, and that which connects these together. In the cooling element there is a fluid which is converted into steam, and when this occurs it takes heat from the cooling compartment which is thereby made colder. The steam is, with the aid of the compressor, sucked away from the cooling element via the pipe a, which keeps the heat in the steam from returning to the cooling compartment. The whole consisting of the compressor, pipe a, the cooling element, and the cooling compartment performs the function of cooling down the compartment. The function is a temporally and spatially inclusive property which extends itself in space from the compressor to the cooling compartment. T o cool something down always means cooling something down to a particular temperature (just as pumping heat always means pumping a certain quantity of heat), and to cool something down to a certain temperature is an accomplishment. This point is not

240

PARTS AND WHOLES WITHOUT INTENTIONALITY

affected by the fact that temperature is a determinable whose determinates form a continuous scale. In the case of an action such as chopping wood there is a smallest accomplishment, the splitting off of a single piece of wood. To chop wood is to perform this accomplishment over and over. But in the case of the function cooling something down there is neither a smallest unit nor any other given inclusive units, but only continuous temperature changes. This, however, is no argument against cooling down being an inclusive property and an accomplishment. For exactly the same is true of motion. There exists no smallest motion, but every motiondeterminate is an inclusive property. There exists no smallest temperature drop but each specific cooling is an inclusive property, and an accomplishment. In the example we started with we should use 'one cooling down' to refer to the accomplishment consisting of the cooling substance in the cooling element being converted to steam as far as possible. Since for the time being we are ignoring the connection via the condensor, the conversion to steam must cease rather quickly, if only because there is no more cooling substance. Like all inclusive properties, the function cooling something down has parts. The natural spatial parts are the cooling element, the compressor, and their connection (the pipe) respectively. The partfunctions in question are taking up heat by converting the fluid, the cooling substance, to steam (the cooling element), sucking in heat by sucking in steam (the compressor), and leading the heat away by leading the steam away (the pipe). All of the functions, both the whole function and its part-functions, are inclusive in time. When heat is transferred it is moved. Transport of energy conducted is motion, and all motions are temporally inclusive. And sucking, of course, also involves motion in this way. That a fluid changes state, on the other hand, does not imply any (temporally) inclusive jiwt-order universal; states are temporally exclusive. But the actual change is a temporally inclusive second-order universal. Even if it is often easy to demonstrate that functions are inclusive in time simply by drawing attention to the motions involved, these motions should not be thought to be the most essential aspects of the functions. The functions in question cannot be reduced to these motions, or to these motions together with diverse physicochemical things and/or properties. That the compressor sucks in heat is not the same as heat moving into the compressor; that a pipe a 241

ONTOLOGICAL INVESTIGATIONS

conducts heat is not the same as heat moving through the pipe; and that the cooling element takes up heat is not the same as heat moving from the cooling compartment to the cooling element. The difference which manifests itself in the three cases is the same sort of difference as that which has preoccupied writers on 'philosophy of action' between 'He lifts his arm' and 'His arm moves upwards'. The difference between the assertions 'The cooling element takes up heat' and 'Heat moves from the vicinity of the cooling element to the cooling element' is, on the surface, only a difference between two assertions about one and the same state of affairs. The functional assertion points to an active pole, the cooling element, in that which is described. The second assertion describes what happens. But a deeper difference between these assertions is involved. The two assertions describe different ontological levels. The level category presented in chapter 2, will now be put to work. In the presentation of the refrigerator's functions, nothing (either in the present chapter or in chapter 5) has been said about the physicochemical composition of the cooling element or of the other parts. In Figure 5.2 very little indication is given of how the material shape represented there is composed. Such indications would be irrelevant to understanding the functioning. Cooling elements which are completely different from a physicochemical point of view can, in spite of that fact, fulfil precisely the same functions, i.e. be functionally equivalent. The physicochemical substratum is related to functions in the same way as linguistic signs are related to concepts. The functions cannot exist without a substratum, but exactly how the substratum looks is not significant. The substratum can, on the other hand, exist by itself without sustaining any functions. Functions are founded upon physicochemical substrata. That a cooling element can be made of different materials, that it can have pipes of different shapes, and that the pipes can contain different sorts of cooling substance, yet in spite of all these variations perform one and the same function is uncontroversial. What may be controversial is the claim that a function concept corresponds to a universal which is one-sidedly existentially dependent on a substratum level. Might it not be a nominalistic concept which only refers to a disjunction of physicochemical descriptions? I count as physicochemical descriptions also all descriptions of pure motions, including heat motions; the fact that 242

PARTS AND WHOLES WITHOUT INTENTIONALITY

heat moves from the vicinity of the cooling element to the cooling element is a substratum for the fact that the cooling element takes up heat. My position about the concept of function and the corresponding universal is anti-nominalist. The main reason for taking functions to be universals is that they can be added. Just as part-shapes can be added to a whole shape, part-functions can be added to a larger function. The functions 'to take up heat by vaporizing a liquid', 'to suck in vapour', and 'to lead steam' can be added to the function 'to cool off'. And this addition can be directly understood in the same way as one can understand the addition of ultimate shapedeterminates and ultimate mass-determinates. In order to perform the addition one does not require ever to think about the corresponding disjunction of physicochemical descriptions. If the function concepts were really the way the nominalist makes them out to be then they would refer to a disjunction of such descriptions, and the addition I spoke of would have to be thought of in the following way. Assume that we have a disjunction of mass-determinates '5.0 g υ 5.1 g ν . . . ν 6.0g', which are to be added to the disjunction '8.0 g ν 8.1 g ν . . . ν 9.0 g'. The only way in which an addition can be performed is to add in turn one determinate from the first disjunction to one determinate from the second, and in this way to go through the remaining combinations. The sum then becomes a new disjunction of mass-determinates. As I see it, one does not need, in the addition of functions, to make use of such a procedure; and that suggests that functions are true universals. Just as a concept such as the sense of 'blue' cannot be identified with the disjunction of all its substrata, i.e. 'blue', 'azzurro', 'bleu', 'blau', and so on, a function cannot be identified with the disjunction of its physicochemical substrata. This difference in ontological levels explains a number of peculiarities which distinguish functions from universals on the physicochemical level. Functions (and actions) often have a 'by-on' structure or 'by-with-on' structure. A function is performed by something on something else (e.g. the function of taking up heat is performed by the cooling element on the cooling compartment) or by something with something else on a third thing (e.g. the function of taking up heat is carried out by the cooling element with the help of the cooling medium on the cooling compartment). The universal which 243

ONTOLOGICAL INVESTIGATIONS

a function with such a structure exemplifies always extends itself over a number of different things. These things have different characteristics on the two different ontological levels. O n the lower physicochemical level the things are independent of one another. But on the upper level, the function level, it is different. What makes a thing a cooling element is its carrying out certain functions, namely those of cooling something else, i.e. the cooling compartment. It is logically impossible for a thing to be a cooling element if there do not also exist things which are cooling compartments, and vice versa: the substances in question are mutually existentially dependent. This dependence is mediated by the dependence relation between the substances and entities which have an on-off structure (functions and actions). But all of the existential dependences mentioned here exist only on the higher level. O n the lower level the corresponding things with their substances are independent of one another. If one does not realize that two ontological levels are involved, it is impossible to understand functions. T h e remarks just made can be illustrated graphically (Figure 14.7) and illuminated if we use a double-headed arrow to symbolize mutual dependence and a single-headed arrow for onesided dependence (cf. page 38). T h e distinction between function and function substratum which I am trying to clarify is of course intimately related to the distinction made in both functional social science and systems theory between function and structure, but the distinctions do not coincide. There is not space here to go through all the different ways in which social anthropologists, sociologists, biologists, and so on have used a distinction between structure and function. In order to understand the claim that functions imply and cannot be understood without a level-ontology, some of these distinctions

Functional level:

Cooling element

Substratum:

Physico-chemical thing]

^

Cooling compartment

Ψ Figure 14.7 244

Physico-chemical things

PARTS AND WHOLES WITHOUT INTENTIONALITY

must be described and contrasted with the present account. I have chosen to contrast my point of view with that of Nagel in his The Structure of Science. Nagel says that in biology morphology is the study of structures and physiology the study of functions. He writes: Let us first remind ourselves in what way a morphological study of some biological organ, say the human eye, differs from the corresponding physiological investigation. A structural account of the eye usually consists in a description of its gross and minute anatomy. Such an account therefore specifies the various parts of the organ, their shapes and relative spatial arrangements with respect to each other and other parts of the body, and their cellular and physicochemical compositions. The phrase 'structure of the eye' therefore ordinarily signifies the spatial organization of its parts, together with the physicochemical properties of each part. On the other hand, a physiological account of the organ specifies the activities in which its various parts can or do participate, and the role these parts play in vision. . . . If this example is typical of the way biologists employ the terms, the contrast between structure and function is evidently a contrast between the spatial organization of anatomically distinguishable parts of an organ and the temporal (or spatiotemporal) organization of changes in those parts. What is investigated under each term of the contrasting pair is a mode of organization or a type of order. In the one case the organization is primarily if not exclusively a spatial one, and the object of the investigation is to ascertain the spatial distribution of organic parts and the modes of their linkage. In the other case the organization has a temporal dimension, and the aim of the inquiry is to discover sequential and simultaneous orders of change in the spatially ordered and linked parts of organic bodies. 1 For Nagel, a thing or phenomenon's structure is its spatial parts together with the relevant spatial relations, and its function is its temporal changes. This differs in many ways from my distinction. A function can, as I see it, have both spatial parts which are functions and spatial parts which are not functions but function substrata. Similarly, a function can have two sorts of temporal parts: those that are themselves functions and those which are 245

ONTOLOGICAL INVESTIGATIONS substrata. Consider Nagel's example, the eye. O n e can specify how parts of the eye can move without thereby saying anything about which function these motions have. T h e compressor in the refrigerator exemplifies the difference between my functionsubstratum distinction and Nagel's function-structure distinction. A compressor has a certain shape, a certain mass, and a certain colour pattern; and it can be broken down into parts which also have the same sorts of properties. According to Nagel, these parts ought to constitute the compressor's structure; but on my account they belong to the compressor's substratum. I also count as part o f the substratum certain motions, for example the rotation of the crankshaft and the up-and-down motion of the piston. These motions are a sort of temporal change and as a consequence cannot be counted by Nagel as part of the compressor's structure. T h e motions belong to the compressor's substratum; they are necessary but not sufficient for the compressor's function. T h e crankshaft can rotate and the piston go up and down without any gas being compressed (the valves being open), but these motions must be performed if a compression is to take place. T h a t functions exist on an overlying ontological level means that one can speak of functions in two senses: about functions in themselves and about functions with their substrata. Since functions can be added in time and space, this means that one can also speak of addition in two senses. Addition of functions in themselves is always a theoretical operation; the additions are carried out without thinking of any substratum; but without a substratum the functions cannot be instantiated. All real or practical additions are, however, additions o f both substrata and functions. Let us return to the refrigerator, but this time to the function o f the whole refrigerator. T h e refrigerator is a machine which in the terminology o f systems theory contains a 'closed loop' (see Figure 14.6). I f we go round the loop we can describe the refrigerator in the following way. T h e cooling element takes up heat from the cooling compartment via the vaporizing of a fluid in the cooling element. Heat/steam is led via pipe a to the compressor, which pumps heat/steam from the cooling element and a to pipe b, at the same time as it increases the pressure and temperature o f the steam. T h e heat/steam is led to the condensor via b. T h e condensor gives off heat by allowing the steam to condense to a liquid, and the liquid

246

PARTS AND WHOLES WITHOUT INTENTIONALITY cooling

refrigerating element I 'o

t td Figure 14.8

is led via с to the regulating valve. With the help of the regulating valve the temperature of the liquid can be lowered, whereupon the liquid is led by pipe d to the cooling element, i.e. to our starting point. In the above description we followed a process in time and space, a process where difTerent functions are added to one another. The cooling element, the compressor, the condensor, and the regulating valve, as well as the connecting pipes between them, exist at different places in space, and must be added in space in order for a spatial whole to exist. But in the description we also followed the change of a volume of liquid over time - it was vaporized, moved, it changed temperature, and was condensed. The function which the whole instantiates is a universal which is inclusive in both time and space. It could be reproduced graphically as in Figure 14.8. W h a t happens when this open system is closed to form the system constituted by the refrigerator (Figure 14.6)? When one comes back to the starting point in the line of reasoning, to the cooling element, it is important to see that, due to the system's being closed, the cooling element is able to perform its function not only during the period of time when we were there in our thoughts, but also when we were in the second element. Moreover, one must bear in mind that precisely the same thing holds for the other elements, i.e. all the elements perform their functions during the whole time fa-t&. It is only when one sees this that one grasps the temporally and spatially inclusive universal that the refrigerator is a manifestation of. T o understand the principle for a machine is fundamentally, as I pointed out earlier, the same as to understand how an inclusive universal (like a certain shape) is constituted. But 247

ONTOLOGICAL INVESTIGATIONS

inclusive properties can be more or less complex. There is in this respect a great difference between shapes and functions, as well as between functions of open as opposed to closed systems. When one has understood the refrigerator principle in the way I have just described, one has understood it in abstraction from its substratum. Understood in this way, the refrigerator becomes an endlessly repetitive system. The cycle in question can be repeated over and over again. But if one takes account of the substratum level, this possibility of endless repetition disappears. The physicochemical substratum wears out. The cooling element can break down and the cooling medium can leak out; the compressor's valves can wear out and the compression disappear, and so on. Once again we see the importance of keeping the different ontological levels separate. Ever since Kant's Critique of Judgement, countless philosophers have been concerned with analyses of mutual means-end relations (reciprocal action); an example would be the activities of the heart and the lungs. 2 Each is simultaneously a means for the functioning of its counterpart. The same kind of mutuality can be found in a machine like the refrigerator. The cooling element takes up heat and makes the condensor receive heat which it can give off; but at the same time the condensor's giving off of heat functions as a means enabling the cooling element to take up heat. The condensor and the cooling element function as means and ends for one another. That they are means for one another does not imply that they are necessary or sufficient conditions for each other's function, but only that they are contextual conditions. The contextuality which manifests itself here and which has perplexed many philosophers has nothing mystical about it at all. The type of mutuality which can hold between functions is fundamentally the same type of mutuality which obtains when two shapes fit one another completely, for example when a ball fits precisely into a hollowed sphere. In this last case we have a mutuality between properties which are inclusive only in space. In the case of the cooling element's and the condensor's functions, we have the same type of mutuality, but here between entities which are inclusive in both time and space — properties moreover which belong to an overlying ontological level. If one accepts the existence of temporally inclusive universals mutuality is easily understood. 248

PARTS AND WHOLES WITHOUT INTENTIONALITY

pipe

speed regulator

Figure 14.9

The type of additions of inclusive spatio-temporal entities which I have concentrated on, gives rise to other peculiar effects beside the mutual means—end relation. One well-known but misunderstood effect is the feedback mechanism. 3 As an example of such a mechanism I shall take the steam engine's classical speed regulator, the centrifugal regulator. It is reproduced in Figure 14.9. The function of the steam boiler is to create steam which is then led to the actual steam engine, whose function is to transform heat energy in the steam into mechanical energy in the form of a rotating crankshaft. In order to obtain a constant velocity of rotation of the crankshaft, a system with a speed regulator is attached. The function of the speed regulator is thus to hold the velocity constant. It does this by transferring information from the crankshaft to the steam pipe between the boiler and the engine. How this occurs in the case where the regulator is a centrifugal regulator is shown in Figure 14.9. The rotation of the crankshaft is 249

ONTOLOGICAL INVESTIGATIONS

carried over to that of the regulator's axle with the help of a bevel gear (1). When this axle rotates, the two pendula (2) are thrown outwards, and the regulator sleeve (3) is raised to a certain position. The regulator sleeve affects the steam valve (5) via the lever (4); the valve opens or closes the supply of steam, which increases or decreases the rpm of the steam engine. T h e centrifugal regulator is a mechanism whose function is completely based on motion. The bevel gear transfers the crankshaft's rotation to the regulator axle. T h e transfer function is normally carried out in this way, but might also have been carried out in other ways. Similarly, the other constituent elements' functions can be carried out in many different ways but are carried out here in one particular way. Since all of the regulator's functions are carried out with the help of mechanical motions one can easily see that the reduction problem for inclusive universals described earlier (section 7.2) arises here too. All five of the regulator's constituent elements (the regulator axle, the pendula, the regulator sleeve, the lever, and the steam valve) have, at any given moment, a particular position, a particular velocity, and a particular acceleration. If one knows the value of these variables over a certain period of time, one can, with the help of integration, obtain the motion of the elements, i.e. properties which are temporally inclusive. But such an integration procedure does not give us any knowledge about whether the motions in question can be combined in the way in which they are combined in the centrifugal regulator. O n e sees that this is possible only when one sees that the motions, i.e. the inclusive properties, can be connected (added) to one another in time and space in this specific way. And such an insight does not demand knowledge of the corresponding exclusive properties. Functions require substrata and in reality always involve two ontological levels. When we speak of functions we sometimes speak of both the functions and their substrata, and sometimes of the functions in themselves. If the above example had been concerned with functions in themselves, then we could have restricted ourselves to considering the additivity of the functions. But the description given seems also to be an account of causal pathways, since the substratum is brought into the description; and on the substratum level we have forces and efficient causality. To talk of the motions (in the feedback mechanism constituted by the 250

PARTS AND WHOLES WITHOUT INTENTlONALITY

centrifugal regulator) in themselves is to refer only to facts such as that the increased velocity of the crankshaft fits an increased velocity of the regulator axle, that this fits the raised pendula and raised regulator sleeve. This last-mentioned motion can thus be seen to fit in with the sinking of the right-hand side of the lever and with the pushing down of the steam valve. This 'fitness description* might just as well have been taken in the reverse order. We could have started with the downward motion of the steam valve, and thus have seen that it fits in with the right-hand side of the lever moving downwards, and so on until we came to the crankshaft moving faster. T o say that the motions in question fit each other is not to say that the one causes the other. Whether this occurs depends on the substratum. If the velocity of the crankshaft increases but the attachment of the regulator axle to the bevel gear slips, then the velocity of the regulator axle does not increase; if the lever becomes loose in its attachment, then the regulator sleeve can move upwards without the lever having moved downwards, and so on. The mutuality relation I claim to exist in the feedback mechanism might seem difficult to reconcile with the observation that one cannot begin to describe the mechanism anywhere one pleases. The normal sequence is for the description to commence with an equilibrium situation, i.e. a situation where the lever and steam valve do not move at all. After that, one begins by imagining an increase in the velocity of rotation of the crankshaft, and then imagines that this change reproduces itself — in different ways via the elements in the feedback system - in the steam valve, which, by being pushed downwards, strangles the supply of steam and lowers the crankshaft's rate of rotation to the equilibrium state. This sort of account is straightforward. But to begin the thought experiment with the steam valve's being pushed down seems to be much more difficult. Such a change cuts off the supply of steam and leads to a lessening of the crankshaft's velocity of rotation. This then leads to a lowering in the velocity of the regulator axle, which makes the pendula sink. When, as a result the regulator sleeve sinks, the lever arm which controls the steam valve is raised, and the steam valve is pushed upwards. If we conceive of the whole as a momentary change it becomes impossible. The steam valve cannot at one and the same time be pushed both upwards and downwards. This problem is solved by allowing causal reasoning, i.e. reasoning which, among 251

ONTOLOGICAL INVESTIGATIONS

other things, makes use of the category of tendency. In order for the steam valve to be able to be moved downwards, the lever and pendula must move upwards, but the position of the pendula should also fit in with the velocity of the regulator axle. What happens is that tension arises in the lever, i.e. the lever receives two tendencies, one to go downwards and one to go upwards, given that it does not bend. The tendency concept actually exists implicitly also in the description which begins with a raising of the crankshaft's rotation velocity. This increase led, as we saw, to a lowering of the velocity of rotation, which means that if all changes are momentary, then the changes under discussion are impossible. What makes the process imaginable as a causal one with stepwise changes is a tacit assumption that the crankshaft can increase its velocity without there being any change in the amount of steam streaming into the steam engine. It is presupposed that there exist at least two factors which can in co-operation determine the velocity of rotation, which in its turn means that the factors in question must be some kind of tendency. The feedback mechanism described here is a so-called negative feedback. Here, the purely functionalist description takes on the appearance of an 'impossibility theorem': It is impossible to change the crankshaft's velocity in the functional system described. But if one also takes account of the substratum level, then the relevance of traditional causal considerations becomes apparent. If we change the feedback to a positive feedback, which we can do by imagining the lever lengthened and fastened to something on the right-hand side of the steam valve, then we can say the following: An increase in the crankshaft's rotation velocity affects the lever which raises the steam valve, which makes the stream of steam increase and the crankshaft increase its velocity even more. This in its turn makes the steam valve let out even more steam, and so on. Here, causal reasoning is once again involved, but even here it can be removed in favour of a purely functional description. Such a description would show that it is possible for the crankshaft and regulator axle to increase their velocities at the same time as the pendula, regulator sleeve, lever, and steam valve move upwards, and that this can continue to a certain extent until one of the motions is blocked. We obtain a 'fitness theorem' of the same sort as when we point out that it is possible for certain shapes to combine without any space between them. 252

PARTS AND WHOLES WITHOUT INTENTIONALITY

The type of diagram which I have used to demonstrate the closed loop in the refrigerator and the feedback mechanism in the centrifugal regulator, blossoms everywhere within cybernetics and systems theory. Parts of my reasoning are clearly also identical with system-theoretic reasoning. The difference between systems theory and my account is pardy that the former does not have a level-ontology (and consequently does not distinguish between function and substratum) and partly that writers on systems theory have not understood the distinction between universals which are inclusive and exclusive in time (and consequently do not see the specific type of addition which their systems involve). Systems theory claims to be a general theory of systems and the relations between parts and wholes. But apart from a number of meritorious classifications of the type I have made use of: 'negative feedback', 'open system', 'closed system', and so on, it has in no way met the expectations it at one time evoked. Instead of a fruitful theory which is applied to more and more new areas, systems theory seems to be a set of general indications about interaction, supplemented with illustrations from the most diverse fields. All the illustrations have the character of box diagrams with arrows between them. The lesson which I believe the development of systems theory teaches us accords well with my view of theories of categories. Like the system theoreticians, I try to say something general about the part-whole relation. But I think that what I am saying on the categorial level is of such an abstract nature that it gives at best general clues to the structures dealt with by the special sciences. There probably exists an unsurveyable number of spatio-temporally extended universals, and the study of these specific universals and their relations to one another must be studied in a number of different special sciences. On the categorial level one can only argue for the actual existence of this type of universal and for the actual possibility of defining addition operations on them. On the level at which systems theory tries to place itself, one can only study the various universals discovered in various areas, see similarities and differences amongst them, and classify them under nominal concepts of the type I mentioned ('negative feedback', 'open system', etc.). It is often said that 'function' has two senses, the 'effect sense' and the 'purpose sense'.4 In the former case, functions are 253

ONTOLOGICAL INVESTIGATIONS

understood as a sort of cause; in the latter, as a form of intentionality. According to the present analysis, functions are neither the one nor the other. Functions are universals which are temporally inclusive without either striving for a goal (the 'purpose sense') or bringing about an effect (the 'effect sense'). The traditional discussions about functions lack the concept of temporally inclusive universals, and can thus not see the possibility of a function concept which is tied neither to purposes nor to causes. This chapter began by setting aside the intentionality category, which is why the examples used have had nothing to do with purposes. The category of causality which has been presented (see chapter 12) presupposes the tendency category, but tendencies are always temporally exclusive and cannot be functions, which are temporally inclusive. There can at best exist causes which explain functions in the same way as the temporally exclusive Newtonian forces, acceleration, and velocity can explain temporally inclusive motions. One can say that functions are 'motions', but 'motions' on overlying levels. Functions are, like motions, extended in time, without it being the case that a later motion/'motion' is brought forth by an earlier, or that a later motion/'motion' is the purpose of an earlier one. The main claims and conclusions of this section are: (1) Functions belong to upper levels; (2) functions are inclusive in both space and time; (3) machines, organisms, and many other natural processes are systems or 'aggregates' of functions; (4) such systems can be understood (as when one understands 'the refrigerator principle') without knowing anything about their substrata. These conclusions are likely to seem counter-intuitive as long as one does not remember that most functions are described and 'added' in everyday language under unfinished descriptions (cf. 14.4). This fact means that it is possible to be familiar with a machine or organism principle without knowing all the ways in which it can be realized.

254

PARTS AND WHOLES WITHOUT INTENTIONALITY

14.6 FUNCTIONAL WHOLES AND NORMATIVE PRINCIPLES The views I have put forward in the preceding section are in several respects similar to views of Michael Polanyi. He has been one of the chief advocates of the view that machines and organisms manifest principles and so cannot be reduced to physical and chemical laws. Technology and biology have to do with other types of universals ('operational principles') than physics and chemistry.5 That is my opinion too. But our views do not coincide. Polanyi has a level-ontology where operation principles refer to upper levels, and of course I agree with him on this point. But Polanyi believes that knowledge about certain levels must have a normative character. According to Polanyi, every machine is defined through its 'operational principle' and its aim. The latter is built into the operation principle itself, which at one and the same time is normative and gives teleological explanations. Knowledge of machines can thus not be scientific knowledge. The cleft between knowledge and technology is retained. And the same reasoning, according to Polanyi, applies just as well in the case of organisms. We must take a normative standard and see all organisms as goals in themselves: Once more, as in morphology and morphogenesis, the existence of every living being is acknowledged as an aim in itself; however nasty a flea or liver fluke may be to us, we recognize the rational functioning of its organs in their own interest. The purely scientific interest of physiology depends, therefore, ultimately on the passions which make us pursue Natural History. It relies on the passions which account for the importance which we attribute to a living being in itself; on its intrinsic interest, and our contemplation of it as it is and as it ought to be.6 As regards the operation principles for machines, Polanyi writes: The conceptions of machines in good working order form a system which ignores the particulars of failures - in the same way as geometrical crystallography ignores the imperfections of crystals. The operational principles of machines are therefore 255

ONTOLOGICAL INVESTIGATIONS

rules of tightness, which account only for the successful working of machines but leave their failures entirely unexplained. 7 The principle of the refrigerator shows how a refrigerator works — to return to my example — but it does not show why a certain refrigerator does not work. It presents the norm for the refrigerator - this is how I think Polanyi's view can be represented. I think that Polanyi is here guilty of circular reasoning. Given the interest in creating a cold space, the corresponding operation principles are normative, i.e. they are normative in the sense of being parts of the interest. But apart from the interest, the principle in itself is only about cold spaces. In other words, if the operation principles are seen as norms, then there must also exist a corresponding interest whereby the principles (can) appear as teleological. Polanyi's mistake is similar to the one made by the thinkers of antiquity when they took certain shapes to be mere deviations from others, which were taken to be norms. Such an understanding is partly grounded in the real difficulty of getting hold of all of the shape-determinates (see section 4.1). O u r limited cognitive faculties make it natural to choose a small number of shapes which we try to understand first, and then we try to understand the other shapes in the light of the first. The half circle, for example, is first understood in relation to the whole circle, not in itself. It is as though the half circle in some way is a deviation and lower form of the whole circle. But with increased mathematical knowledge, this peculiarity disappears, and we accustom ourselves to seeing all shapedeterminates as being on an equal footing. The same is true of machines and organism principles. The universale which the latter refer to are difficult to grasp clearly, which means that, when we have a grasp of some of them, we try to see other universals as deviations from them. In this way we are led to see refrigerators as 'better' or 'worse', as corresponding to or falling away from an ideal norm, the refrigerator principle, instead of being just different universals. But even machines which run, but do not work in the sense that they fulfil their aims, instantiate universals. Even if we do not concern ourselves about these spatio-temporal inclusive universals (just as little as people earlier concerned themselves with peculiar shapes), they are there all the same. Polanyi says that an operational principle cannot explain mistakes and breakdowns of machines, and that is of course true. 256

PARTS AND WHOLES WITHOUT INTENTIONALITY

But this is no more peculiar than that a circle shape cannot explain a square shape. They are simply two different shapes. And in the same way a functioning and a broken machine quite simply instantiate two different four-dimensional universals. Polanyi also maintains that operational principles do not appear at all in physics and chemistry, and this opinion is in a way a logical consequence of his view that these principles are normative. But since the latter view is mistaken, we can raise the question whether some sorts of operational principles really do not appear in physics and chemistry. In fact, they do. To understand a mechanism is to understand an operational principle; in other words, some operational principles refer to mechanisms. And physics and chemistry will never be able to do completely without mechanisms.

257

15

MAN AND SOCIETY: NESTED INTENTIONALITY

In the introductory chapter I described my theory of categories as, among other things, an attempt to transcend the opposition between atomism and holism. I shall now repeat and summarize what is of immediate interest in relation to such a distinction before tackling the main theme of the present chapter: nested intentionality. 15.1 FOUR CONCEPTS O F 'PARTS' As regards the question of atomism and holism, one can begin by distinguishing four different basic senses of 'part'. The first three are: (1) part as element in a set, (2) part as moment of a whole, and (3) part as a spatio-temporal part. Part-whole relations in the first sense lack ontological interest and are mentioned here only for the sake of completeness. The element-set relation is either a purely linguistic creation as when one constructs a set of completely arbitrary objects, for example the set consisting of the sun, the first word on this page, and the chair I am sitting on. Nothing other than their being members of the same set unites these three elements in a 'whole'. Or the relation may be based on something which truly exists and which unites the elements in question, for example a certain property such as when one constructs the set of all red things. Ontologically interesting whole-part relations can exist here, but they concern the universale used to construct the set, not the actual relation of belonging to a set. That something is part of a whole in the sense of a moment means that the part cannot exist without being a constituent of a more comprehensive whole. Such a part (aspect) is thus not, as atomism would have us believe, existentially primary in relation to 258

NESTED INTENTIONALITY

the whole. On the other hand, neither can the comprehensive whole in question exist without (at least some of) its moments. Part and whole depend on one another. In this sense even Democritus's indivisible atoms have parts. A spatially indivisible corpuscle/atom has as moments (parts) both shape and volume. Without these parts, the atom could not exist. There is no absolute atomism. No matter how the atoms in an atomistic world view are specified, they must contain moments, i.e. one sort of part. It is true that it is only philosophers within some traditions who have been happy to talk of 'moments' as 'parts' but this is a merely verbal matter. The important point is that atoms and moments are connected together. There are many sorts of parts in the third sense, spatio-temporal parts. One thinks first, perhaps, of aggregates of things. One can also distinguish spatio-temporal parts amongst instances of single universals. Most interesting in this context are the inclusive universals example. We must also remember the distinction between parts and lower-dimensional parts. Normal parts are parts with the same number of spatio-temporal dimensions as the viniversal in which they are included; lower-dimensional parts are those which have fewer dimensions. With regard to normal parts the whole is one-sidedly existentially dependent on the included parts. The situation is more complicated as regards lowerdimensional parts. In at least certain cases there is a relation of mutual existential dependence: two-dimensional surfaces cannot exist other than as lower-dimensional parts of three-dimensional volumes, and of course three-dimensional volumes cannot exist if they do not have two-dimensional surfaces as lower-dimensional parts; they have to be limited by such surfaces (see also the discussion on page 88f.). Patterns and Gestalten have spatiotemporal parts and addition can be defined even for these parts. There is no reason to employ the expression 'the whole is more than the sum of its parts' in connection with patterns and Gestalten in so far as the parts in question are equi-dimensional parts. When one looks at lower-dimensional parts then it is always true that 'the whole is more than the sum of its parts'; lowerdimensional parts cannot, by definition, via addition result in a whole with an increased number of dimensions. The claims I have put forward about parts in the sense of moments, and parts in the sense of normal and lower-dimensional 259

ONTOLOGICAL INVESTIGATIONS spatio-temporal parts, go beyond w h a t traditional atomists usually accept without, for this reason, coinciding with w h a t holists usually say. Atomists usually do not accept, for example, internal relations, i.e. m u t u a l existential dependence, a n d holists usually d o n o t accept one-sided existential dependence. And neither atomists nor holists have clearly understood the existence of inclusive universals. In spite of this, w h a t I have said so far does not lead to a radical break with the atomistic viewpoint. Spatio-temporal parts a r e ontologically p r i m a r y in relation to t h a t type of spatio-temporal whole (things, aggregates of things, machines, organisms, a n d n a t u r a l processes in general) which they constitute. T h e radical break comes with the existence of p a r t s in the fourth - as yet u n n a m e d — sense. Parts in the sense of moments are based on the category of existential dependence, a n d spatio-temporal parts are based on the category of container space. Parts in the fourth sense, however, a r e based on the category of intentionality. T h e whole of which such p a r t s are parts is an intentional whole with the two features: 'pointing' and 'connection at a distance'. Parts in the fourth sense I shall call 'participants' — these parts participate in a whole. O n e example of parts of this last-mentioned type are t h e m e m b e r s in a n organization. T o be a m e m b e r of a n organization is not the same as (1) being an element in a set, or (2) being a m o m e n t , or (3) being a spatio-temporal part of a n organization. O n e can of course construct the set of all m e m b e r s of a n organization, b u t this set is not identical with the organization. A n organization exists in space and time, while a set is an abstract entity. T o be a m e m b e r of an organization cannot, either, be the s a m e as being a m o m e n t of the organization, since the organization can persist even if one m e m b e r has died, and vice versa. Both the m e m b e r s a n d their organization exist, in space a n d time, b u t this does not imply that the relation 'to b e a m e m b e r o f ' is identical with the relation 'to be a spatio-temporal part o f ' . T h e spatial inclusion relation is, for example, transitive, but the m e m b e r s h i p relation is not. If a is a spatial p a r t of b, a n d b a spatial p a r t of c, t h e n a must be a spatial part of c. B u t if α is a m e m b e r of (the organization) b, a n d b is a m e m b e r of the organization c, then it does not follow that a is a m e m b e r of c.1 O n e thus sees, as a result of simple considerations, that the m e m b e r s h i p relation is a very specific part—whole relation. 260

NESTED INTENTIONALITY

Similarly, it is easy to see that the intentionality category here plays a decisive role. In order for someone to be a member of an organization, he must be seen and conceived as a member, if not by himself then by someone. Membership without some sort of intentionality does not work. The intentionality category can, under certain conditions, and like the space category and the existential dependence relations, function as a 'unifying category'. Just as existential dependence, by linking substances and properties, gives rise to the new category of state of affairs, so too, as we shall soon see, intentionality gives rise to nested intentionality. There is intentionality of two kinds: presentational and representational. It is presentational intentionality which forms the basis of the most interesting part—whole relations, but representational intentionality also gives rise to special types of wholes. Assume that five people who know each other but live on different continents think of each other at the same time. Person A thinks of В, C, D, and E; person В thinks of A, C, D, and E, and so on. Obviously some sort of unity among the five persons is created; a unity that goes beyond the external spatial relations which naturally connect them and make them constitute a pure spatial pattern. The example is trivial in the sense that normally we encounter similar situations all the time. It obtains philosophical significance through the analysis of the intentionality category that I have carried out. I would like to stretch Leibniz's terminology a little and say that representational intentionality can give rise to 'monadological wholes'. In the above example we have five persons (with spatial bodies, that is to say, they exist in container space), all of whom have an intentionality which is founded upon their bodies. Each intentional state or act points to the other persons, but each intentional phenomenon is spatially distinct from the others. Each person becomes a monad that mirrors the others. T h e situation is quite different in the case of presentational intentionality, since I have argued that in such cases we can be in direct contact with the intentional goal. Intentionality and its correlate here make up a unit and a wholeness in space that goes far beyond purely spatial relations; we get a much stronger type of connection than that which gives rise to 'monadological wholes'.

261

ONTOLOGICAL INVESTIGATIONS

15.2 AGENTS AND SUBJECTS T h e whole of which participants are parts is always a whole consisting of at least two participants who are subjects. A subject is a spatio-temporal whole consisting of a body which has, or has a disposition to have, intentional acts (cf. section 13.1). Before commencing on a discussion of this concept of 'participant', a few points should be made regarding both the subject category as such and the intentional connection as such. We shall analyse some intentional phenomena which only include one subject and an intentional correlate which is an object or a naturalistic state of affairs. Certain emotional and intentional states, like happiness, sadness, and fear can have naturalistic states of affairs as intentional correlates, and the intentionality in question can be presentational. I can be happy about the sunrise, sad about the rain, and frightened about the avalanche which is approaching me. T h e sunrise, the rain, and the avalanche exist independently of my perception of them, but they nevertheless become connected to me - and I to them - through perception. We become connected by intentionality. 2 The claim that these states are intentional does not contradict our earlier analysis of states in general (sections 7.3 and 11.2). A state of happiness, sadness, or fear contains as do other sensory states, a tendency to carry out certain actions. Emotional states like intentions (see pages 106-7 and 208—9), contain both intentional acts and tendencies to act in certain ways. Subjects who are in emotional states like happiness, sadness, and fear exist in the space-time taken up by their substrata. T h a t which the emotional state is about is, in the case of presentational intentionality, in some sense a part of the state; but due to the intentionality 'hop' in space, it would be mistaken to call it a spatio-temporal part of the subject. The most adequate description possible is to say that the subject and the naturalistic state of affairs ( = the intentional correlate) are connected by intentionality. Normally we understand emotional states as something we are affected by, not as something we choose. One is not an agent simply because one has conscious experiences of happiness, sadness, fear, and such. T h e agent category does not coincide with the subject category; agents are a speacial sort of subject. These subjects, 262

NESTED INTENTIONALITY

however, play an important role in what follows and so a few words should here be said about the agent category. It is best understood by means of the concept of rationality and rational action. Like happiness, sadness, and fear, rationality involves intentionality. Whether it also includes tendencies and is a sort of intentional state is a question which will have to be set aside for the moment. To be rational is to be rational in a situation or to be capable of being rational in a situation. The one who is rational must have his situation as an intentional correlate. In order to make the issue more precise, I shall return to chapter 5 and the presentation of methodological individualism. That I make use of this '-ism' as an example is of no great consequence. The type of explanation pattern which I presented under the label 'methodological individualism' can be found in such widely disparate traditions as neo-Kantianism, hermeneutics, existentialism, and marxist praxis-philosophy. Here, there is a hidden meeting-place for apparently contradictory traditions. But this is not strange. All these traditions take as their starting point the everyday category of agent, i.e. the fact that we often act after having made a choice in a situation. Noretta Koertge has suggested, as a first approximation, a model of methodological-individualistic explanations. 3 I shall not reproduce the model in order to comment on Koertge's analysis, but because the model provides a good background to the remarks I should like to make. This is how the model looks: (1) Description of the situation: Agent A was in a situation of type C. (2) Analysis of the situation: In a situation of type C, the appropriate thing to do is d. (3) Rationality principle: Agents always act appropriately to their situations. (4) Explanandum: (Therefore) A did d. In the model the explanandum is a logical consequence of the first three points. I shall go through them one at a time, and begin at the bottom with point three. In the formulation that the principle of rationality has received here, it is obviously false. Normally we are affected by diverse factors which often lead to our irrationally diverging from the reasonable action we had decided on. Agents only seldom act completely adequately in relation to their 263

ONTOLOGICAL INVESTIGATIONS

situation. Many of the philosophers and scientists who have advocated this model of explanation have pointed out how frequently such irrational divergences occur. 4 Rational actions, they have pointed out, must be seen as an ideal type. The above model must be supplemented with a clause saying that no irrational factors play a role in the situation in question; otherwise the explanandum cannot be derived. The principle of rationality also contains an unclarity. As it has been formulated, it leaves a categorial question open: Do agents always act rationally because they have a tendency to act that way (a tendency which can then complement or counteract irrational factors); or do they always act rationally because they spontaneously (in the sense of the spontaneity category of chapter 7) choose rational actions? In the formulation given here, the rationality principle says only that agents act rationally; but their actions must be either spontaneous or occur causa stii and be conditioned by an inertial tendency. In the latter case rationality can be put on equal footing with whatever intentional state one pleases. Rationality becomes a special kind of emotional state. The majority of those who have advocated a rationality principle seem to have had the former case in mind. One takes for granted that the agent could have acted otherwise. He could have acted otherwise by choosing to try to realize another intention, and that choice or decision is to be understood as a change which falls under the spontaneity category. I shall quite simply define the agent concept in such a way that an agent is a subject whose intentions (tryings) could spontaneously have been different. Here, everything turns on the category of spontaneity that was introduced in section 7.1. An agent is such that he can have a spontaneous intention to do something. This means that it is impossible to predict with certainty an agent's intention: one may only speak about normal intentions. Of course, these intentions also have a tendential side, which explains the realization of the actions and makes possible their combination with irrational factors. The rationality principle should therefore be reformulated in the following way: (3) Rationality Principle·. (a) Agents normally decide to act in a way appropriate to the situations in which they find themselves. (b) Agents can be affected by irrational factors. 264

NESTED INTENTIONALITY

As point (2) (Analysis of the situation) has been formulated, it presupposes that the rationality of the situation is independent of the agent. No agent is even mentioned. But this must be wrong. Let us once again look at the classic example of the consumer in the market - more particularly, in the market described earlier (see page 62fT.) where there are only two commodities at given prices, food baskets and fun baskets. We can also regard the quantity of money the agent has at his disposal as part of the situation. But these facts do not tell us in any way what the agent's rational purchasing activities in the situation will look like. In order to find out about that, we must also know something about the agent, for example that he is something of a hedonist and has well-defined indifference curves. One must be acquainted with the agent's preference system. If this has a certain character, then a certain way of acting is rational in the situation; if it has some other character, then some other way is rational. Two sorts of agents can find themselves in the same situation without the same sort of action being rational for both of them. Point two must be given the following formulation: (2) Analysis of the situation: In a situation of type C, the appropriate thing for an agent of type A to do is d. An action is thus never rational in itself, but rational for a certain sort of agent in a certain sort of situation. But more needs to be said about the agent. What relation is there between the agent and the situation? Points (1), (2), and (3) above say nothing about this. Point (1) says, for example, only that (given the modification just introduced): 'Agent a of type A was in a situation of type C.' But the answer to the question about the relation of mediation between agent and situation is not hard to find. It is the intentionality category which supplies this relation - i.e. usually perceptions, but also other intentional acts and states. Point (1) must be rewritten in the following way: (1) Description of the situation: (a) Agent a of type A was in a situation of type C. (b) Agent a was intentionally connected with situation с of type C. It is not sufficient that a is in the situation c, he must also have 265

ONTOLOGICAL INVESTIGATIONS

the situation as his intentional correlate. We can now describe the consumer example in the following way: (1) Description of the situation·. (a) Agent a of type A (hedonist with specified indifference curves) was in a situation с of type С (= found himself in a market with only food baskets and fun baskets at given prices, as well as having a sum of money to spend). (b) Agent a correctly apprehended с as С and correctly apprehended himself as being an A. (Agent a: 'Given that I am a "foodie" and see a jar of päte de foie gras in the food basket what shall I do?') (2) Analysis of the ntuation: In a situation of type C, the appropriate thing for an agent of type A to do is d ( = buy χ food baskets andj> fun baskets). (3) Rationality principle-. (a) Agents normally decide to act in a way appropriate to the situations in which they find themselves. (b) Agents can be affected by non-rational factors. (4) Closure clause·. (a) a decided to act rationally (b) a was not affected by non-rational factors in c. (5) Explanandum: (Therefore) a did d. In the structure given here, certain things are easily seen. The remarks made about irrational factors and about spontaneity lead to a completely new premise, point (4), and that in its turn makes point (3) superfluous. Thus we see by chance why actions and explanations of actions in many traditions are understood as unique and as more bound to individual situations than are explanations in the natural sciences. Point (3) is universal, but point (4) is determined by the situation. One can also see that a's considerations (i.e. some of a's relevant mental actions), which lead to his analysis of the situation, are not included as particular actions in the schema. They ought to be, but their omission does not affect the conclusion I wish to draw with the help of this schema. The conclusion is that the agent concept does not require us to introduce any part—whole relations over and above those required 266

NESTED INTENTIONALITY

by the subject category. The intentionality category is already a defining characteristic of the subject, and the spontaneity category does not lead to any new part-whole relations. Before we get involved in some of the part—whole relations based on intentionality, a few words should be said about misintentionality and contradictory intentionality.

15.3 MISINTENTIONALITY AND CONTRADICTORY INTENTIONALITY In the example of the previous paragraph the satisfaction modality of intentional acts and states always had the value 'satisfied'. We shall now see what happens if we allow it to receive the values 'partially satisfied' and 'unsatisfied'. The interesting cases involve agents. Then we sometimes get what can be called the 'glass-door effect'. If we are walking along a corridor which we perceive correctly until we come to a glass door which we do not see at all, then we walk into the door. Such an action or occurrence can only be explained with the help of misintentionality, in the present case a misperception. We walked into the door because we did not see it. The explanation has a structure which links up with the structure of the position of methodological individualism, but which extends it as follows: (1) Description of the situation·. (a) Agent a of type A was in a situation с of type C. (b) Agent a was intentionally connected with с as being of type C 1 and with himself as being of type A. (2) Analysis of the situation: (a) In a situation of type c, the appropriate thing for an agent of type A to do is d. (b) In a situation of type C 1 , the appropriate thing for an agent or type A to do is d 1 . (3) Rationality principle·. (a) Agents normally decide to act in a way appropriate to the situations in which they find themselves. (b) Agents can be affected by irrational factors. (4) Closure clause: (a) a decided to act rationally. (b) a was not affected by non-rational factors in c. 267

ONTOLOGICAL INVESTIGATIONS

(5) Explanandum: (Therefore) a tried to do d 1 . In the glass door case, С is to be equated with the corridor as it actually is, i.e. with a glass door. And the rational action (d) for the given person is to go through the corridor and open the glass door. C 1 , on the other hand, stands for the corridor without a glass door, and d 1 is the action of simply walking through the corridor. In the schema itself there is only the conclusion that a can try to walk straight through the corridor. But actually the schema contains more. С and d 1 can be compared. It is this comparison which leads to the conclusion that a is going to collide with the door. However, there need not always be a collision, literally or metaphorically, between an intended action and a misperceived situation. Let us look at the example of the consumer and assume that the agent misunderstands the prices, i.e. has an act that is characterized by misintentionality. He then buys ζ food baskets and и fun baskets, which is rational in a situation of type C 1 ; not χ food baskets and у fun baskets, which would be rational in a situation of type С. A comparison of d 1 (the buying of ζ food baskets and и fun baskets) and С can be made, and its result is that d 1 is perfectly feasible. The conclusion here is not that d 1 is impossible, but that a performs a nonoptimal act because he misunderstands the situation. T h e comparison between d 1 and С is a comparison of a potential action with an actual situation. The type of universale which actions and naturalistic situations instantiate has been discussed in earlier chapters. The type of comparison that is made between d 1 and С is not a categorially new type of comparison. It leads to the same type of possibility or impossibility theorems as was discussed in the last chapter. In the same way as we can see that a ball with a 2-cm diameter cannot go through a hole with a 1-cm diameter, and can see that in a thermostat system with negative feedback the temperature cannot vary, so we can see in the glass door case that a collision is going to occur. In the consumer example we can see that a non-optimal action will be carried out. T h e comparison between the real situation С and the situation which is the (not yet realized) intentional correlate is of course a comparison made by an external onlooker. The agent himself only acts, and then the collision just happens. The agent cannot make the comparison made by the external onlooker. 268

NESTED INTENTIONALITY

Presentational intentionality is in several senses primary, but representational intentionality is just as real. If we assert that at the same time both p and not-p is the case, where p is a naturalistic state of affairs, then we are asserting something contradictory. The assertion itself (the representational intentionality) occurs, but it is logically impossible for its intentional correlate to exist. Contradictions do not exist, but intentional acts directed to contradictory states of afTairs do occur. When I now speak of contradictions I mean logical contradictions in the extended sense of logical I earlier (in section 8.2) described. Logical insights are not confined to insights which are based on formal—logical relations. There is material logic or existential dependence as well; one case is provided by the inclusion relations between inclusive determinates of one determinable. In this sense, one has a logically impossible intentional correlate if one believes that a 2-cm ball can go through a 1-cm hole, or that a volume of 5 cm 3 includes a volume of 10cm 3 . All of the impossibility theorems I spoke of in the previous chapter actually say that a certain intentional correlate is logically impossible. I would thus like to extend the concept of a contradiction, or, more correctly, return it to approximately the position it occupied before modern formal logic replaced it by a very watered-down counterpart. 5 O n the other hand, I hold fast to the view that contradictions cannot exist in the non-intentional part of the world.

Figure 15.1 (Opus 1, 1934 Design Oscar Reutersvärd)

269

ONTOLOGICAL INVESTIGATIONS

Since presentational intentionality claims to be in direct contact with the world, the conclusion is that only representational intentionality can be directed toward contradictions. But, as we shall now see, representational intentionality actually contains perception as well as thinking in the ordinary sense. All of the impossible figures due to artists such as Escher and Reutersvard (see Figure 15.1) are such that when we direct our attention to them our acts have contradictory correlates. Their intentional correlates are logically impossible. This point is easily misunderstood if presentational and representational intentionality are confused. Is the impossible figure the correlate of a presentational or of a representational intentional act? The solution of the problem lies in the insight that seeing a picture, like most uses of ordinary language, contains something which is, so to speak, presentationally given, without for that reason ceasing to be a representational act. The purely graphical pattern (like the purely linguistic sign) is presentationally given (i.e. it makes no claim to go beyond the act in question), but there is no presentational act directed to the graphical signs. To see a picture is to have a representational intentional act. In an impossible figure different parts of the drawing itself point toward different states of affairs, states of affairs which are logically impossible to add to one another. In the same way different statements may simultaneously point toward states of affairs which it is logically impossible to combine. I believe, however, that picture-contradiction is genetically primary and that from a developmental perspective linguistic contradiction should be seen as secondary. 15.4 NESTED I N T E N T I O N A L I T Y We shall now look at the interaction between two subjects or two agents. I would once again like to remark that our discussion takes place on the categorial plane. It is therefore very abstract. We shall clarify some different types of possible intentionality structures; we shall investigate whether there appear new such structures when an intentional act is directed toward another act. I shall approach the abstract relations we want to isolate by looking at Jean-Paul Sartre's analysis of the phenomenon of shame. Sartre asks his readers to imagine a situation where one makes a vulgar gesture: 270

NESTED INTENTIONALITY

I have just made an awkward or vulgar gesture. This gesture clings to me; I neither judge it nor blame it. I simply live it. I realize it in the mode of for-itself. But now suddenly I raise my head. Somebody was there and has seen me. Suddenly I realize the vulgarity of my gesture, and I am ashamed. 6 We shall try to place this simple but keen observation into a more abstract structure. At least the following moments belong to the situation described: (1) the agents A and B, as well as the action (= the vulgar gesture of A = Ad); (2) the fact that A knows that he is performing action d, and the fact that В sees and despises A for doing d; (3) the fact that A sees that В sees that Ad. This structure can be made clear if we extend the notation used above and use an arrow to represent intentionality in general; specific kinds of intentionality are represented by an arrow with the specific character described above it. With such a notation we obtain the following: (1) (2)

(3)

A Ad knows A *Ad sees despises A *{B iUAd)

В В despises В——*Ad

One of Sartre's main conclusions can now be stated in the following way. Shame can never be found on plane (1) or (2), but requires the structure which exists on plane (3). Shame cannot be identified with the action Ad, and neither can it be identified with i4's being aware that he is doing d. Shame cannot exist in me as a monad, but presupposes two subjects in intentional contact with one another. Sartre himself gives a concise formulation of the points: 'Thus shame is shame of oneself before the other.'1 On the first plane there exist only states of afTairs without any intentionality, and shame cannot be found there since shame essentially involves intentionality. On the second plane we find intentional acts directed towards non-intentional states of affairs, but shame cannot be found here either since shame implies that one is ashamed about something before someone else. Shame includes an intentional act 271

ONTOLOGICAL INVESTIGATIONS

directed towards another intentionality. Shame is a type of whole which presupposes at least two participants. In order to understand this analysis, certain things must once again be stressed regarding the relation between representational and presentational intentionality. Sartre's example falls under the category of presentational intentionality. However, this does not imply that one cannot feel ashamed when alone, but rather that, when one is ashamed and alone, the same fundamental structure obtains; the only difference being that person В is then not a real person but an imagined one, perhaps even an imagined undetermined 'Mr Somebody'. Sartre emphasizes this point and maintains not only that the same structure is to be found in presentational and representational shame, but also that representational shame is only to be found if it has actually been preceded by presentational shame: Yet although certain complex forms derived from shame can appear on the reflective plane, shame is not originally a phenomenon of reflection. In fact no matter what results one can obtain in solitude by the religious practice of shame, it is in its primary structure shame before somebody.8 With this in mind, we can return to the abstract structure described above. It is not only the case that there exist three distinct planes, there also exist relations among the planes. What exists or occurs on plane (3) implies that which exists on planes (2) and (1). That A —• (B —» Ad) entails A, B, and Ad as well as (B—* Ad), ought to be clear without further comment; but it also entails A —» Ad. A cannot see that В sees that he does d without thereby becoming aware that he does d. This means that planes (1) and (2) are necessary conditions for A —* (B —> Ad), or that the latter is one-sidedly existentially dependent on the former. The existential dependence relation between planes is a dependence relation between satisfied presentational intentional acts. Where the acts are not satisfied the dependence relation obtains within the act of a single agent; the structure described is then not a structure between two real agents but merely a structure that connects this agent and an imagined agent. The naive realism I defended in section 13.6 allows me to regard real agents as normally nested. When an intentional act of a subject is existentially dependent 272

NESTED INTENTIONALITY

upon intentional acts of other subjects, we have a very specific kind of ontological unity which I shall call nested intentionality. Such unities can be more or less nested, but the fundamental structure of nested intentionality is A —* (B —* A). It should be compared with intentional phenomena where an agent is directed towards a naturalistic state of affairs (= A —* O, where О stands for object), or where an agent sees that another agent sees something; either a naturalistic state of affairs (= A—*(B—*0)) or a third agent (= A->0 A (В -> O) A (B —• C) A (B -*• A)

-| IJ

= non-nested intentionality = nested intentionality

The last structure above is an abstract intentional structure which is exemplified by the intentional moment of shame (shame of course involves more than intentionality; cf. page 262). But this nested structure is more abstract than the particular example we have described; there can exist other examples of the same structure. The structure in question delimits the intentional moment of a whole set of emotions and intentional states. Apart from shame, it also fits its opposite, pride. To be proud is primarily the same as being proud of oneself before another, to paraphrase Sartre. Pride, too, is a whole which has participants as parts. Emotions like the opposites pride-shame thus distinguish themselves from emotional opposites like happiness-sadness, pleasure-horror, desire-repulsion, admire-despise, in that the latter pairs can (they need not) have non-nested intentional structures like A —* О or A —» В while the first pair necessarily involves the nested structure A—* (B —* A). This also means that one kind of emotion may be existentially dependent on another kind of emotion. Shame depends upon despising someone and pride depends upon admiration: sees despises A is ashamed: A * (B *A) sees admires A is proud: A * (В *A) It is not only mental states which fit into such structures. A person cannot have the property of being a teacher in splendid isolation in the way that he in himself has a certain length and a 273

ONTOLOGICAL INVESTIGATIONS

certain shape. Nor can he have the property of being a teacher by himself in the way in which he can perform actions like sitting, walking, drawing, and speaking. Assume, for example, that I day after day act just like a teacher. At some time every day I go to a certain place and give lectures. But nobody sits and listens, and I am not paid by any institution. In that case I am not a teacher in spite of my manifesting precisely the same behaviour as a teacher. If there had been students sitting there and had I been paid by a school or university the same behaviour would have made me a teacher. The property of being a teacher is, like shame and pride, dependent upon something beyond behaviour itself; it presupposes certain types of nested presentational intentionality. In order to be a teacher I must see that there exist students who see me. The teacher must see that there are students who see him as a teacher. We obtain the structure: T—* (S—* T). Nor, in a corresponding way, can one be a student per se, but must fulfil the intentionality structure S—> (T—* S). I n chapter 8 the 'teacher-student' couple was used as an example of internal relations. Together with 'preacher-congregation' and 'capitalist—worker' it was the only example where there was a spatial gap between relata. We see now what this gap depends on. The existential dependence relations are in these cases based on the intentionality category. The relata are participants. It might seem a,mazing that social roles or 'character masks', to borrow an expression of Marx, exemplify the same structure as emotional states. Social roles and emotional states seem otherwise to be categorially distinct. That this is so depends on the fact that we analyse only the intentionality moment of the phenomena; we have left aside the tendential moment with its accompanying actions. And it is on this latter side that the differences exist which we so clearly experience. Emotional states fall under the tendency category, while roles seem to fall under the spontaneity category. Of course roles are completely deterministically understood in many sociological role theories, i.e. they become tendencies, but in everyday speech associations which are connected with the spontaneity category (free will, etc.) seem to have the upper hand. The structure we have now studied, A —* (B—*A), can be easily built into similar but more comprehensive nested structures. Let us return to Sartre's example of the man who is ashamed. If, when he is ashamed, he begins to blush about his own shame, the observer 274

NESTED INTENTIONALITY

can also perceive that the man is ashamed. The man, in turn, can discover this and even begin to be ashamed of being ashamed. He thinks it contemptible to be ashamed. We have a shame of second order whose intentionality is nested according to the following structure: A —* (B—> (A —* (B —» Ad))), i.e. A is ashamed because A sees that В despises A for being ashamed that В despises what A is doing. Another complex nested structure is the intentional moment of being ambitious. A man who is ambitious wants or strives for situations where he can see that others (B) admire him. We get: A is ambitious: A

wants sees admires * (A > (В * Л)) A is proud

What has been said about pride, shame, shame of second order, and ambitions obviously shows the possibility of a formal study of emotions and other intentional states; a study based on the concept of nested intentionality. Descartes and Spinoza tried to carry out such analyses of the soul's passions and emotions, but since then it has not, to say the least, been comme il faut amongst philosophers, except for work by Brentano and his pupils. All of the intentional states hitherto mentioned are states which via intentionality depend upon and contain other persons, but which, as states, exists in one particular person. Let us now investigate a phenomenon like friendship. That A and В are friends seems to mean more than that A likes В and that В likes A (row (1) below). There must be some sort of mutuality that goes deeper. 9 In analogy with earlier analyses, we can perhaps say that A in addition must see that В likes him, and that В must see that A likes him (row (2) below). If we return to the abstract notation given earlier, and abbreviate 'likes' with Τ and 'sees' with V, we can write the following with A and В as agents:

(1)

(2)

A I

В I В—* A s i B—*(A—*B)

В s i A—*(B—> A)

To my mind, not even this amendment is enough in order to capture the intentional moment of friendship. A must want В to see 275

ONTOLOGICAL INVESTIGATIONS

that A likes him; and В must want A to see that В likes him. We get yet another row ('want' = 'w'): (3)а^(В^(АЛВ)) (= A3)

B^> (A—>(B—> A)) (= B3)

If there is real friendship, Β should, I think, see A's want as described by (A3) and A should see B's want as described by (B3). Furthermore, A should want В to see this, and В should want A to see this. We have to add another two rows: (4) A—*B3 ( = Л4) w (5) A—*B4

Я Л А З (= 54) w B-+ A4

When one knows that p, reflection shows that one also knows that one knows that p, as well as that one knows that one knows that one knows that p. There is no end to this regress, and in the same way the intentional moment of friendship involves an infinite series. Every other row gets Ά sees . . .' and 'B sees . . .', and every other row gets Ά wants . . .' and ' В wants . . where the intentional correlate is to be found in the opposite column on the row before. This regress has obvious similarities with the regress which occurs in two mirrors which reflect one another, I shall thus call it 'intentional mirror infinity', or, for short, 'intentional mirror'. M y thesis is quite simply that intentional mirrors constitute a special sort of infinity. They are of course potential infinities in the same sense as a spatial extension is potentially divisible to infinity. But spatial extension and intentionality are as different as categories can be, and so we must have two completely different kinds of infinity. T h a t intentionality can involve infinity is neither more nor less disturbing than the role of infinity in other categories. An analysis of the intentionality moment of phenomena of the mutual type — like friendship, love, and loyalty - must result in an infinite regress of the type described. Even if A likes В and knows that В likes him, friendship is more than this. It contains the possibility that В should know that A knows that В likes him, and the possibility that A should know this, and so on. Since the intentionality category is such that the subject-pole is anchored in only one body-substratum, there can be no question of any 276

NESTED INTENTIONALITY

absolute mutuality in the sense of the complete merging into one another that mystics claim to describe. The closest we can come to one another as subjects is to stand in the mutuality relation exhibited by intentional mirrors. T h e fundamental intentional mirror on which all other social phenomena are at bottom founded, is the following (p = 'perceive'): А В (1)

Ρ A^B

Ρ B-* A

(2)

Ρ Ρ лД(ВДл)

Ρ Ρ В—* (А—* В)

(3)

Ρ Ρ Ρ А Д (В Д (Α Β))

Ρ Ρ Ρ Β-^(Α^(Β-^Α)) [etc.]

Let us abbreviate this structure by A «-»· B, and expand the example by bringing in a third agent, C. We thereby also get nestings of A and С as well as of В and С according to the same structure, i.e. A (Б*~*С) ( 2 С )

)

B-+(A~Q В-^{А-*{ЕПС))

В С C—>(A~~B) C-*{A-*{Bt~C))

(3) А—* (В—* (А—* (В" С)))

[etc.] T h e principle generating the successive higher order intentional levels is that A on level (2) takes as intentional correlates B's and C s intentionality on level (1), that A on level (3) takes as intentional correlates B's and C s intentionality on level (2), and so on; and in a corresponding way for В and C. Each level will in this way contain different intentionalities for each person, where η is the number of the level. This infinity we can abbreviate by A,B,C 277

ONTOLOGICAL INVESTIGATIONS

Now we bring in a fourth agent, D. First, this person is nested in A, D"" В and D *""* C. Secondly, he is nested in D,A,B, D,B,C and D,A,C but also, thirdly, nested in a regress beginning with D^(A,B,C) and ending in an infinity which we can symbolize by A,B,C,D. This kind of nested intentionality gives us the real structure of intersubjectivity. 10 My introduction of the category of nested intentionality took its departure in Sartre's analyses of emotions. In order to bring out some other features of nested intentionality it is pertinent to begin with H.P. Grice's attempted analysis of meaning. 11 1 say 'attempted' because damaging criticisms have been made of Grice's analyses, but such criticism should not (and usually has not tried to) conjure away the type of nested intentional phenomena described. 12 I am not here interested in the specific details of different speech acts, I am interested in showing the existence of different general structures of nested intentionality. A common way of restating Grice's analysis very briefly is the following: S meant something by (or in) uttering χ iff S uttered χ intending (1) that his utterance o f * produce a certain response r in a certain audience A; (2) that A recognize 5"s intention (1); (3) that Л'в recognition of S?s intention (1) shall function as at least part of 4 ' s reason for ^4's response r. 13 As indicated above this analysis has flaws as a general analysis of meaning, but, in my opinion, it captures the phenomenon of giving an order very neatly. Assume that A says to В: 'Open the window!', or via gestures transmits the same message. If В willingly follows the order, he performs an action the reason for which is that another person (Л) wants him to do it. If В has understood the order, he has not only 278

NESTED INTENTIONALITY

understood that A wants him to open the window, but also that A wants it to be the case that В will realize that A wants him to open the window. A does not say that В should open the window for an arbitrary reason, but intends the order in itself to be a sufficient reason; so we must add that A wants В to open the window because В realized what A wants him to realize. The structure in the order fits the following schema: A orders В to do d if and only if (1) A intends В to do d; (2) A intends В to recognize (1); (3) A intends В to do d for the reason that В recognizes (2). If we symbolize B's doing of d with Bd, and the fact that Bd is carried out for a certain reason with Bd [r . . .], we can capture the structure of obeying an order (an execution of the behaviour commanded) in a way which shows its resemblance to the other structures discussed above. Let 'i' = 'intends' and V = 'recognizes' when V stands above the arrows).

(1)

(2) (3)

The complex state of affairs described here is a whole which has A and В as parts in the sense of participants. The main differences between this nested intentionality and the structures described earlier are (a) that it contains reasons for actions; (b) the peculiar Gricean kind of self-reference which is to be found exemplified by A at level (3). In 15.8 ('Macroagents') the importance of the latter will become clear.

15.5 NESTED MISINTENTIONALITY I have earlier shown, in connection with the discussion of the 279

ONTOLOGICAL INVESTIGATIONS

explanation schema of methodological individualism, how misintentionality in the form of a misperception can lead to non-optimal or unsuccessful actions ('the glass door effect'). The correlate of the intentional act was a naturalistic state of affairs. Let us now look briefly at the corresponding phenomena where we have two agents who stand to each other as intentional correlates. I shall now take my examples from a philosophically oriented psychologist partly inspired by Sartre, R.D. Laing. Before I get involved in examples, it may be of interest to cite from Laing's foreword to his book Knots. He writes: T h e patterns delineated here have not yet been classified by a Linnaeus of human bondage. They are all, perhaps, strangely, familiar. . . . I could have remained closer to the 'raw' data in which these patterns appear. I could have distilled them further towards an abstract logico-mathematical calculus. 14 I should add that it is not my aim to emulate Linnaeus. O n the other hand, I do want to bring out a little of the abstract calculus Laing also believes possible. By raising Laing's example to a higher level of abstraction, one sees similarities in structure to the analyses of Sartre and Grice, but also dissimilarities. Laing has actually captured a type of intentionality phenomenon which the others have not noticed. As we shall see yet again nested intentionality provides an area of study which is well worth philosophical cultivation. Now to the first example: She wants him to want her He wants her to want him To get him to want her she pretends she wants him T o get her to want him he pretends he wants her Jack wants Jill wants Jill's want ofJack Jack's want of Jill so so Jack tells Jill Jill tells Jack Jack wants Jill Jill wants Jack a perfect contract 15 J a c k ' s intentional correlate is 'Jill wants me', but he does not reach that correlate. The state of affairs which exists is rather 'Jill wants 280

NESTED INTENTIONALITY J a c k to want her'. In the same w a y and for the same reason Jill does not reach her correlate either. T h e two misintentionalities bind each other, however, in such a w a y that they all the while apparendy confirm one another. Nested misintentionality will come to determine many of their actions. Laing calls such cases as these 'fantasy circles'. T h e circles are seldom seen through even by outside observers, which is, I think, the reason for our still lacking labels for this type of social—psychological phenomenon. Behind Jack's (and Jill's) pretending there lies a form of strategic thinking which is connected to the schema of explanation in methodological individualism. Jack understands the situation as though it were rational for him to trick Jill in order to reach his correlate. His goal is that she will want to have him, and he analyses the situation as being such that if Jill believes that Jack wants to have Jill, then Jill will come to want to have Jack. But since J a c k does not primarily w a n t to have Jill but only Jill's wanting-to-have-Jack, J a c k has to lie, i.e. create a misintentionality in Jül. In the earlier examples of misintentionality in agents, an understanding or grasp of that intentionality was a condition for understanding certain unsuccessful actions. In these examples intentionality was directed toward naturalistic states of affairs, but now w e have intentional acts directed towards intentionality and misintentionality which are themselves directed towards intentional acts. W e obtain a situation where the actions can be described neither as unsuccessful nor as irrational. T h e actions cannot be called unsuccessful since in one sense J a c k and Jill both obtain what they want; and the actions cannot be called irrational for there exists no knowledge which A and В had but did not use. 1 6 T h e notation I have used in order to represent nested intentionality can be enlarged so that some features of nested misintentionality can also be represented. W h e n we have misintentionality we shall write a dash ( - ) under the arrow which represents the intentional act in question. Laing's example then contains the following structure ('want' = 'w' and 'sees' = 's'): A

В

(1)

(2) 281

ONTOLOGICAL INVESTIGATIONS

(3)

W S W А - > ( В г * { А - * В ) ) S

W

S

(4)

W S W ( А ^ ( В ^ А ) )

В - ^

W

(Л-»(В->Л))) В [etc.]

S W S W = * { А ^ > ( В - ^ ( А - > В ) ) )

The structure above should be compared with that below which represents the nestedness of two agents who both really want each other (cf. the analysis of friendship on pages 275f.). А

В

(1)

W А—* В

W В —> A

(2)

S W A—*(B—*A)

(3)

W

S

S

W

W

A ( В ( A - * В)) S

W

S

w (A—*B)

B-+ S

W

В—* {A—* {В—* A))

W

S

W

s

w

(4) A ( B ( A { В - * Л))) В—* ( A ( B - * (A—> B))) [etc.] s Rows number (2), (3), and (4) contain instead of the first occurrence of in rows number 2, 3, 4. But since this difference is not visible to the agents themselves, the similarity between these corresponding levels can hide the important fact, namely the difference between rows 1 and 1. Jack and Jill, consciously or half consciously, have the intention of tricking one another, but the same type of bad (or good?) interpretation-circle can arise even when the agents' acts exhibit misintentionality which is not at all based on any intention to mislead. What follows is an example which contains no hidden intentions, but which nevertheless literally leads to a vicious circle — a vicious circle which is actually a spiral. In the example a temporal development of nested misintentionalities is described involving 'mismatched interpretations, expectancies, experiences, attributions, and counter-atributions'. 'It', to continue the quotation, 'starts to whirl something like this': Peter:

Paul:

(1) I am upset.

(1) Peter is upset.

(2) Paul is acting very calm and dispassionate.

(2) I'll try to help him by remaining calm and just listening. 282

NESTED INTENTIONALITY

(3) If Paul cared about me and (3) wanted to help he would get involved and show some emotion also. (4) Paul knows that this upsets (4) me.

He is getting even more upset. I must be even more calm. He is accusing me of hurting him.

(5) If Paul knows that his (5) I'm really trying to help, behaviour upsets me, he must be intending to hurt me. (6) He must be cruel, sadistic. (6) He must be projecting. Maybe he gets pleasure out of it, etc.17 15.6 NESTED CONTRADICTIONS Earlier in this chapter (15.3) I made the point that not only linguistic phenomena but also perceptions may be contradictory. Now, after having introduced the idea that emotions contain as an essential part a moment of intentionality (15.4), I can add that even intentional states may be contradictory. Once again I shall take an example from Sartre. His analysis of love is a very illustrative example.18 Sartre says that when one loves someone, then one wants the loved one in two completely incompatible ways. A partly wants the loved one (B) to have freely chosen to love him, i.e. the loved one's love ought to fall under the spontaneity category. But in part and at the same time, A also wants the loved one В to love him as a result of a passion which she finds irresistible. In the latter case, A wants to function as an efficient cause for the loved one's response. But a phenomenon cannot at one and the same time arise both spontaneously and as the result of an efficient cause. Love is a contradiction, an impossible figure. Our analysis does not make love out to be something which, so to speak, only occurs inside a person's head. The contradiction characteristic of love is to be found in one's very perceivings. I remind the reader that I maintained that perceptions are a form of presentational intentionality and that they contain representational intentionality. When we perceive a tree, for example, its front side 283

ONTOLOGICAL INVESTIGATIONS

is presented, but its farside is represented. (The way the tree has just been — immediate past — and is about to be are also represented.) In the same way, causes of actions are often represented at the same time as the actions themselves are presented. Without coming into conflict with the thesis that contradictions can only be found in representational intentionality (cf. page 270) (not presentational), I can therefore maintain that we can have contradictory perceptions. T h e situation is the same as in the case of the impossible figures. T h e r e a graphic pattern was presentationally given and a representational intentional act was directed towards two incompatible figures. In the same way love, too, can be an impossible figure. In the perception of the beloved one's behaviour there is a representational intentionality which is contradictory. A contradiction between spontaneity and efficient causality manifests itself in many human contexts. It is in fact the contradiction which, in the present theory of categories, represents the old conflict between natural-causality and freedom-causality, to borrow terms from Kant. 1 9 An agent who has contradictory intentional acts strives after the impossible. The correlate cannot be reached, but his intentional states and acts nevertheless affect his actions. The exact way in which this expresses itself cannot be understood from intentionality alone. It requires studies of each individual case, even if certain clues can be obtained from the contradictory correlate itself. Someone who believes that square circles can exist may perhaps be led to draw different figures for the rest of his life in a vain search for a satisfying representation. And he who seeks love in the contradictory form described by Sartre will perhaps spend his life going from one woman to the next, or spend his whole life in dissatisfied nagging at the woman he nevertheless cannot bring himself to leave (as the poet writes: 'We can die by it, if not live by love'). 2 0 But one cannot determine in advance how an impossible intention will try to realize itself in detail. It might be useful here to mention Gregory Bateson's famous studies of the 'double-bind' phenomenon. 2 1 I shall take a simple example. A mother who, after a long absence, welcomes her child by saying, 'Oh, finally. O h but I love you!', but who with her whole body shows abhorrence and dislike towards her child, gives the child a contradictory message. T h e child perceives a contradiction. And it can, as Bateson has shown, have both long-lasting and devastating consequences if it is 284

NESTED INTENTIONALITY

a frequent occurrence. Contradictions do not only exist in the sphere of the intellect and logic, but also in the sphere of feeling and perception. However, contradictions do not appear outside the sphere of intentionality. The Bateson example conforms to the following structure ('sees' = У and 'loves' = V): s I A-*(B^A&B-*A)

-I

Sartre's contradictory love has a structure which is somewhat more complex since it involves reasons, and causes behind the state of love. But if we build the latter into the symbolism - we let '/„' mean 'loves spontaneously', and 7 p ' loves because of an overwhelming passion - we obtain a similar structure since /, and l p exclude one another. 22 We get: s L Л->(ЯАЛ

&

L ВЛА)

In both cases we have nested intentionality and we have contradictory intentionality, but neither case exemplifies what I want to call nested contradiction. Nested contradiction is a special kind of contradictory intentionality which, unlike the two examples above, does not have the form A —» \p & not-p). In nested contradictions the contradiction itself is nested. The structure can be found by means of substitution in the Gricean analysis of ordergiving earlier presented (see pages 278-9). When A orders В to do d the following intentional structure was said to obtain where 'intends' = Y, 'recognizes' = V when above the arrows, and 'B does d for reasons r" = 'Bd [r: . . .]':

(1)

(2) (3)

(1)

(2) 285

ONTOLOGICAL INVESTIGATIONS

(3)

τ г Bd [г.В —* (А(В—>

(АBd)))]

In this structure we shall now try to insert the specific command 'Forget C!'. Or, to avoid possible misinterpretations, A orders B: 'You must forget CI'. It is thus a matter not only of forgetting, but of forgetting because of Л'в authority. T o have forgotten something implies that one lacks intentionality toward that thing. If we symbolize the lack of intentionality with an arrow with a minus sign above it we can exchange 'Btf for ' B С in the structure above. We then get this structure:

(1) (2) (3)

(1) (2) (3) There are no contradictions to be found on the first two levels, but the third level exemplifies a contradiction. First, there is a contradiction since A requires that В shall be both aware and not aware of С; В is supposed to remember to forget C. Second, the contradiction does not appear in the form A—*(p&. not-/»), i.e. A does not give В the order 'Remember to forget C!'. This means that we have a contradiction but not an ordinary contradiction of the type explored by formal logic; we have a nested contradiction. I f agent В does not realize the contradictory character of the command he will strive after the impossible. He will be like the man who tries to draw a square circle. Another example of the same structure is the order 'Be spontaneous!'; to be spontaneous here means carrying out actions without bothering about the reasons for these actions. If A directs such an order at В we should exchange A-X (B—> C)

286

NESTED INTENTIONALITY

on the first level in the last schema for А Л The latter means that A intends that В should not think at all of giving reasons. On the third level we then get a nested contradiction. Nested contradictions are perhaps the grain of truth contained in Hegel's famous master-slave dialectic. The master (A) wants (1) the slave (В) to regard himself as a thing, i.e. as something that lacks intentionality: w A-*(B-+x) A also wants В to (2) recognize this intention and (3) to turn himself into an intentionless thing because of this recognition. If w A —> (B-*x) is inserted for Л —• (.ß —• C) in the last schema, we get the corresponding nested contradiction on the third level. I shall end this section by quoting yet another 'knot' from Laing. Think about this: A son should respect his father He should not have to be taught to respect his father It is something that is natural That's how I've brought up my son anyway Of course a father must be worthy of respect He can forfeit a son's respect But I hope at least that my son will respect me, if only for leaving him free to respect me or not 23 15.7 MARXIST DIALECTICS AND INTENTIONALITY This theory of categories of mine is marxist in inspiration and (to a lesser extent) as regards the problems dealt with, but somewhat analytic in its way of handling these problems. 24 The resulting ontology, however, is most closely related to the positions of two philosophers who belong neither to marxism nor to analytic 287

ONTOLOGICAL INVESTIGATIONS

philosophy: Aristotle and Husserl. In this section I shall comment on a problem within marxism, the problem of dialectic. This problem is discussed by a philosopher whose views do not have much in common with this theory of categories: Hegel. It was Hegel who put forward the view that in the world there exist actual contradictions, and that it is a dialectic between contradictory opposites which explains the world and its history. Both the idea of nested intentionality and the idea of internal relations are to be found in Hegel. But the account given in chapter 9 of internal relations shows that the latter are a special case of Husserl's existential dependence relations. These are all passive, not productive in Hegel's sense. Moreover, Husserl allows for onesided dependence relations. And, last but not least, they have nothing to do with contradiction in Hegel's sense. J o n Elster, whose work is analytic in orientation, has suggested that there exist interesting observations in what Hegel and M a r x say about dialectics and contradictions, but that what is valuable must be separated from its mystifying packaging. 2 5 In this way it can be placed into a traditional analytical framework where all types of contradiction are defined in terms of logical relations between sentences. Elster's analyses are very valuable and I shall take my point of departure from them, even though my conclusions diverge somewhat from his. Elster's 'vindication of dialectics' consists of a denial of the two theses: (1) 'There are contradictions in reality', (2) 'An adequate description of reality must contain self-contradictory propositions'; but a defence of a third thesis : (3) 'There are situations in reality that can only be described by means of the concept of a logical contradiction.' 2 6 He also makes a distinction between 'contradictions of the mind' and 'contradictions of society', and in the former group between 'contradictory desires' and 'contradictory beliefs'. W h a t I have said so far about 'contradictory intentionality' is, as can be seen from the footnotes, partly inspired by Elster's analyses. O n e such case is my reference to Sartre's analysis of love, an analysis Elster cites as an example of contradictory desire. H e who loves in the way described by Sartre finds himself, according to Elster, in a real situation 'that can only be described by means of the concept of a logical contradiction'. O n e respect in which Elster and I differ is that I regard the concepts of intentionality and existential exclusion (D9.5) as the 288

NESTED INTENTIONALITY

basic concepts needed in order to understand contradictions, while his basic concepts are taken from formal logic. This difference is often of no significance, since logical contradictions belong to the area of the intentional (cf. sections 15.3 and 15.6); but significant differences sometimes show themselves. One such is my emphasis on contradictory perceptions. Perceptions are intentional phenomena, and, as such, they can in principle include contradictions. But for Elster, who begins with descriptions, all contradictions come automatically to be placed on the linguistic level; this is implied by thesis (3) above. The distinction between 'contradictions of the mind' and 'contradictions of society' is not essential. What is essential is rather to show that classical examples of supposedly dialectical phenomena can be understood with the help of the intentionality category. For reasons of space I shall here take up only a couple of marxist examples. Certain supposedly dialectical phenomena can, as Elster has shown, be understood by means of 'non-dialectical' concepts like 'counterfinality' and 'suboptimality'. Both concepts belong to the intentionality category. I shall not try to summarize those of his arguments which relate to these concepts (since I agree with what he says), but shall comment on what he calls the 'fallacy of composition' 27 and discuss an example he does not deal with at all. 'The fallacy of composition' is the mistaken belief that because something is possible for each entity of a certain kind, then it must also be possible for all of them simultaneously. If we use formal notation this means that the following is not a logical truth: (Vx) (P(Fx)) —» P(Vx) (Fx). It is easy to find examples where the antecedent is true but the consequent false. It is possible for each person who starts in a marathon race to win it, but it is not possible for all to do it together. There is only one winner, but any of those who start can be the winner. It is possible to read any of the books in my library in one day, but it is not possible to read all of them in one day. Some such examples may be called learning or reproduction impossibilities. It is possible for anyone at any time to break his promises, but it is not possible for everyone always to break their promises, for then the 'institution of promising' would not be able to reproduce itself, and no one would be able to learn what a promise is. A similar point can be made about the statement that not everyone can lie, for if that were the case, no one would be able to learn a language. 289

ONTOLOGICAL INVESTIGATIONS Elster's point is that a fallacious conclusion of this sort, which is a purely logical mistake by a certain person, can a p p e a r in a societal f o r m . It is then not a question of a logical fallacy by a certain person, b u t rather a question of a societal situation which ' c a n only be described by means of the concept of (formal-) logical contradiction'. O n e of Elster's examples of this type of situation is w h a t M a r x calls 'the contradiction in the general formula of capital'. A s s u m e that we have a society containing a group of producers a n d a g r o u p of merchants w h o m a k e a profit by buying c h e a p and selling dear. H e r e it is possible both for each merchant a n d for all the m e r c h a n t s simultaneously to m a k e a profit in this way. T h e s e m e r c h a n t capitalists, w h o m we can assume to have become very well-off, might easily imagine the 'good society' to be a society consisting only of merchant capitalists. As a m a t t e r of fact a n u m b e r of early economists m a d e suggestions in this direction. S u c h a state of affairs is, however, impossible. H e r e we have a traditional fallacy of composition. It is possible for each m e r c h a n t capitalist to buy cheap a n d sell dear, b u t it is not possible for everyone to buy and sell goods a n d to be m e r c h a n t capitalists a n d live on such a profit-margin. Someone somewhere must lose w h a t the other gains. So long as merchant capitalists only constitute a part of a society their profit can arise by being taken from those w h o are not m e r c h a n t capitalists. But a society consisting only of m e r c h a n t capitalists is a contradiction. O n e should not, however, think that in the situation assumed above, the m e r c h a n t capitalists themselves, or anyone at all, c o m m i t the fallacy of composition a n d actually believes that everyone in the society could be m e r c h a n t capitalists. Each person can believe t h a t he himself can live by buying c h e a p and selling d e a r , a n d not think at all about w h a t is possible for society as a whole. T h e n there is no contradiction in the relevant intentional acts of any single person. There a r e no 'contradictions of the mind', b u t the description of the society includes a contradiction in the sense t h a t a description of the aggregate of individual intentions cannot possibly be realized. If, in our imaginary society, those w h o already are m e r c h a n t capitalists strive to retain their status, a n d everyone else strives to acquire this status, then the contradictory 'intentional aggregate' leads to the occurrence somewhere of conflicting tendencies. W e see h e r e how the categories of contradictory intentionality a n d

290

NESTED INTENTIONALITY

counteracting tendencies are both exemplified. The social contradiction expresses itself in the fact that A wants to buy B's goods below their value, while В wants to buy A's goods below their value. There are two conflicting tendencies, and presumably there will be no exchange. The tendencies cancel one another. It is only this connection between contradictions and tendencies which makes Elster's thesis interesting. As I see it, Elster does not himself make this sufficiently clear. If such a connection does not exist, the thesis that 'There are situations in reality that can only be described by means of the concept of a logical contradiction' is reduced to the triviality that if one person believes p, and another not-/», then the whole situation in which they partake cannot be described in any other way than by means of a formal—logical contradiction. If one is to understand all the peculiarities of 'dialectics' more is needed than a distinction between contradictions on the one hand and conflicting tendencies on the other. One must also see the connection between them, i.e. see that intentions have both an intentional and a tendential aspect. Were this not the case, Elster's analysis would lack interest. Marx himself speaks of the general formula of capital as M-C-M1, where С stands for a commodity, Μ for a sum of money, and M 1 for a sum of money greater than Af.28 The general form represents the 'intention of capital', which is also the intentions of individual capitalists. One invests money ( A f ) in buying commodities (C) in the hope of thereby obtaining even more money (M 1 ). Assume that all the individual capitalists are agents with this intention. But assume also that all these capitalists are merchant capitalists. Then yet another aspect of the 'contradiction in the general formula of capital' reveals itself. Since the formula M-C-M 1 is valid for each individual, but is not generalizable to all simultaneously, there must somewhere exist a limit to the number who can live with this formula as their guiding star. If the merchant capitalists have a conscious tendency to transform society in such a way that it consists only of merchant capitalists then they thereby unconsciously manifest a tendency to cease to be merchant capitalists. Marx can now be interpreted as saying that the formula Αί-C-M 1 also has a contradiction-^« form. Industrial capital invests money (Af), as does merchant capital, in the buying of goods (£), with the aim of obtaining more money (M 1 ). But here, according to Marx, there is no 'contradiction'. One way of realizing 291

ONTOLOGICAL INVESTIGATIONS the aim Л / - С - М 1 is to be a merchant capitalist, another is to be an industrial capitalist. I f the first w a y contains an 'Elster-contradiction', then a tendency presumably arises to realize the same aim in another w a y , i.e. the merchant capitalists remain capitalists but change over to being industrial capitalists. T h i s point must be borne in mind if one is to understand the whole o f the 'dialectic' in M a r x ' s example. Elster picks out three examples of 'contradictions' in M a r x ' s writings. I h a v e now commented on one of these. T h e two others are 'the contradiction of the law of the falling rate of profit' and 'the contradiction between the forces of production and the relations of production'. W i t h regard to these, however, I shall only refer to Elster's analyses, 2 9 and instead go through the example w h i c h Elster, remarkably enough, does not mention: M a r x ' s famous 'derivation' of money in the first part of the first volume of Capital. It is here that one can see, more clearly than anywhere else, M a r x ' s need of the category o f existential dependence and of a well-delimited intentionality category. M a r x distinguishes a m o n g four forms of value which, from a chronological point of view, have followed upon one another in history, but w h i c h also in some never really clarified Hegelian (or Hegel-turned-upside-down) way, have been assumed to follow logically from one another. 3 0 T h e first form is the simple form of value. Historically speaking this form is connected with societies w h e r e the production o f goods is only temporary and w h e r e trade is by direct barter. E x c h a n g e according to the simple form of value can be written in the following w a y : χ units of commodity A = у units of commodity В T h e person w h o wants to sell A and b u y В has no use for more .As but is in need o f Bs. Commodities o f type В have, in M a r x ' terminology, use value for that person. T h e situation is correspondingly similar for the other person. Say that the two persons exchange food for clothing - the one is a farmer and the other a tailor. T h a t the goods are exchanged in just the proportions given means that χ units of Л have the same exchange value as у units of B. T h e second form is the expanded form, w h i c h can be written in the following w a y :

292

NESTED INTENTIONALITY

у units of commodity В χ units of commodity A = ζ units of commodity С и units of commodity D Here there are a number of different sorts of goods such that he who wants to sell A can immediately set A in relation to many other types of commodity. The exchange value can as a consequence be measured with the help of many alternative types of commodity. The third form, which Marx calls the general value form, is merely the reversal of the expanded value form (i.e. the second form); thus it has the following form: χ units у units Ζ units и units

of commodity of commodity .. of commodity of commodity

A В ,. „ _ = ν units of commodity Ε С ' D

The last schema is typical for societies where the exchange of goods has an even greater intensity. The normal procedure when someone now wants to exchange commodity A for commodity В is that he first exchanges A for the universal equivalent E, and then with the help of Ε obtains B. Now the exchange value of all commodities are measured in one and the same commodity, E. In all three commodity forms all commodities have a use value for someone; this also holds for E. In the last case one can think of societies where a certain kind of cattle is used as currency. In the fourth and last form of commodity, the form of money, this relation is done away with. We have: * units у units . Ζ units и units

of commodity of commodity . ,. ' of commodity of commodity

A В . = ν units of money С D

Money does not have use value in the normal sense, i.e. normally one cannot use money directly to satisfy human needs. It is not edible, it cannot be used as clothing, and so on. If one wants to speak of the use value of money, one might somewhat paradoxically say that its use value is to be the opposite of use value, namely, exchange value. Hegelian marxists have tried to read Hegelian conceptual development into the above four schemata. They begin with a supposed 293

ONTOLOGICAL INVESTIGATIONS

conflict/contradiction between two properties, use value and exchange value. (In my opinion these two kinds of value are quite simply two mutually dependent moments of each commodity; but Hegelians see them not only as distinct but as contradictory.) This conflict/contradiction is then assumed to be 'transcended' (or aufgehoben) in that the one side of the conflict, the exchange value, necessarily releases itself from the commodity and takes on an existence of its own, namely, as money. T h e appearance of money is to be understood on this Hegelian interpretation as arising not from situationally determined intentions with accompanying nonintended consequences, but as a necessary consequence of a contradiction between use value and exchange value which inheres in commodities. This, however, is Hegelian nonsense. T h e more developed forms do not follow logically from the less developed. But their less developed forms are conditions of existence for the developed form; the latter contain the former. In the states of affairs represented by the schema for the form of money, there necessarily exist the states of affairs represented by the schema for the general form, and so on. At the end of this chain of one-sided dependence there are the states of afTairs which correspond to barter. These one-sided existential dependence relations are wrongly interpreted by Hegelians as a developmental schema (cf. page 133). Money, however, cannot be completely ontologically understood in terms of these existential dependence relations. Special states of affairs connected with the intentionality category also play an important role. Marx himself strongly emphasizes that neither use value nor exchange value are properties inhering in things, but are relations. What kind of relations he has in mind he never makes clear; this lack I shall now try to remedy. No commodity has use value in itself; use value is always use value for someone. One has to make a distinction between direct and indirect use value which Marx himself never explicitly made. I f a commodity has direct use value, it can be used by the buyer in order to satisfy (in and of itself) some of the buyer's need(s). But if a commodity has indirect use value then it cannot in and of itself satisfy the buyer's need(s), but may well in another situation be exchanged for something which has direct use value. Direct use value might possibly be understood as a form of causal relation (i.e. the commodity satisfies a desire), but indirect use value can only be understood in terms of intentionality. I f a commodity has

294

NESTED INTENTIONALITY

indirect use value for someone, then that someone understands that the goods are exchanged with the aim of making yet another exchange. Let us look at some examples. We call the buyers and sellers P\, P2, etc. An exchange according to the simple value form can then be characterized in the following way (cf. the schema for ordergiving on page 279): (1) Ρ ι knows that В has direct use value for him, and believes that A has direct use value for P2. (2) Ρχ wants P 2 to recognize point (1). (3) Ρ ι wants Pi to exchange В for A because P 2 recognizes point (2)·

If the general value form is to fit into a similar schema, such a schema must look a bit different. We assume now that P\ is in possession of the universal equivalent E, and wants to have commodity B: (1) Pi knows (a) that В has direct use value for P\; (b) that Ε has indirect use value for P 2 ; (c) that Ε has direct use value for P 3 . (2) Pi wants P 2 to recognize point (1). (3) Pi wants P 2 to exchange В for Ε because P 2 recognizes point (2).

The buyers and sellers in these examples are ideal-typical in the normative sense. They do not intend to trick anyone, but want everyone to receive a use value at the right price. If we continue with this assumption, ordinary business with money (M) can be said to follow the schema below: (1) Pi knows (a) that В has direct use value for P\, (b) that Μ has indirect use value for P 2 ; (c) that some commodity A,B,C,D,. . ., has direct use value for P 2 ; (d) that Μ has indirect use value for P x who sells the goods which have direct use value for P2; (e) that some commodity A,B,C,D,. . ., has direct use value for P x .

295

ONTOLOGICAL INVESTIGATIONS

(2) Ρ ι wants P2 to recognize point (1). (3) Ρ ι wants P2 to exchange В for Μ because P2 recognizes point (2).

If Ρ ι is to be the ideal-typical merchant we have assumed him to be, he must have intentionality directed at an infinite progression. Since money does not have direct use value, it must be exchanged for something. He who has exchanged goods for money and cannot use the money has basically given his goods away without compensation. In reality, this need not be a question of an infinite progress, but only of a progression into an unknown future. But this is quite sufficient for the point I wish to make. Money is, like both language and machines, a human creation, but money is, from a categorial point of view, distinct from both of these in an interesting way. 31 Money is distinct from machines in that the potentially endless interaction which is inherent in money can only be understood via the intentionality category. One can easily imagine machines (or natural processes, for that matter) which, given a certain supply of energy, consist of an eternal repetitive process, for example a machine which endlessly moves circular metal plates from place to place. But these discs cannot be money, for the machine cannot have intentionality which is required in order for indirect use value to exist. That which is specific to money is not a potentially endless interaction, but an intentional endless progression. Without intentionality currency notes are only paper, just as a letter without intentionality is only a pattern of ink strokes. Money is distinct from language in that the material substratum plays a more important role. Money is a material part of a material exchange in a completely different way than the material substrata of linguistic signs are parts of linguistic interactions. Such substrata have no communication value in a communication. In the case of oral communication the substratum disappears more or less simultaneously with the arrival of the message. In the case of a written message, of course, the substratum remains, but the message itself can be read off by anyone who knows the language without that person's owning or standing in any intimate relation to the substratum. But in the case of buying and selling, as distinct from linguistic interaction, it is very important that the substratum is not dissipated and that it belongs to only one party in the interaction at a time. 296

NESTED INTENTIONALITY

Money is, it is worth repeating even today, a more noteworthy invention than most philosophers have realized.

15.8 M A C R O AGENTS I began this chapter with a few words about atomism and holism. It is now time to deal with a specific part of this general issue, namely the difference between microagents and macroagents. By 'microagents' I mean individual persons, and by 'macroagents' such entities as organizations, companies, classes, and nations. When one nation declares war on another and when a company raises the price of its goods, it is macroagents who are acting. But these macroagents have, in a way, microagents as parts i.e. participants. How do microagents constitute macroagents? The intentionality category is, in a sense, always connected to individual persons. There is no primitive macrointentionality. The subject-poles never unite into a single macrosubject-/>o/< directed towards a certain intentional goal. A microagent is always founded on the same substratum as a subject pole; it is the spontaneity category which distinguishes agents from subjects in general. The actions of microagents can of course be aggregated in the traditional sense. If 10 individuals bake 3 loaves of bread each, then the 'unit' consisting of these 10 individuals seen as an aggregate, can be seen as having baked 30 loaves of bread. But this does not make the bakers into a macroagent. When we now seek an understanding of what unites microagents to make a macroagent, we seek an understanding which refers to the macroagents' intentionality. On the pure action plane, as it manifests itself via the type of four-dimensional universale which was discussed in chapter 14, there is nothing special which distinguishes macroagents from patterns of a purely naturalistic type. Without intentionality, macroagents are reduced to a type of machine or organism. In the case of machines and organisms the parts add themselves to a whole with the help of spatio-temporal inclusive universals and the category of efficient causality. Macroagents, on the other hand, appear when a group of microagents has specific nested intentionality structures. As I have already indicated, a macroagent cannot have a subject-pole of its own which is distinct from the microagents's subject-poles. A macroagent cannot be defined in any other way than via the 297

ONTO LOGICAL INVESTIGATIONS

specific nested intentionality structures of the constitutive microagents. There are many different types of macroagent. I shall only bring out the structure of a few important cases which have a prominent place in political philosophy. Not in order to bring out everything that is involved in the cases at hand, but in order to show the importance of the category of nested intentionality. By means of this category one can explain the real difference between a society ruled by 'the general will' and a society ruled by 'the will of all'. T h e last distinction is due to Jean-Jacques Rousseau. Rousseau has often been accused of being an unclear thinker, and many have complained about unclarities in the concept of 'general will' and its relation to the 'will of all'. I shall not try to refute these partly correct accusations, but I shall show that two specific intentionality patterns can be connected in a natural way with the two concepts. Neither of these patterns, I shall also show, fits an absolutely totalitarian society, a society ruled by a 'Leviathan' in Thomas Hobbes's sense. However, it will also be shown that there are an astonishing number of similarities between the nested structures of 'the general will' and absolute authority. But first let us look at Rousseau's distinction. In what follows I shall only discuss cases where the macroagent in question has but two participants (= microagents). The generalization to more participants is in principle simple but cumbersome to write out, and I have earlier sketched the way such a generalization can be made (see pages 277-8). T h e intentionality structure of the general will is closely related to that of friendship discussed in 15.4. Assume that the microagents A and В both want В to do a certain thing d (= Bd), and that they want this because they are part of a macroagent with a general will. Then the following obtains (w = 'wants'):

(1) (2) (3)

A w A-*Bd WW A^>(B-+Bd) w w w A —» Bd-*))

В w B-+Bd WW В—* (A—* Bd) w w w B-* (£->· Bd)) [etc.]

O n level (2) A wants В to want what A wants on level (1); and 298

NESTED INTENTIONALITY

vice versa for B. Furthermore, on level (3), A wants В to want that A wants what he wants on level (1); and vice versa for B. This should be compared with the following, the structure of the will of all.

(1) (2) (3)

A w A-+Bd WW A-+{A^>Bd) w w w A {A -»(А-» Bd-+))

В w B—*Bd WW B^>{B-^Bd) w w w (B-* Bd)) [etc.]

In a structure like this A's want never becomes nested with B's want. A wants what he wants and В wants what he wants, and there is nothing more to say. I have chosen an example where both microagents want the same, they both think that В ought to do d. If their wants had been different they would have had to construct or rely on some mechanism, for example democratic voting, in order to get a want for the macroagent. The 'will of all' has a rock-bottom, or rather two, the wants of Л and В on level (1). But the corresponding wants of the general will (level (1)) are not such rock-bottom wants, since the will of Л is also supposed to conform to that of В and vice versa. There is no individual starting point, and that is the characteristic peculiarity of the general will. It ought to be noted, however, that such a will need not be static. The circle that A wants what В wants, who wants what A wants, requires only that A's and B's wants change simultaneously in the same way. It ought also to be pointed out that the structure of the general will presumably never occurs in a pure form. The structure is to this extent an ideal type. There exist, however, it seems to me, many real groups which have a half-conscious striving towards the structure of the general will, or would at least like to appear outwardly to have this structure. In such a group conflicts are not allowed, and no one can really distinguish his individual will from the will of the others. In some way there just arises a general will for the group. I now turn to the kind of macroagents where one microagent has absolute power and the other participants blindly follow. The latter have internalized the leader's authority in such a way that they 299

ONTOLOGICAL INVESTIGATIONS

want him to decide what they shall want to do. If A is the ruler and В the ruled, and A wants В to do d, then similarly В wants В to do d. This is level (1) in the structure for the general will. Let us therefore see whether, on the higher-order levels, we can discern the difference which obviously must exist between the general will and an absolute authority. On the second level in the general will, A wants В to want Bd, but that is also what our totalitarian leader wants, because he wants В to want to do what he is ordered to do. Similar arguments apply to the third level. In order to make apparent the difference we are interested in, we have to bring in the reasons A and В have for their wants. We shall once again use the notation [r. . . .], only this time for symbolizing reasons not for an action but for a want.32 Taking such reasons into account, the general will looks as follows:

(1) (2) (3)

w w (A —* Bd) [r.B —» Bd] WW w A-> (B-* Bd) [r.A -> Bd] w w w w A—*(B—*{A—* Bd)) [r.B—* Bd] [etc.]

(1) (2) (3)

В w w (B-*Bd) [r.A-^Bd] WW w B-* (A-* Bd) [r.BBd] w w w w B—*(A—> (B-* Bd)) [r.A Bd] [etc.]

The nested intentionality structure of a macroagent with absolute authority is similar, but differs in one important respect.

(1) (2) (3)

w w (A^Bd) [r.A^Bd] WW w A-* (B-* Bd) [r.A -> Bd] w w w w A -> ( £ - » [A Bd)) [r.A -* Bd] [etc.] 300

NESTED INTENTIONALITY

(1) (2) (3)

В w w (B-+BJ) [r.A^Bd\ WW w B-* {A—* Bd) [r.A -> Bd] w w w w B—> (Л —» {B-* Bd}) [r.A Bd\ [etc.]

The first schema brings out very clearly the circularity in the general will. A wants something because В wants it, but В wants it became A wants it. Such a circularity does not exist in the latter structure. The difference between the structures is that in the former case we have both w [r.A -> Bd\ and w [r.B —* Bd] but in the latter case we have only w [r.A-* Bd]. This means that in the authoritarian structure the only kind of reason that exists is a reference to the want of A, to Leviathan himself. The abstract schemas above do not say anything about the concrete social relations which will normally clothe the intentionality structures described. An absolutely authoritarian macroagent, as well as a macroagent with general will, can be based on tradition, i.e. A and В may be socialized by tradition and come to adopt their respective intentionality patterns in this way. But, in the former case, it may simply happen that a very strong person creates an intentionality pattern within a group by means of his charisma. One can also imagine the structure arising via a contract of the type described by Hobbes. To avoid a war of all against all, everyone decides to give A absolute power, and make a contract to this effect. Yet another thing to note is that both of the intentional structures I have presented allow that the 'power', of the absolute sovereign or the general will, may be employed in an egoistic or in an altruistic fashion. There is nothing in the schema for the absolute 301

ONTOLOGICAL INVESTIGATIONS

authority which says what he is to will. He can thus satisfy the structure by taking into account either his own welfare or that of everyone else, when he decides what everyone is going to do. T h e same holds for the general will. T h e schema above sets out the form of the general will. This makes it possible to see that this form is neutral with respect to any questions about the people in whose interest the decisions are made. As I see it, in the discussions of Rousseau's general will, too much weight is usually put on the fact that the general will must make decisions which are for the good of the entire group concerned. The form for the general will can in fact be satisfied in a group which is completely oriented toward serving one particular person. T h e formal similarity which exists between the two structures discussed is probably of sociological interest. It is a remarkable fact that many religious and political sects whose original ideal was that the group should be a general will and have power-equality among its members, have smoothly turned into strongly authoritarian groups. O n e aspect of this phenomenon is that the formal similarity between the structures in question helps the members to overlook or repress the real change involved. I have already pointed out that the structure for the general will is a n ideal type but the same goes for the other structures as well. Most actually existing macroagents exist as such only with respect to certain types of action, while I have here described them as existing unconditionally for all types of action. This conditional existence is obvious in the case with organizations and associations, but it is also true of the modern nation-state with its distinction between public and private life. Most macroagents also have delegated authority; someone has power via a superior power. I have not taken account of this aspect of macroagents. Moreover, macroagents are often rooted in (the threat of) pure violence. M a n y microagents become part of a macroagent not because their intentionality fits the intentional structure of the macroagent, but because they are forced to obey the macroagent and be a part of it. O n e variation of the same theme is obtained when a microagent himself consciously chooses to be part of a macroagent in spite of the fact that his intentional structure diverges from that of the macroagent. H e becomes part of the macroagent because he directly benefits from it. His actions are part of the macroagent's, but his soul is not in it. 302

NESTED INTENTIONALITY

If we return to ideal-typical macroagents we may note another point of sociological and anthropological interest. Let us first once more write down the first level in the 'general-will structure' (1) and in the 'Leviathan-structure' (2), and then in the same notation also the 'will-of-all structure' (3). We get:

(1) (2) (3)

А W W {A —* Bd) [r:B —* Bd] W W (A -> Bd) [r.A Bd\ W W (A Bd) [r.A -+Bd\

В W W {B -> Bd) [r:A Bd] W W [B Bd) [r.A Bd\ W W (B-* Bd) [r.B Bd\

In the last macroagent it is impossible for the microagents to see themselves only as part of a macroagent. They must also see themselves as microagents, i.e. they must as individuals have contentdetermined intentions and expressions of will. In the formal structure this is revealed by the fact that i4's and B's reasons for their wants are their own will. The first two structures behave in this respect differently. The circularity in the intentional structure of the general will clearly has the contrary character; it hides the fact that a microagent can have a determined will as a microagent. In a macroagent with absolute authority, the sovereign must see himself as a microagent, but his subjects need not see themselves as microagents; they can see themselves as parts with no will of their own. (The ruler can, of course, see himself as the representative of God, or some such thing, but then it becomes instead a Active person, God, who must conceive of himself as a microagent.) This last point is intended to show that certain macroagents can in one sense be primary in relation to their constitutive microagents, namely in the sense that the macroagent's specific wants have to be fixed before the miroagents' specific wants are determined. The microagents are however always ontologically primary, for they constitute the subject-poles in the nested intentionality which constitutes a macroagent. The distinction made between these priority orderings is sufficient for my ontological system to harmonize with the view that, historically, man first saw himself as part of 'non-individualistic macroagents', and only gradually began to see himself as part of 'individualistic macroagents'. All of anthropology testifies to the view that man saw himself as part of a group before he began to see himself as an 303

ONTOLOGICAL INVESTIGATIONS

individual in the modern sense. In spite of this, the tendency still exists amongst many philosophers to regard the microagent who sees himself as a microagent as in all respects the fundamental agent. In game theory, to take one example, such a view expresses itself in the belief that 'co-operative games' are always parts of larger 'nonco-operative games'. 33 It is high time for game theory to accept anthropology's results. 15.9 NATURE, MAN, AND SOCIETY It is now time to summarize the theses of the present theory of categories. In the remaining sections I shall then discuss some epistemological problems. The point of departure for this ontological system has been the difficulty of uniting in a reasonable way the modern conception of nature (nature = 'dead' matter) with the belief that certain concepts which do not refer to such a nature, (e.g. 'action' and 'intention') also refer to something which really exists, and as such demand a place in science. T h e conception of nature contained in this system is founded on the fundamental category of space. Space-time is the only entity which fulfils the classical definition of 'primary existents': it exists in and through itself. Everything else is dependent on space-time for its existence. Things cannot exist if they do not exist in container space. The same holds for man and society. Nature, man, and society exist in space; and so nature cannot be delimited with the help of the category of space. Any such delimitation has to be made with the help of the intentionality category: Nature is identical with the non-intentional part of reality. This theory of categories diverges from the conception of nature which in modern physics is assumed to imply on two points. First, nature does not consist only of things, substances, (genuine) properties, and relations, but also of tendencies. Second, nature exemplifies universale which are necessarily extended in time. Both organisms and machines are thus, from this perspective, parts of nature. They exemplify the same kind of universale in spite of the fact that their way of coming into existence is completely different. Machines, as distinct from organisms, are constructed by people, they are artefacts; but that is the only ontological difference so long as one assumes that organisms lack intentionality. This equality between machines and organisms is completely clear on the 304

NESTED INTENTIONALITY

practical level within modern medicine; I am thinking of the use of artefact organs. The conception of man presented here takes its departure in the claim that what we call actions are a truly special type of phenomenon. First, they are specific types of four-dimensional universals. In this sense animals, too, can perform actions. This aspect of actions does not tell us what is distinctive about man, but it is none the less important for an understanding of man. The most important category for the understanding of man is intentionality. This I maintain in spite of the fact that intentionality cannot be identified either with consciousness or with rationality. It is a deeply rooted prejudice that what is most important for the conception of man must be what is characteristic only of man. Our modern understanding of man as a natural being owes much to the categories employed within physics, chemistry, and technology. The way to a deeper understanding of man as man will, I believe, depend on an exploration of the category of intentionality. Whether this will, at the same time, lead to a deeper understanding of the higher animals is in this context a subordinate question; much speaks for these animals having intentionality at least in the form of perception. We understand, literally, mortally important phenomena by having knowledge of our partial identity with dead nature. An increased understanding of the type of intentional structures and patterns I have described may, I hope, in a similar way create preconditions for a better human life whether or not the same structures and patterns exist among other animals. One specific aspect of intentionality is its ability to be a 'connection at a distance'. That there is such a phenomenon as connection at a distance is actually one of the things which are most obvious in everyday life. But from a modern natural-scientific point of view this appears to be quite mysterious, or indeed occult. As I have drawn the lines of demarcation, this type of connection does not belong to that part of reality which nature constitutes. But reality includes intentional phenomena as well. That which is real exists in space and time, and intentional phenomena exist in space and time just as much as non-intentional phenomena. The soul (= the subject-pole in intentional phenomena) occupies the same place in space as does the body. Irreductive materialism solves in an elegant way the mind-body problem. The solution is based on the category of one-sided existential dependence. 305

ONTOLOGICAL INVESTIGATIONS

I have now mentioned three ways in which the concept of man represents something which cannot be captured in the framework of traditional natural science: (1) actions exemplify universals which are necessarily extended in time, (2) the intentionality category, and (3) the level ontology, which makes it possible for the 'soul' to be placed in space and time. 34 Yet another point of the same degree of importance ought to be mentioned. The present ontological system also contains (4) a category of spontaneity. This makes it possible to ascribe to man a certain amount of so-called free will and an ability to transcend what actually exists. Man need not be seen as completely determined by his biological nature and social situation. That man is partly so determined, I regard as an undeniable fact. 35 It is important to guard against an easy misunderstanding regarding (1) and (2). It is tempting to see all real actions as the outflow of a preceding intention ('prior intention'). Such a perspective tacitly presupposes that intentional phenomena always constitute a kind of state, and that they in themselves cannot be actions. However, I have nowhere adopted this perspective. Quite the contrary - it seems to me to be obvious that certain intentional phenomena are actions and in every respect irreducibly extended in time. Many actions such as thinking and deliberating are actions which cannot possibly be regarded as outflows of a 'prior intention'. To try to explain what one is thinking by making a reference to a prior intention to think what one is thinking, leads nowhere. There must exist thinkings which do not have their basis in a prior intention. 36 Certainly, one often first has to have an intention in order then to be able to act, but sometimes one must act without intention in order for new patterns to become visible, patterns which when they have become visible can make up the content of intentions. The intentionality category is not only the key to an understanding of man, but also the key to the understanding of society. Man and society cannot be understood independently of one another. As Marx points out in the famous passage from the sixth of his Theses on Feurerbach: 'the human essence is no abstraction inherent in each single individual. In its reality it is the ensemble of the social relations.' 37 Can this aphorism be given any reasonable meaning? Can the essence of man exist outside the separate individuals? The only 306

NESTED INTENTIONALITY

category which can give meaning to Marx's thought is the intentionality category. The essence of human beings clearly must in some way be anchored in individual people, otherwise it would be meaningless to call it the essence of man. Intentionality and the accompanying distinction between the subject-pole and the intentional correlate can help us here. The capacity to have intentionality, i.e. the capacity to be the subject-pole in an intentional phenomenon, must be 'inherent in each single individual', but the human essence perhaps includes some states of affairs which belong to the intentional correlates of presentational intentionality. In that case the human essence does not exist only in the single individual. In order to get a grip on this problem one has to distinguish between two human natures, a natural nature (or biological nature) and a social nature. The former exists by definition only in the spatio-temporal volume occupied by an individual body during its existence. Man's biological nature includes not only such things as genetic structure and the species-specific ability to reproduce, but also the capacity to perform or have intentional acts, which of course is not identical with intentionality as such. The latter belongs to an upper ontological level, while the capacity belongs to the substratum. Presentational intentionality can have both nature and other intentional phenomena as intentional correlates. It is only in the latter case that there are states of affairs containing human properties which partly lie outside a single person. It is only then that the type of nestedness occurs which makes certain states (e.g. shame and pride) wholes which must include at least two participants. This characteristic is most clearly seen in a group with a general will: A wants what В wants, and A wants В to want what A wants, etc., and vice versa (see the analysis in 15.8). If a determinate content of the will has appeared here, it will be reproduced. The will then exists in both A and B, but it is nevertheless no will 'inherent in each single individual' but a will which is 'the ensemble of the social relations'. The will can only be understood as a nested intentionality pattern which contains several participants. This line of thought can also be illuminated by looking at the conventionality of linguistic signs. Today, when a large part of the world's population knows at least a little of a foreign language, the 307

ONTOLOGICAL INVESTIGATIONS

thesis of the conventionality of the linguistic sign appears to be a truism. Most people understand that the same concept or thought can be related to completely different graphic symbols or completely different sounds. But recognition of this truism can easily obscure an important problem. For a bilingual person the linguistic signs are genuinely conventional. He easily exchanges one type of sign for another. But for the person who is not bilingual, linguistic signs are conventional only in a theoretical sense. They are substitutable in principle, but that person cannot actually make the substitution. This fact requires an explanation. Why can we not easily exchange linguistic signs in the same way we can change clothes? To a certain extent this non-substitutability can of course be explained in the traditional way by appealing to diverse conditioning mechanisms, i.e. by giving an explanation on the substratum level. But may there not also exist an explanation on the intentionality level itself? Let us compare the linguistic situation with a situation involving a general will. It is not clear from the structure of the general will what it is that the general will wills. The will's content is in this sense conventional. It is not, however, genuinely conventional for the microagents who normally constitute a general will, but one can imagine a secondorder general will - one where those involved not only have but also consciously want the whole of the general will's structure. They see the structure from the outside in the way in which it has been described in the present chapter. If they are able to do this, then they are like a bilingual person. They can choose a certain determined content for the general will, or consciously go from one group with a certain general will to another group with another particular general will. The conventionality in the general will is genuinely conventional and not merely theoretical. T h e conventionality of linguistic signs is similar to the conventionality of the content of the general will. There was a time when people quite simply could not imagine languages with signs which differed from their own. Such people, I think, were in the same type of situation as members of a group with a general will where the structure of the general will is not recognized. That this is the case is due to the fact that the linguistic signs are in a nested intentional structure similar to the one the general will has. If a certain graphical pattern is to be a sign for A, then it must also be a 308

NESTED INTENTIONALITY

sign for В, С, D, etc. And it becomes a sign for A because A sees that it is a sign for В, C, D, . . ., and sees that В, C, D, . . . see that it is a sign for A. This explanation is very sketchy, but it is an explanation on the intentionality level. I shall not even try to enter into a discussion as to how this sketch is to be filled out in more detail. I want only to explain the way in which it is reasonable to state that a certain language is not 'inherent in each single individual' but is 'the ensemble of the social relations'. A language is a human 'property' which at one and the same time exists both within and without each single person. When one understands the conventionality of the linguistic sign via structures on the intentionality level itself, one obtains at the same time a certain insight into the way a person's nature can be a 'sum' of social relations, where relations are something completely different from external, internal, and grounded relations. In biological nature there is a general speech capacity in the individual person; but this capacity can only find expression via conventional signs. A social nature thereby comes into existence, a language which cannot be said to exist only in individual persons. In principle one can regard the basic human drives, e.g. hunger and sexuality, in the same way as we regarded language. In biological nature there exists not only an undetermined linguistic ability, but also an undetermined drive for food and similarly an undetermined sexual drive. The objects of the drives are, from a biological point of view, conventional. This is not a fact which is stranger than the fact that a machine can be lubricated by many types of oil, and nevertheless function perfectly. But this indeterminacy can be determined, partly on the basis of the nestedness of certain intentional patterns. And this occurs in such a way that many biologically possible ways of satisfying the drives become more or less impossible. It is only in very extreme crisis situations that the pattern is broken. Who in our culture can eat raw mice? How many can be sheer will-power change the character of their sexuality? Such questions can be multiplied. 'Above' our biologically given nature there is another, a social nature, and that nature cannot be understood without the concept of'nested intentionality. I should like to rewrite Marx's aphorism about the social nature of the human essence in the following way: 309

ONTOLOGICAL INVESTIGATIONS

It is of the human essence to have two essences, one which is inherent in each single individual, and one which is not. The latter, in its reality, is an ensemble of relations of nested intentionality. T h e point I have made can be made more simply if one speaks not of the human essence, but of one person's nature, thereby indicating that different people can have different natures. Biological nature can then be identified with all of the properties and functions of a specific body, while social nature becomes something like the cultural, national, or class identity of the person in question. Discussions about the relation between the individual and the society are nowadays, more and more, carried on in terms of the three models below (Figure 15.2), which I shall briefly comment on. Following Roy Bhaskar, I shall simply call them Models I, II and I I I , but for those knowledgeable about sociology, they may be associated in turn with Max Weber, Emile Dürkheim, and Peter Berger. 38 In the first model, arrows represent intended actions, which does not mean that the intentions are always realized. But if something is to be a social object, it must be the result of meaning-laden h u m a n behaviour. What is wrong with the first model is that the link between the individual who acts and the result of his action is too weak. The relation between the action and the result is like that which obtains when someone waters a lawn. The result, the wet lawn, might just as well have been achieved by nature herself. T h a t there was no rain and that therefore some human intention was required to make the lawn wet is an external contingent aspect of the

Model I

Model II

Society

Society

Individual

Individual

Figure 15.2

310

Model III

Society

Society

Individual

NESTED INTENTIONALITY

result. Social objects and social individuals are spatially distinct; and a philosophy without the intentionality category will never be able to connect individuals and their products so as to obtain a social unit. Model I has, moreover, the problem of making social objects too voluntary. There seem to be no societal impediments to action. The arrow in Model II does not represent the same thing as in Model I. Now the arrow represents a causal relation. Society causes the individual to behave in different ways. Here, too, the bond between society and individual is too weak. Causality is conceived of as efficient causality, i.e. the cause is completely distinct from its effect. In Dürkheim himself this is very clear. When he argues for the existence of social facts distinct from psychological facts, he emphasizes strongly that social facts are external in relation to the individual. 39 Linguistic signs, currency, and so on — all such things exist outside of me in space. In this, Dürkheim is of course right, but he seems to be committing a simple 'fallacy of composition'. It is possible for each individual in a society to see society as something which exists both outside and independently of him, but it is not possible for all individuals collectively to do this. 40 To imagine a society completely without people is impossible. Both Models I and II are of course based on and get their plausibility from a number of true observations. We are sometimes aware that we freely create products which become social products, and we often experience just as clearly the kinds of social constraints which Model II takes as its point of departure. But both models seem to be based on the (mistaken!) classical conception on which intentional phenomena are conceived of as thoughts in the heads of people, and where all relations amongst people, and between people and nature, are either purely spatial or variants of efficient causality. In such a conception one can either emphasize free creation of intentions 'in the head', intentions which then, via efficient causality, realize themselves in material products distinct from the intention; or one can emphasize the pressure exerted on the individual person by that which is spatially external. If the society employs pressure, the pressure must be spatially external. To combine the arrow in Models I and II in order to get Model III solves nothing. I agree completely with Bhaskar's comment: 'In seeking to avoid the errors of both stereotypes, Model I I I succeeds 311

ONTO LOGICAL INVESTIGATIONS •

Society

II socialization

7Ϊ II reproduction/ •I transformation

''

II

II

Individuals

Figure 15J only in combining them.' 41 Peter Berger, who is one representative of Model III, describes the arrow from the individual to the society as symbolizing 'objectivation', while the arrow from the society to the individual is to symbolize 'socialization'. The problem is that the socialization must take place completely within the single individual, while the objectified product must be placed completely outside the single individual. It once again becomes impossible to understand what constitutes a societal unit. In Berger's case, one finds only external and grounded relations among the individuals, objects and states of affairs which constitute a society. The difference between machines and society turns out to be that society contains a type of parts which, in the space-time they occupy, also contain thoughts. Bhaskar, who has made critical comments on all three of these models, has himself suggested a fourth alternative which he calls 'The transformational model of the society/person connection'. Graphically, it looks like Figure 15.3.42 This model is, in some respects, an advance over its competitors; but some of the problems I have pointed out for the other models are problems for this one, too. I shall comment first on its strong points and then on its weakness. According to Bhaskar, it is a mistake to say that man creates society; man only 'reproduces' or 'transforms' society. This line of thought can be expressed by saying that for man there always already exists a societal matter in the Aristotelian sense. Just as people cannot create statues from nothing, but require a matter to transform, so they cannot create a specific societal institution from nothing, but require some societal matter to transform. As a description of how changes actually occur, Bhaskar's perspective is useful. Changes of language, of food-eating customs, of sexual patterns, even large political changes, are better understood as transformations of pre-existing social structures than as totally new 312

NESTED INTENTIONALITY

creations. But Bhaskar avoids going into the question about what this societal matter consists of. Such a matter must of course be in some way distinct from natural matter, and it cannot, like nature's prime matter, have existed always. It cannot have arisen historically before man. Another strong point of Model IV — one which, however, cannot be read off from the figure — is the assertion that, when a social agent produces something, he often reproduces at the same time a certain societal state of affairs. Bhaskar calls this the dual character of praxis. Production and reproduction are moments of the same action. When you buy something you produce private satisfaction if you are lucky — but at the same time you participate in the reproduction of the currency used. When you talk to someone you produce messages and help to reproduce language. If you talk in a new but yet understandable way, you still produce messages but now you help to transform language. Currency and language set limits on what can be done with their help, but they do not determine exactly what will be done. They can thus be reproduced or transformed at the same time that something specific is produced. The dual character of praxis has its counterpart in a duality of structure. Currency and language, to continue with the same examples, are at one and the same time both condition (or medium) and outcome of actions. Without currency some kinds of transactions would not be possible, but without actions which make use of the currency the currency would not exist. Without a specific language some kinds of communication would not be possible, but without speech acts making use of the language the language would not exist. This kind of conditioning, it is important to note, is both enabling and constraining. A certain language enables types of communication which would be impossible without language, at the same time as it may make some other types of communication totally impossible. In this respect it is similar to inanimate things. A roof enables you to get shelter from rain but limits the possibilities of sun-bathing at home or of throwing balls up in the air. Actions (praxis) are performed by persons (agents) and persons have intentions and make deliberations. A person could, most of the time, have acted otherwise than he acted. The same thing, however, is not true of social structures or societal matter in 313

ONTO LOGICAL INVESTIGATIONS

general. T h e y do not act; they are conditions and outcomes of actions. There exists, Bhaskar says correctly, an 'ontological hiatus between society and people'. 4 3 Bhaskar describes a difference between persons and society, but, unfortunately, he does not probe deeply enough into the connection between agents a n d societal matter. A specific form of natural matter is very often, like societal matter, a condition for a certain action. But Bhaskar says nothing about the way in which societal matter is distinct from natural matter. From an intuitive point of view, societal matter ought to have a more intimate relationship to agents than does the latter. I n spite of Bhaskar's effort, his analysis also means that in the end he is committed to individuals and society standing against one another, not in one another. (This is also clear from the diagram of Model IV.) W h a t is the alternative to Models I - I V ? I can only repeat what I have now said many times. There is only one way to unite individuals and society, and that is with the help of the intentionality category and different types of nested intentionality. Without nested intentionality there would be no societal matter in contradistinction to natural matter. Both Berger and Bhaskar are (like myself) partly inspired by marxism. Such an inspiration, however, has its dangers, and this might be the explanation why neither of them gives intentionality the same importance as I do. Marxism is materalistic, and a solution of the problems of the sort I have discussed in terms of intentionality can easily appear to be far too idealistic for a marxist. But such a marxist has not grasped the content of irreductive materialism — and many marxists have unfortunately not done so in spite of the fact that just such a n ontology belongs to this tradition. A subject is something with both a body and intentionality. And my descriptions of the intentionality patterns connected with the objects of drives and linguistic signs have, I hope, made it clear that both the objects of drives and linguistic signs require material substrata. T h e material and the intentional are woven together. T o satisfy my hunger I need natural matter, but a natural matter which is mediated by my societal nature. 4 4 T h e category of intentionality is fundamental to the existence of m a n and society, and it does not imply idealism.

314

16

EPISTEMOLOGICAL POSITIONS

The aim of this book is to present an ontology, not an epistemology, but (to quote Nicolai Hartmann): 'in the last analysis the cognitive relationship is itself an ontological relationship.' 1 Epistemology has to be about existing cognitive types of acts, which means that these acts must belong to one or several categories in a theory of categories. Epistemology is not something outside ontology, it is rather a regional ontology. I believe in ontology-centred philosophy. In what follows I shall comment upon three epistemological issues in the light of the ontology set out in previous chapters: (1) the correspondence theory of truth; (2) the claim that there are truths which are synthetic a priori, and (3) the question of the epistemological status of the ontological system itself.2 16.1 A CORRESPONDENCE T H E O R Y O F TRUTHLIKENESS A theory of categories is so abstract a theory that it subsumes scientific theories under it. From this point of view it is not necessarily a problem that, through history, different scientific theories have replaced each other, and, in all probability will continue to do so. If all the (competing) scientific concepts used refer to phenomena which fall under the categories of a certain ontological system, then the historical relativity of past and present scientific truths does not affect the truth-claim of the ontology. However, there may be clashes between science and ontology, a problem which will be discussed in section 16.3. In this section I shall discuss the correspondence theory of truth itself.

315

ONTOLOGICAL INVESTIGATIONS

T h e immanent realism of my system, together with the category of intentionality, implies some kind of correspondence theory of truth. Intentional acts may be directed towards phenomena which exist independently of intentionality, and intentionality has a modality of satisfaction. Of course, this is true even for representational intentionality, which is the kind of intentionality at work in scientific theories. T h e relativity problem I have in mind can be stated in the form of a question: Is the correspondence theory of truth valid, if we take it for granted that not even our current scientific theories can be considered to be true? We cannot regard them as true if we think they are going to be revised in the future, but it seems equally impossible to regard them as false! Must we not declare them neither true nor false? Must we not throw the correspondence theory of truth away? T h e solution to this conflict between a historical and a presentday perspective is easy to find. Intentionality's modality of satisfaction allows of degrees; besides satisfaction and nonsatisfaction there is partial satisfaction. In the case of statements this means that there can be degrees of truth. T h e theories we normally speak of today as being absolutely true, should really be considered more true than the superseded theories, but less true than some possible future theories. The old polar concept of truth should be replaced by a comparative concept of truth, 'truthlikeness'. In the last decades or so there has been a revival of epistemological relativism (especially among social scientists and people within the humanities), but all the same the overwhelming impression is that a concept of truthlikeness lies implicit in most discussions when scientists talk about scientific development. It is therefore surprising that few materialistic philosophers have found it worthwhile to stress the notion of truthlikeness. I shall mention and quote two who, on this issue, make u p an unholy alliance. One is an outstanding marxist, the other an outstanding anti-marxist. Both Lenin and Popper have argued for the importance of 'truthlikeness'. 3 Lenin uses the terms 'relative truth' and 'approximative truth', Popper is most fond of 'verisimiltude'. T h e idea of truthlikeness is a very straightforward idea, but there are some stumbling blocks which may explain why it has not hitherto become a common or accepted philosophical notion. A confusion of truthlikeness with criteria for determining truthlikeness presents a stumbling block to many. Both Lenin and Popper are 316

EPISTEMOLOGICAL POSITIONS

clear about this problem. Both of them reject what Popper has nicely labelled 'criterion philosophies'. Popper writes: I believe that it is the demand for a criterion of truth which has made so many people feel that the question 'What is truth?' is unanswerable. But the absence of a criterion of truth does not render the notion of truth non-significant any more than the absence of a criterion of health renders the notion of health nonsignificant. A sick man may seek health even though he has no criterion for it. An erring man may seek truth even though he has no criterion for it.* Within classical dialectical materialism philosophers have often spoken about 'the criterion of practice', but what exactly is meant by a 'criterion' is usually unclear. Lenin, however, is clear. It is not a criterion in the ordinary sense, i.e. something which decides conclusively for or against. It is an 'indefinite' criterion. He writes: Of course, we must not forget that the criterion of practice can never, in the nature of things, either confirm or refute any human idea completely. This criterion is sufficiently 'indefinite' not to allow human knowledge to become 'absolute', but at the same time it is sufficiently definite to wage a ruthless fight on all varieties of idealism and agnosticism. 5 The two quotations underline the fact that in philosophy the distinction between criteria and what they are criteria of is very important. Two objections might be made to the Lenin-Popper position: (1) can we really understand a concept for which we have no criteria?; and even if this is possible, (2) what is the point of having a concept if one does not have criteria for its application? The assumption behind the first question may be that a concept is identical with the criteria for its application. Whether or not this is the case I think that in order to understand a concept one must be able to apply it in at least some cases. The scope of a concept I have will probably not be restricted to my criteria of application, but in order to grasp the concept I must nevertheless have access to some criteria. In the case of 'truthlikeness' I claim that we have enough criteria to understand the concept. The truthlikeness of a statement is a case of partial satisfaction of representational intentional acts. To establish the value of the satisfaction modality of a statement like 'He is in room number 4' 317

ONTOLOGICAL INVESTIGATIONS

is always a two-step affair. The first step is to relate representational intentional acts to presentational, for example in the case at hand by going to the room to see whether he is there. The second step is to find out the real value of the satisfaction modality of this presentational act. Now, my claim is that independently of the epistemological problems involved in the second step, the first step is as clearcut as anything can be. We can without trouble establish a hypothetical satisfaction value. If the observations are correct then the truthlikeness of the statement is such and such. In this sense we do have criteria for truthlikeness, and that is all we need in order to understand the concept. This conclusion enables us to deal with the second question. Usually when one speaks of 'truthlikeness' with regard to scientific theories, the intention of the term goes well beyond the criteria by means of which it has been understood. So, what use does it have in science? My answer, in short, is that we need an explanation of the fact that discussions about competing theories are meaningful. They are meaningful because the theories may have different degrees of truthlikeness; this goes for contemporary theories as well as ancient and future ones. The concept of truthlikeness affords us no tool with which we can settle scientific discussions, but nothing requires that one and the same concept shall both explain why discussions are meaningful and how they are to be settled. 'Truthlikeness' is a relatively empty, but nevertheless necessary notion. Popper has rightly likened it to a regulative idea. 6 Look at the following inference which might be used in an attempt to show that Thomas Kuhn's incommensurability thesis combined with the correspondence theory of truth yields a relativistic world view; A means Aristotelian physics and G Galilean physics: 7 (1) There is no way to translate A into G, though a believer in A could learn G. (2) Therefore, the falsity of Л cannot be argued for in G. (3) This means that both A and G have to be regarded as true, i.e. the notion 'true' is always relative to a conceptual system. (4) The world makes beliefs true. (5) So different worlds must make A and G true. 318

EPISTEMOLOGICAL POSITIONS

A conclusion may be blocked in several ways. In this case, one of them is simply to substitute 'truthlikeness' for 'truth'. Then, without bothering about this presumed inference, one can maintain that it is the same world which makes both A and G truthlike. The concept of truthlikeness differs in important respects from that of truth. The polar concept of truth may lead to relativism in the way indicated, but it may also lead in the opposite direction. It may make a person believe that if he is right his opponents must be totally wrong. But if one considers a case of theory conflict armed with the notion of truthlikeness, one can without any contradiction maintain the following five propositions: (1) Both one's own theory and that of one's opponent have probably a non-zero degree of truthlikeness. (2) One's own theory has a greater degree of truthlikeness than the other. (3) The other theory, that with the lesser degree of truthlikeness, captures something in reality which one's own theory does not capture. (4) Therefore, one's own more truthlike theory has something to learn from the other less true one. (5) As a consequence of (4), it might be well worth continuing the discussion. Amazingly few philosophers, particularly within the postpostivist and post-structuralist traditions, seem to have realized the simple fact that the concept of truthlikeness is quite compatible with pluralism in both science and philosophy. A correspondence theory of truth and truthlikeness is compatible with pluralism. Another argument against my account of truth is the rejoinder that a correspondence theory of truth, even of truthlikeness, presupposes a pre-established harmony between the world in itself and our cognitive acts. And, the argument may continue, the theories of categories presented above seems not to presuppose any such harmony. The last statement is true in the sense that the category of intentionality could be cancelled without any contradictions arising in the rest of the system. There is nothing contradictory in the idea of an unknowable world or a world without intentionality. But my ontological system is, so to say, a system after the event. It is an indubitable fact that there are

319

ONTOLOGICAL INVESTIGATIONS

intentional phenomena in the world. With intentionality comes knowledge, or at least presumed knowledge, which makes it look as if there is a pre-established harmony between the world and the knowing subjects. This point has been made by Collier: It is a useful exercise for an epistemologist to reflect on the manner in which Darwin's theory dealt a death-blow to the idea of teleology in nature. There appears to bean 'original complicity' between species and their environment - because where the two are ill-suited, the species die out. Nature will always produce the appearance of design if it produces the appearance of anything, because where it doesn't produce that appearance, there will be nothing for it to appear to. There is no necessity about the knowability of nature, any more than about the existence of Descartes; but whenever Descartes thought he existed, he existed sure enough. And anyone who asks: Is the universe knowable? lives in a knowable universe alright. 8 The last remark I want to make about truthlikeness, is that this idea has nothing at all to do with a rejection of absolute truth. If a statement has a certain truthlikeness, then it is absolutely true that this truthlikeness obtains.

16.2 A N E W WAY O F L O O K I N G A T T H E S Y N T H E T I C A PRIORI In modern philosophy there is a line going from Kant to Quine which could be described by a Quinean as follows. Kant thought it necessary to distinguish between three kinds of truths: analytic, synthetic a priori, and synthetic a posteriori. According to Kant, both Euclidean geometry and the fundamental principles in Newtonian mechanics were synthetic and true a priori, but scientific development has shown that both are false. T h e philosophical upshot of this is that there can only be one kind of synthetic truth, a posteriori truths. Later on Quine showed that it is also impossible to distinguish analytic and synthetic statements from each other, 9 which means that there can only be one kind of truth, a posteriori truths. I do not believe in this Quinean story, and I shall sketch an alternative view which allows for the existence of analytically true statements as well as for the existence of statements which are 320

EPISTEMOLOGICAL POSITIONS

synthetic and true a priori, albeit a priori in a modified sense. In whatever way Quine's general arguments about analyticity are understood, there is a kind of indubitable counterexample to the thesis that all truths are true a posteriori. I am thinking of abbreviations established by means of stipulative definitions. If, for instance, a physicist writing a book stipulates that 'cm' equals by definition centimeter and that '#-' equals by definition electron, then nothing in the world could make his statements 'cm = centimeter' and = electron' false. Quine himself in a way admits this. He writes: 'In formal and informal work alike, thus, we find that definition - except in the extreme case of the explicitly conventional introduction of new notations hinges on prior relations of synonymy.' 10 If one has accepted that stipulative definitions can give rise to analytically true statements, then it seems difficult to reject in principle the possibility that a natural language can gradually develop or introduce a synonymy between two terms which functions in exacdy the same way as an explicitly introduced stipulative definition functions. Language then works as if there were such a definition. However, it seems that such synonymies are very short-lived. Natural languages seem to have a horror synonymi. As soon as there is an opportunity the different terms begin to take on different meanings. The classical example of synonymous terms in introductory books in analytic philosophy is 'bachelor = def unmarried man', but these terms are beginning to lose their synonymy — at least in Sweden. 'Bachelor' tends to refer only to men who never have married, whereas 'unmarried man' tends to refer both to those who have never married and to those who have married but then get divorced. Most of the terms which in chapter 1 I called 'conventionalist general terms' can be regarded as stipulatively defined, and so as terms that give rise to analytic statements. If 'blue' by stipulation, or naturally, has become synonymous with a disjunction of terms referring to different hues of blue, then statements like 'everything light blue is blue' and 'everything which is turquoise is blue' are analytically true statements. One of Quine's arguments in 'Two Dogmas of Empiricism' is that if one tries to define 'analyticity' by means of 'synonymy' the procedure is viciously circular. According to Quine, one cannot make 'synonymy' clear without presupposing the concept of 321

ONTOLOGICAL INVESTIGATIONS

'analyticity'. In a way I agree with Quine. The concepts 'terms' and 'statements' are correlative concepts, i.e. they presuppose each other mutually without thereby being identical. Synonymy is a feature of terms and analyticity of statements, and 'synonymy' and 'analyticity' are also correlative concepts. It is impossible to understand one concept in a correlative pair without having some understanding of the other. One has not understood the concept 'left' if one has not understood that what falls under the concept 'left' always falls under the concept 'to the left of something'·, i.e. one has at least a partial understanding of 'right'. This means that it is impossible to give a definition of the one term in terms of the other, without going in a circle. However, this is consistent with the fact that one can get increased understanding of the one concept by talking about the other. I do not think that analyticity can be defined by means of stipulative definitions and/or synonymy, but I claim that analyticity may be illuminated by means of the latter. My thesis that stipulative definitions and synonymies can give rise to statements which are analytically true, and consequently that there are analytic statements, gets some general support from my ontological system. One distinction put forward there as an absolute one is the distinction between intentional acts as such and their modality of satisfaction. Certain features of intentional acts are totally independent of the satisfaction modality (e.g. the distinction between presentational and representational intentionality) whereas others only concern this modality (e.g. whether a wish is fulfilled or not). Corresponding to this, there are, I claim, two different concepts of 'true' (and 'false', but I shall for the most part restrict the discussion to the concept of truth); (1) 'true' as dependent upon the satisfaction modality; (2) 'true' as independent of the satisfaction modality. Truths of the first kind are synthetic truths, those of the second kind are analytic. I am not saying that analytic statements lack a satisfaction modality, in that case we would have fictional intentionality (cf. section 12.2), not representational intentionality. A statement like 'All bachelors are unmarried' is satisfied in the sense that there really are bachelors in the world; the concept 'bachelor' has reference. The point is that even if the statement was not satisfied, it would still be true. It is not true because there is a certain state of affairs which satisfies it, i.e. because its whole intentional correlate exists. It is true because of something in the statement itself. 322

EPISTEMOLOGICAL POSITIONS

In order to make the specific distinction between a priori and a posteriori truths which I am going to propose, something must be said about the parts of a statement. It is worth noting that Kant took as his starting point a subject-predicate logic whereas much of the criticism by logical positivists of the synthetic a priori referred to statements as such. The analysis of (atomic) statements was not regarded as being of importance for the discussion. I shall not try to reintroduce subject-predicate logic as something that might replace the logics of relations and propositions, but I shall claim that in most statements one can distinguish one or more referring expressions. This modest claim is enough for my purposes. By 'referring expression' I mean of course ordinary nominal expressions but also the term A in sentences of the form 'All A are B\ In the sentence 'The elephant is grey' the referring expression is 'The elephant'; in the sentence 'a is taller than b' there are two referring expressions 'a' and 'b'. Definite descriptions are normally referring expressions, and it is tempting to follow Strawson's analyses and say not only that statements like 'The king of France is bald' lack truth-value if the definite description has no reference," but quite generally that statements whose referring expressions fail to refer lack truth-value. For a statement like 'All electrons have the same charge, 1.6-ΚΓ19 coulombs' this means that if it should be discovered that there are no electrons at all, we should not say that the statement is false, but that it lacks truth-value; since its referring expression ('electron') fails to refer. In one way such an option seems odd, in another way it does not. Since by hypothesis it is discovered that all our theories about electrons have been about fictitious entities, it is natural to apply the label 'false'. But, on the other hand, after this discovery we can begin to speak about electrons in the same way as we speak about persons in fairy-tales. We can very well look upon superseded theories not so much as false theories but as theories that lack truth-value (i.e. regard them as belonging to fictional discourse). Much speaks in favour of the Strawsonian option. I think it comes closest to everyday language even among scientists, but in order to bring out the contrast between my view and traditional views on the synthetic a priori it is more convenient to adopt the other option for a while. Therefore, let us now regard as false those statements whose referring expressions fail to refer. 323

ONTOLOGICAL INVESTIGATIONS

Table 16.1

Applicable Non-applicable

Analytic statement

Synthetic a priori statement

Synthetic a posteriori statement

NT NT

NT F

Τ F

T h e next step in the analysis is a sweeping generalization: since so many 'indubitable' scientific theories have been dethroned in the course of history, it must always be an a posteriori matter whether any referring expression fails to refer or not. This gives us the hint we need. We should try to distingish between truths which are analytic, synthetic a priori, and synthetic a posteriori in a way which is neutral with respect to the question about the reference of the referring expressions. T h e term 'neutral' is perhaps misleading, because what I mean is that the three 'kinds of truth' should be defined by taking into account both the case where there is failure of reference and the case where there is no such failure. If this is done, one discovers the similarities and dissimilarities necessary and sufficient for a tripartition. M y views can be put forward in the form of a table (see Table 16.1), and I shall give it at once and explain it afterwards. I shall say that a statement whose referring expressions succeed in referring is an 'applicable statement'; consequently, a statement lacking such reference is 'non-applicable'. I n the table Τ is an abbreviation for 'true', NT for 'necessarily true', and F for 'false'. An analytic truth is characterized by the fact that it is necessarily true whether it is applicable to the world or not. It is necessarily true that all bachelors are unmarried whether or not there any any bachelors. Analytic truths are true even in fairytales. It should be noted that all I want to say is that there exist analytic truths, something which is quite consistent with certain kinds of mistakes. Analyticity is not the same as infallibility. If, for instance, 'bachelor' and 'unmarried m a n ' are not synonymous (cf. page 321) then the statement in question is not an example of an analytic truth. T h e column for synthetic a posteriori truths is quite simple; Τ means of course not only 'true', but 'true yet not necessarily true'. T h e statement 'All electrons have the same charge, 1.610 - 1 9 324

EPISTEMOLOGICAL POSITIONS

coulombs' is false if there are no electrons, but true if there are electrons and if they all have the charge described. If there is some electron which has another charge, then the statement is of course also false. When we come to the synthetic a priori, we come to my most controversial claim. This column combines necessary truth and falsity! Examples falling under this heading abound in the present theory of categories. All the existential dependencies in chapter 9 belong here: 'everything coloured is extended', 'every tone has both an intensity and a pitch', and so on. If the statements are applicable they are necessarily true, if they are not applicable they are false. Should someone feel it odd to say, in the latter case, that the statements in question are false, I must remind him of what I said about the 'Strawsonian option'. Our intuitions work along two different directions, and I have decided to follow the antiStrawsonian line for the time being. Another reason why it may look odd to say that the statement 'Every tone has both an intensity and a pitch' can be false, is the fact that the referring expression 'tone' so obviously has a reference that the opposite seems inconceivable. Let us therefore take a look at Newtonian physics and at what happened when it was replaced by Einstein's theories. I regard some of Newton's fundamental laws as entailing laws which are synthetic a priori but nonetheless false. It is logically impossible that there exists a Newtonian force if there do not exist mass and acceleration (this is the second law of motion which the numerical relationship omitted). And it is logically impossible that one body afTects another with a Newtonian force if the other body does not aifect the first with a similar force (i.e. the third law of motion).' 2 The referring expressions in Newtonian mechanics refer to bodies which have the property mass, and mass in a definite sense. Mass inheres in bodies and is independent of the velocity of bodies. The clash between Newtonian mechanics and relativity theory is best described by saying that relativity theory denies that Newtonian mechanics refers to the world. It is true that Einstein, too, speaks about the mass of a body, but his notion of mass is quite different, as the debates about the incommensurability of theories have shown. 13 In principle the same fate can befall all our statements, and in that sense even statements which are synthetic a priori can turn out to be false. That is to say, if 'true' and 'false' are used in the way I 325

ONTOLOGICAL INVESTIGATIONS

have hitherto used them. There is another way of making my point which is in accordance with Strawson's line of thought. We can construct hypothetical definitions of the three kinds of truths. We then get the following: (1) True synthetic a posteriori statement = def. a statement which, if it is applicable, is true but not necessarily true. (2) True synthetic a priori statement = def. a statement which, if it is applicable, is necessarily true. (3) True analytic statement = def. a statement which independently of applicability, is necessarily true. In this way I think it is possible and necessary to retain a threefold distinction between kinds of truth. If this is granted, I may speculate a little about which presumed truths really are synthetic a priori. I think there are relations of existential dependence between universals in the core of every so-called paradigm (Kuhn). Every great scientific revolution consists in the announcement of some synthetic a priori statements. Paradigms do not lack truth-value as epistemological relativists claim, they actually contain statements which are true synthetic a priori. Old paradigms are never shown to be plain false, they are shown to be non-applicable. Aristotelian physics (which I have referred to now and then in the book) is non-applicable rather than false, and so on. If conventionalism with regard to space (i.e. the view that any geometry can be made applicable) is, as I think it is, untenable, 14 then my view on synthetic a priori truths seems to get support from geometry. It fits very well with the way mathematicians speak about Euclidean geometry. The discovery of non-Euclidean geometries, and the subsequent discovery that the universe is nonEuclidean, have not led mathematicians to say that Euclidean geometry is false. They distinguish instead between pure and applied geometry, and say that Euclidean geometry is not applicable to the universe as a whole, although it is a true or consistent mathematical theory. It is false that our universe has an Euclidean geometry, but Euclidean geometry is not false. It consists of a number of synthetic a priori truths; remember the hypothetical moment in definition (2). In 'Two dogmas of empiricism' Quine mentions the proposal that quantum mechanics requires a revision of logic, and he 326

EPISTEMOLOGICAL POSITIONS

interprets this as showing the possibility that the law of the excluded middle in formal logic may be false. Such an interpretation of course supports his view that there is only one kind of truth, a posteriori truths. Another interpretation, and one conforming to the ideas I have put forward, is to say that logicians should learn from geometers. One has to distinguish also between pure logic and applied logic. There can never be any question of declaring the law of the excluded middle to be simply false, but one may consider whether one should stop applying it in the domains of quantum mechanics. I do not know whether any logician has made use of the terms 'pure logic' and 'applied logic', but the growth within formal logic of non-standard and deviant logic seems implicitly to point towards a similar distinction. I tend towards the view that the whole of mathematics and logic, as well as the existential dependence relations in scientific paradigms, are synthetic a priori truths in the sense I have defined. A comparison with Kant may serve to bring out some interesting features of my position. According to Kant, the analyticity of analytical statements is grounded in logic, whereas synthetic a priori truths are connected with three different areas; arithmetic, geometry, and fundamental scientific postulates. Today there is no reason to differentiate between arithmetic and geometry in the way Kant did. The development within mathematics has here had repercussions within philosophy. But, if my speculations are correct, then the development within logic also has negative repercussions as far as Kant's philosophy is concerned. Logic can no longer be grouped together with the stipulative definitions and synonymies that give rise to analytically true

Analytical truths

Logic Arithmetic

Scientific postulates Figure 16.1 327

ONTOLOGICAL INVESTIGATIONS Synomies and stipulative definitions

Analytical truths

Logic Mathematics

Synthetic a priori truths

Scientific postulates Figure 16.2

statements. The Kantian figure (16.1) should be replaced by Figure 16.2. Whatever may be said about the position put forward, I think the main burden of proof rests on those philosophers who deny the prima facie distinction between (a) totally empty statements, (b) abstract statements put forward by logicians, mathematicians, and theoretical scientists, and (c) statements that are obviously empirically testable. 16.3 EMPIRICALLY CRITICIZABLE METAPHYSICS What is the epistemological status of this theory of categories? Is it (1) transcendentalist? Is it (2) rationalist? Is it (3) empiricist? Or, might it even turn out to be (4) part and parcel of the latest relativistic fashion — sociologism?15 First, it is not transcendalist. As we saw in the last section there are only hypothetical synthetic a priori truths. What goes for the synthetic a priori in general also goes for this ontological system: If it is applicable, it is necessarily true. It is at one and the same time both a priori and a posteriori. In order to decide whether a metaphysical system is applicable or not, an interplay between theory and observation is necessary. And it is a kind of interplay which is overlooked by and often in conflict with rationalism as well as with empiricism. Let me argue by analogy. Before scientists could see that Euclidean geometry is not applicable to the universe as a whole, it was necessary to have 328

EPISTEMOLOGICAL POSITIONS

a competing theory, a non-Euclidean geometry. Observations in themselves were not sufficient. But neither was it possible by rationalist means to decide from the geometries themselves which one is true of the world. Now if we consider different metaphysical systems, it seems that some systems, e.g. those naturalistic ones that deny the intentionality category, are flatly contradicted by experience — our experience contains intentional phenomena. The same fate, however, does not befall dualist systems. T o see that Cartesian systems are inapplicable, systems like the one presented here are needed. Conversely, this means that in principle this ontology can be challenged by other metaphysical systems. Empiricists will welcome the claim that a metaphysical system may be incompatible with certain observations. Both metaphysical systems and scientific theories, they may say, are not independent, merely underdetermined, by observations. However, it is nowadays a commonplace among philosophers of science that all observations are theory-laden. 16 This position fits well with the views put forward above (chapter 13) about intentionality. Acts of presentational intentionality are dependent upon both the subject and the state of affairs making up the act's intentional correlate. And so there is no kind of observation which the (empiricist) philosopher can pick out beforehand as being a true apprehension of objective reality. Since the observation is dependent upon the subject, it may be influenced by the subject's theories. It is important to note that, in spite of the fact that all observations are theory-laden, there may occur observations which seem to be, and perhaps are, in real conflict with the theory which is implicit in observations. This is the lesson of Kuhn's concept of anomaly, although it has not been paid the attention it deserves. 17 If the existence of anomalies is recognized, the theory-ladenness of observations does not without further ado lead to relativism. For traditional empiricists anomalies are reality's way of saying 'no' to a theory. But, as just pointed out, things are not that simple. Furthermore, as Kuhn and others have remarked, all fundamental theories seem to encounter anomalies all the time. Such theories are born with anomalies, live with anomalies, and die with anomalies. And what at one time seems to be an anomaly may later turn out not to be an anomaly at all. Relativists have taken such facts to support their position. In my opinion, relativists have not looked carefully enough at the available options. One may 329

ONTOLOGICAL INVESTIGATIONS

easily think that the only possible alternatives in the face of anomalies are the following: (la) It is always the case that: anomalies are reality's protest. (lb) It is always the case that: anomalies are internal defects of a theory. If we take the first phrase in (la) and (lb) to be a modal operator, we can write (la) as Ap and (lb) as Ai, and we get the disjunctive option (Ap υ Ai). The first disjunct is the empiricist answer and the second the idealist answer. The funny thing now is that empirical data tell against empiricism. The history of science invalidates (la), and idealism with the relativism that accompanies it remains. However, let us compare the following options: (1) (ApvAi) (2) A(p υ i) T h e important thing to note is that these disjunctions are not equivalent. In the second option the modal operator operates on the whole disjunction. Translated into ordinary language it says: 'It is always the case that: either an anomaly is the protest of reality or it is an internal defect of the theory.' There is here a meta-option between (1) and (2), and if one opts as I do, for (2), one is not faced with the dilemma between empiricism and idealism. Nothing in the history of science blocks the second alternative. The low-level option then only holds for each individual case. We get: (2a) This anomaly is a protest of reality. (2b) This anomaly is an internal defect of the theory. When the theory encountering the anomaly is a scientific theory, the answer to the last option should not be given by philosophers at all, but by the scientists who are concerned with the case at hand. From the purely philosophical point of view the anomaly is just something which may be a sign that the theory is not truthlike. One cannot be sure, but neither is it possible to be sure of the opposite. The meta-option (2) above might be called a realistfallibilist conception of anomalies. According to this conception, we can never know for sure when one theory is more truthlike than another. Yet anomalies affords us one area where we can discuss 330

EPISTEMOLOGICAL POSITIONS

which theory is closest to the truth. We have got an epistemological position, albeit a non-foundational one. A fundamental scientific theory, a paradigm, may be criticized in two different ways: either by pointing to a competing synthetic a priori truth (paradigm) or by pointing to the existence of anomalies. T h e same, I maintain, is true of ontological systems. An ontology may be contested by another competing ontology, but it may also encounter a kind of anomaly and be empirically criticizable. The main line of thought behind the latter claim is the following. I have now and then stressed the fact that universals can be ordered in a determinate-determinable relationship. Let us now make the false but convenient assumptions that there are only three such degrees of abstraction in the world, and that the lowest determinates are more directly observable than the determinables. Furthermore, let us say that these determinates are described by (theory-laden!) 'observation terms', that the first order determinables are described by means of 'scientific terms' and the upper or second order determinables by 'category terms'. T h e associations I want to create are obvious: a neat tripartition of the activity of seeking knowledge into observation, science, and ontology. It must not be taken too seriously, but let me pursue the example. We assume that some of the observations constitute anomalies for some scientific theory. Such a fact need not be problematic for the ontology at hand. If the anomalies cannot be explained away, then there may be another superseding theory which also conforms to the ontology. But, and now we come to the interesting point, the alternative scientific theory which can take care of various anomalies may presuppose categories which are in conflict with the initial ontological framework. From what has been said about the determinable-determinate relationship, it follows that in our model scientific theories are 'category-laden' in the same way as observations are 'theory-laden'. T h e alternative theory then constitutes a 'category anomaly'. And, as in the case of 'theory anomalies', one of two things may happen. Either, in spite of first appearances, it turns out that the alternative theory does fit the ontology; or the alternative theory comes to be regarded as the more truthlike one, and so implies that the ontology is not applicable. In this way even metaphysical systems are empirically criticizable. 'Criticizable' should be taken literally. Anomalies need 331

ONTOLOGICAL INVESTIGATIONS

not be sufficient reasons for rejecting a theory or an ontology. 18 It might well be argued that the ontology presented in this book is confronted by just such a category anomaly. My defence of naive realism is, as I pointed out (see section 13.6), in conflict with traditional theories in perceptual psychology. According to the latter, we can never in a perception be in direct contact with material reality. All perceptions are regarded as subjectively created interpretations of sensory stimulus. But according to the category of presentational intentionality defended here, such a direct contact is really possible. I shall not, however, dwell on this problem here, since I think I can appeal to the new kind of perceptual theories developed by Gibson 19 - theories which, as far as I can see, are compatible with my metaphysical system. In my view, neither scientific theories nor metaphysical systems can be absolute in the sense of having an absolutely indubitable foundation, but this view does not imply total epistemological relativism. It is consistent both with my concept of truthlikeness and with the kind of naive realism defended in chapter 13. Naive realism implies that observations can be regarded as the 'meeting place' for theory and reality. Perceptions can normally contain both an epistemologically subjective moment and an epistemologically objective moment. The problem for science and philosophy is to deepen the objective moment. It is possible to reject transcendentalism, rationalism, and empiricism, and yet remain an epistemological realist and believe that it is possible to get closer and closer to the truth. A belief in some kind of sociology of knowledge has, together with the recognition that observations are theory-laden, led many a scientist and philosopher to sociologism, i.e. to the view that it is meaningless to speak of truth and try to discover the truth, 'since', as the saying goes, 'anyhow, everything is socially determined'. As is clear from my defence of 'truthlikeness', I am opposed to sociologism. Even if theories are socially determined they do have a certain degree of truthlikeness. Social determination cannot eliminate the fact that by means of the category of real intentionality we are directed towards reality. My theory of categories is not opposed to the sociology of knowledge, only to sociologism. Moreover, the category of nested intentionality can be used to show why the sociology of knowledge does not imply sociologism. Let us look once again at the kind of 332

EPISTEMOLOGICAL POSITIONS

nested intentionality which I associated with Rousseau's concept of the general will (see above, page 298). Where there is such a general will every person wants what the other wants, and all the wants get nested into a potential infinite regress. In the same way as a person may want what others want (instead of wanting something just by himself), so a person may want to perceive nature and society in the same way as others perceive them (instead of trying to perceive them as they are, or according to some wish of his own). We get the following 'general will': A

В

(1) A wants to perceive what В perceives

В wants to perceive what A perceives

(2) A wants (В to want to perceive what A perceives)

В wants (A to want to perceive what В perceives) В wants (A to want (В to want to perceive what A perceives))

(3) A wants (В to want (Л to want to perceive what В perceives))

In this structure we can substitute any cognitive verb whatsoever for 'perceive'. The structure is of course an ideal type, but nonetheless I think it gives a more correct description of our cognitive situation than the traditional individualistic pictures of epistemology which only reckon with non-nested intentionality. Traditionally, the epistemological subject is a single person thinking about and/or observing the world. But the real epistemological subject is certainly a macroagent in the sense given in section 15.8. Whether such a subject is best described by marxist, feminist, Mannheimian, or some other type of sociology of knowledge, is a question which falls outside the very abstract concerns of this book. With regard to sociologism, however, what comes out clearly in the structure above, is the fact that nested intentionality in cognitive situations does not in any way cancel cognitive intentionality itself (i.e. the perception in the case considered). The persons making up a 'cognitive general will' have nested cognitions. But even this sort of cognition still has a directedness and a modality of satisfaction. In other words, there is truthlikeness. 333

NOTES

I ONTOLOGY 1 These are the categories which Aristotle lists in his Categories. The translation is from D. Ross (1968) Aristotle, London, p. 21. In the Loeb Classical Library edition of the work of Aristotle, H.P. Cooke translates Ross's 'passivity' as 'affection (or passion)', 'date' as 'time', and 'posture' as 'posture or position'; see H.P. Cooke (ed.) (1938) Aristotle vol. 1, London, 1973, pp. 79, 81. 2 See D. Hume (1964) A Treatise of Human Nature, London, pp. 11, 19. 3 See I. Kant (1929) Critique of Pure Reason, trans. N. Kemp Smith, London, 1968, p. 113. 4 T h e terms I use are a mixture of the terms J.N. Findlay and C. Taylor use; see J.N. Findlay (1970) Hegel: A Re-examination, London, and C. Taylor (1977) Hegel, Cambridge, respectively. Where I (like Taylor) have 'reality' and 'concept', Findlay has 'actuality' and 'notion'; where I (like Findlay) have 'quality' and 'essence as reflection into self' Taylor has 'dasein' and 'from reflection to ground'. 5 Gilbert Ryle (1963) The Concept of Mind, Harmondsworth, pp. 17ff. 6 L. Wittgenstein (1961) Tractatus Logico-Philosophicus, London, contains what is, to my mind, a very thoroughly worked out atomistic ontology. I have, however, criticized it for committing a category-mistake. Instead of one category Dinge or Gegenstände, there should be two in accordance with E. Husserl's distinction between selbständige und unselbständige Gegenstände; cf. chapter 9 of this book. My criticism is published in my paper 'Ivar Segelberg, Edmund Husserl och Tractatus', in Spräk, kunskap, medvetande. Festskrift tillägnad Ivar Segelberg pä hans 70-arsdag, Gothenburg, 1984, pp. 109-20. 7 With regard to the atomism-holism opposition I would like to mention both a structuralist marxist philosopher, Louis Althusser, and a phenomenological marxist philosopher Karel Kosik. T h e former distinguishes, in For Marx, London, 1969, and Reading Capital, London, 1970, between (1) mechanical causality (associated with Descartes), and (2) expressive causality (Leibniz, Hegel), and (3) structural causality (Marx). The latter distinguishes in The Dialectics of the 334

NOTES

8 9

10

11 12

13

Concrete, Dordrecht, 1976, between (1) the atomist-rationalist conception (Descartes, Wittgenstein), (2) the organicist and organicist-dynamic conception (Schelling, Spann), and (3) the dialectical conception (Heraclitus, Hegel, Marx). In both cases the first conception represents atomism, the second conception holism, and the third the true middle. I disagree with both specifications of the third option, but I share the underlying intention. It ought perhaps also be said that Althusser and Kosik differ between themselves, as is apparent from their classification of Hegel. Among those philosophers who have tried to overcome the rift between pro-subject and anti-subject approaches to philosophy, three books, to my mind, are outstanding. First, in chronological order, comes Jean-Paul Sartre (1976) Critique of Dialectical Reason, London, with its accompanying introduction, The Problem of Method, London, 1963. Second, we have Peter L. Berger's and Thomas Luckmann (1967) The Social Construction of Reality, New York; and third, Roy Bhaskar (1979) The Possibility of Naturalism, London. The latter two are criticized later on in this book. Sartre's views are too vague to function as a foil for my own views, but they are determinate enough to have functioned as one of the sources of inspiration. J . Piaget (1977) Understanding Causality, New York, p. 145. The modern revival of the concept of'tendency' is due to G.E.M. Anscombe and P.T. Geach (1961) Three Philosophers, Oxford, R. Harre (1970) The Principles of Scientific Thinking, London, R. Harre and E.H. Madden (1975) Causal Pouiers, Oxford, and R. Bhaskar (1975) A Realist Theory of Science, Leeds; for criticism see Q. Gibson (1983) 'Tendencies', Philosophy of Science 50:296-308. The concept of intentionality has of course, as it always has had, a prominent place within the phenomenological tradition. However, in later years it has also been discussed in areas like speech-act philosophy and cognitive psychology; see J . Searle (1983) Intentionality, Cambridge, and H.L. Dreyifus (ed.) (1982) Husserl, Intentionality, and Cognitive Science, Cambridge, Mass., respectively. I have written a (forthcoming) paper which applies ideas from this book to quantum mechanics. It is called 'Planck's constant and necessarily time-extended phenomena', and uses the idea of inclusiveness in time put forward in section 4.5. David M. Armstrong (1978) Nominalism and Realism: Universals and Scientific Realism vol. 1, Cambridge, p. 113. Eminent exceptions to the nominalist domination in this century are Edmund Husserl and Samuel Alexander. Neither, however, has given such detailed arguments for universals in re as has Armstrong. Perhaps the most well-known representative of this view is G.F. Stout. Armstrong even calls the particulars in question 'Stoutian particulars'. Another representative I would like to mention is Ivar Segelberg: see the foreword for references. The false identification of 'Stoutian particulars' with the immanent realism of Aristotle and Husserl can be found in W. Künne's (1982) 'Criteria of abstractness: the ontologies of 335

NOTES

14 15

16 17 18

Husserl, Frege and Strawson against the background of classical metaphysics', in B. Smith (ed.) (1982) Parts and Moments: Studies in Logic and Formal Ontology, München, pp. 421-4. Armstrong, Nominalism and Realism, chapter 8. I am fully convinced, though, that this distinction is important in the philosophy of language, too. See my paper (1986) 'Levels of intension and theories of reference', Theoria 52:1-15. David M. Armstrong (1978) A Theory of Universals: Universals and Scientific Realism vol. 2, Cambridge, p. 106. ibid., p. 106. Brand Blanshard (1973) Reason and Analysis, La Salle, 111., p. 406. 2 IRREDUCTIVE MATERIALISM

1 With regard to the different kinds of psyche enumerated, I am following A.N. Leontjev (1975) Probleme der Entwicklung des Psychischen, Berlin. 2 G.W.F. Hegel (1873) Logic, trans. W. Wallace, Oxford, 1975, p. 280. 3 K . Marx (1974) Grundrisse, Harmondsworth, p. 91. 4 For a thorough and detailed analysis of Engels's philosophy, its genesis and development, see S.-E. Liedman (1977) Motsatsemas spei. Friedrich Engels'filosofi och 1800-talets vetenskap vols 1 and 2, Lund; it contains an English summary; German translation (1985) Das Spiel der Gegensätze, Frankfurt am Main. Liedman stresses Engels's interest in the natural sciences and their importance for his philosophy, and stresses the term 'irreductive materialism', a term which Engels himself never used. Views similar to those of Liedman (and mine) can be found also in D.D. Weiss (1984) 'Toward the vindication of Friedrich Engels', in R.S. Cohen and M.W. Wartofsky (eds) Methodology, Metaphysics and the History of Science, Dordrecht. Weiss calls Engels's materialism 'genetic materialism'. 5 S. Alexander (1920) Space, Time and Deity vols 1 and 2, London; N. Hartmann (1975) New Ways of Ontology, Westport; R. Ingarden (1964-6) Der Streit um der Existenz der Welt vols 1, 2, 3, Tübingen; M. Polanyi (1964) Personal Knowledge, New York, and (1967) The Tacit Dimension, London; M. Bunge (1973) Method, Model and Matter, Dordrecht, part 3, and (1977, 1979) Treatise on Basic Philosophy vols 3 and 4, Dordrecht. 6 Alexander's metaphysical system obviously does not fit into the tradition of analytic or linguistic philosophy in Britain. Hartmann is (as is said in The Encyclopedia of Philosophy vol. 3, p. 422) 'not at all typical of recent German philosophers'. G.H. von Wright has in an overview of modern philosophy written (my translation): 'Isolated thinkers of impressing stature but, in my opinion, without really being engaged in the philosophical process of their time, are Samuel Alexander (1859-1938) and Nicolai Hartmann (1882-1950).'; see (1965) Logik,filosofioch spräk, Stockholm, p. 24. Ingarden, who belongs to the phenomenological tradition, sought to combat the development 336

NOTES of phenomenology towards transcendental idealism and/or existentialism. Polanyi and Bunge are most discussed among Anglo-American philosophers of science, but none of them fits into the divide between positivists, Popperians and the so-called Weltanschauungen analyses (Kuhn, Feyerabend, et al.). None of the philosophers mentioned put forward exactly the kind of irreductive materialism which I am going to propose. However, there seems to me to be more affinity between us than just our subscription to level ontologies. All seem to have about the same way of looking at ontology's relation to both epistemology and science; see section 16.3. 7 M. Bunge (1967) Scientific Research I. The Search for System, Berlin, p. 293. 8 Bunge (1973) Method, Model and Matter, pp. 162-4. 3 STATES OF AFFAIRS AND QUALITIES 1 The table is taken from K. Mulligan, P. Simons, and B. Smith (1984) 'Truth-makers', Philosophy and Phenomenologtcal Research 44:287-321, who in turn derive it from I. Angelelli (1967) Studies on Gottlob Frege and Traditional Philosophy, Dordrecht, p. 12. This book contains an excellent survey of Aristode's views on these matters and its subsequent fate in the history of philosophy; see chapter 1 on 'Ontology'. 2 G.E. Moore writes in 1952, for instance, in a comment on a paper of his written in 1910: Ί should now make, and have for many years made, a sharp distinction between what I have called the "patch", on the one hand, and the colour, size and shape, of which it is, on the other; and should call, and have called, only the patch, not its colour, size or shape, a "sense-datum".'; see G.E. Moore (1953) Some Main Problems of Philosophy, London, p. 30. And A.J. Ayer says frankly: 'But now it turns out that for the reasons I have given, statements about physical objects cannot be translated into statements about sensedata.'; see A.J. Ayer (1965) Philosophical Essays, London, p. 142. 3 See J.E. Hochberg (1964) Perception, Englewood Cliffs, N.J., ch. 4. 4 For Armstrong's critical comments on such a stand and for further references to the discussion, see his (1979) A Theory of Universal!: Universals and Scientific Realism vol. 2, Cambridge, section 18.1. 5 The terms 'part' and 'aspect' which are here used as primitive terms will be defined in chapter 9. What I now call 'aspect' will then be called 'moment'. As an undefined term, however, 'aspect' has better connotations. 6 What has just been said indicates similarities between my threefold division of universals and instances of universale and that of Peter Strawson. In his book (1964) Individuals, London, he distinguishes between three kinds of universals which have different kinds of relations to particulars. His characterizing universals corresponds to my properties and his feature universals (e.g. snow, water, coal, and gold) to my substances; the third kind of universal, sortal universals, seems to come very close to my universal state of affairs, which as a species has things and whose instances may have independent existence, i.e. they 337

NOTES may be particulars. Here is a quotation from Strawson (pp. 202-3): Snow, water, coal and gold, for example, are general kinds of stufF, not properties or characteristics of particulars; though being made of snow or being made of gold are characteristics of particulars. Nor are the universal terms introduced into these propositions sortal universals. No one of them of itself provides a principle for distinguishing, enumerating and reidentifying particulars of a sort. But each can be very easily modified so as to yield several such principles: we can distinguish count and reidentify veins or grains, lumps or dumps of coal, and flakes, falls, drifts or expanses of snow. Such phrases as 'lump of coal' or 'fall of snow' introduce sortal universals; but 'coal' and 'snow' simpliciter do not. 7 A good presentation of the technical philosophical concepts o f ' s t a t e of affairs', their history and their problems is to be found in K. Mulligan (1985) ' "Wie die Sachen sich zueinander Verhalten" inside and outside the Tractatus', Teoria 5:145-74. Mulligan does not, however, discuss how the concept has been used in the last decades. In particular he does not discuss the philosopher whose notion of 'state of affairs' I come closest to, i.e. Armstrong (ibid.). T h e differences between Armstrong's views and mine are to my mind very subtle and cannot be adequately treated in a footnote. But that there are differences can be seen from the fact that Armstrong does not regard property-instances as particulars. At bottom, I think, the difference is that Armstrong does not make explicit use of the distinction between dependent and independent particulars which I regard as fundamental. 8 I a m thinking primarily of Saul Kripke and Hilary Putnam, see, for example, the anthology (1977) Naming, Necessity, and Natural Kinds, Ithaca and London, edited and introduced by Stephen P. Schwartz. 9 C.J. O'Connor (1972) 'Substance and attribute', in The Ewyclopedia of Philosophy vol. 8, New York and London, p. 37.

4 EXCLUSIVE AND INCLUSIVE UNIVERSALS 1 See, for example, W.V.O. Quine (1964) Word and Object, Cambridge, Mass., section 19. A related problem has also been discussed in the philosophy of the social sciences; see here, for example, M . Brodbeck (1958) 'Methodological individualisms: definition and reduction', Philosophy of Science 25:1-22. She distinguishes between individual properties and group properties. To be twelve is for Brodbeck a group property, because it can be true of a group but not of its parts. A group property can presumably not be defined in terms of corresponding individual properties, and it was as a 'problem of translation' that the question first arose in the debate about methodological individualism within analytic philosophy. 2 David M. Armstrong (1979) A Theory of Universals: Universals and Scientific Realism vol. 2, Cambridge, p. 68. 338

NOTES 3 Armstrong, who does not allow determinables in the books referred to, ends up by saying that there may in fact exist no homoeomerous (i.e. exclusive) properties; ibid., p. 71. This is one of the reasons why I have not adopted Armstrong's terminology. The main reason, however, is that that which in my account comes out as the opposite of the exclusive/homoeomerous, the inclusive, is for Armstrong merely a special kind of the anomoeomerous. I think my 'inclusive property' corresponds to or even has the same extension as his 'non-relationally structural property'; ibid., pp. 68-71. In this connection it should be pointed out that Husserl, too, comes close to similar distinctions. He uses an operation of'piecing', but he none the less does not make exactly the distinction I have made; see (1968) Logische Untersuchungen vol. 2, part 1, Tübingen, or Findlay's translation (1970) Logical Investigations vol. 2, New York, sections 19 and 25 of the third investigation, 'On the theory of wholes and parts'. 4 My use of the term 'addition' is consciously equivocal. It refers both to ordinary mathematical scalar addition, i.e. addition of pure real numbers, and to a mode of combination of things, i.e. putting things spatially next to each other. This equivocation is chosen because I want to draw attention to the fact that when things are so put together some of the instantiated universale (the inclusive ones) also combine while some of the universale (the exclusive ones) do not combine. And this is in my opinion the fundamental fact, and the fact which makes it possible to really add properties not just pure numbers. In chapter 12 I shall have more to say about quantities. Here I merely want to say that if one believes in determinate properties as universals in re, then specific numerical magnitudes like 4.97 kg must refer to such determinates. This means among other things that quantitative properties cannot be opposed to the category of property. Some properties are quantifiable, some properties are not. But what is important at the moment, is the fact that whether a numerical magnitude is additive or not depends on what kind of universal it refers to. 5 See C. Hempel (1952) Fundamentals of Concept Formation in Empirical Science, Chicago and London, p. 75. 6 I. Kant (1929) Critique of Pure Reason, trans. N. Kemp Smith, London, 1968, the sections on 'Axioms of intuition' (extensive magnitudes) and 'Anticipations of perception' (intensive magnitudes). 7 This thought experiment might appear a little far-fetched. In my opinion such an appearance must be due to the fact that, today, it is really very hard to regard colours in the way I have chosen to do, i.e. as a property which inheres in things in the same way as shapes and volumes do (cf. the beginning of this chapter). Scientific theories can affect even our everyday perceptions, and theories of light have so affected our perceptions that colour is partly perceived as lightdependent; thus my thought experiment partly breaks down. However, if colour is regarded as wholly inhering in things, then I maintain that my line of thought is correct. 8 Cf. note 3, above. 339

NOTES

5 A C T I O N S AND F U N C T I O N S 1 William Dray (1957) Laws and Explanations in History, Oxford, especially p. 124. Compare also section 15.3 in this book, 'Actions and subjects'. 2 Charles Taylor (1964) The Explanation of Behaviour, London, pp. 36-7, and A.I. Melden (1961) Free Action, London, ch. 9, especially p. 101. Very relevant in the present context is section 26 of G.E.M. Anscombe's (1957) Intention, Oxford. 3 Pointed out, for example, by A. Goldman (1970) A Theory of Human Action, Englewood Cliffs, N.J. 4 Gilbert Ryle (1963) The Concept of Mind, Harmondsworth, pp. 130ff. 5 T h e terms 'logic of the situation' and 'situational logic' derive from Karl Popper, although their intent have a much longer history; especially within the neo-Kantian tradition. The terms could very well be compared with Max Weber's concept of Sinnzusammenhang. See Karl Popper (1961) The Poverty of Historicism, London, ch. 31, especially p. 149, and (1966) The Open Society and Its Enemies vol. 2, London, ch. 14, especially p. 97; and Max Weber (1962) Basic Concepts in Sociology, Secaucus, N.J. par. 1. I have myself earlier participated in the discussion on methodological individualism; see I.Johansson (1975) A Critique of Karl Popper's Methodology, Gothenburg, ch. 10, section 4. 6 For an attempted explication of the phrase 'could have done otherwise', see R. Chisholm (1976) 'The agent as cause', in M. Brand and D. Walton (eds) Action Theory, Dordrecht, pp. 199-211. 7 In what follows I presuppose the view that actions and functions are some kind of universal. T o perform an act or to fulfil a function is to instantiate a universal. With regard to actions, this view is defended by Alvin Goldman, op. cit., ch. 1. If, as I shall show, such a view is in harmony with the rest of the Kategorienlehre presented here, this coherence must be taken as a further indication of the correctness of the presupposition. 8 Arthur C. Danto (1968) 'Basic actions', in A. White (ed.) Philosophy of Action, pp. 43-58; the quotation is from p. 54. The paper was originally published in American Philosophical Quarterly 2 (1965). 9 John R. Searle (1979) 'The intentionality of intention and action', Inquiry 22:273; compare also John R. Searle (1983) Intentionality, Cambridge, pp. 99-100. 10 There is more to be said in criticism of Danto's conceptioin. My point is that it is too wide. A. Goldman has pointed out that it is too small in that it rules out some actions as basic which we want to class as basic; op. cit., p. 24 — and its note 2 for further references to similar criticisms. 11 With reference to the earlier quoted example of his, Searle writes in Intertionality (p. 98): Princip, for example, fired the gun by means of pulling the trigger and he shot the Archduke by means of firing the gun. Some but not all of these relations are causal. Pulling the trigger causes the gun to fire; but killing the Archduke doesn't cause a blow to be struck against 340

NOTES Austria, in the circumstances it just is striking a blow against Austria. The members of the list, together with the causal (or other sorts oi) relations between them constitute the conditions of satisfaction of a single complex intention in action on the part of Princip.

12 13 14 15 16

17 18 19 20

21

22

A. Goldman (op. cit., ch. 2) has tried to classify different sort of relations which can obtain between actions A and A where A is done by means of A; but Searle makes no reference to Goldman. The latter distinguishes between four kinds of'level generation' (if A is done by means of A, A is level generated by A): causal, conventional, simple, and augmentation generation. All the cases I am going to discuss are cases of'causal generation', and so I shall not discuss Goldman's opinions. ibid., p. 100. Alasdair Maclntyre (1958) The Unconscious, London, p. 52. Cf. the concept of pattern in chapter 6. See Polanyi's work on 'tacit knowing', in, for example, (1964) Personal Knowledge, New York, ch. 4 on 'Skills'. Goldman (cf. note 10, above), who has not captured the idea of inclusiveness, makes some gross mistakes which lead him to the conclusion that if A is level generated by A then A and A must be performed during the same interval of time (op. cit., p. 21) - in flat contradiction with my view. Goldman rightly notes that A ends later than A (p. 21), but then slurs over this fact by taking notice - again rightly - of another fact, namely that one cannot say that someone first does A and then does A (p. 22). The latter fact, however, does not imply that A and A must be performed during exactly the same time. It is enough that A is included in A in such a way that they start at the same time but A ends later than A; see Figure 5.4. Gilbert Ryle (1963) The Concept of Mind, Harmondsworth, ch. 5, pp. 143-7. Zeno Vendler (1967) Linguistics in Philosophy, Ithaca, N.Y., ch. 4 on 'Verbs and times'. ibid., pp. 101-2. Here it should be noted that I analyse things as ordinary things appear to us in everyday vision, i.e. as continuously coloured. According to modern physics, however, if we make the parts small enough we come down to particles like electrons, and these do not have a colour at all. A similar ambiguity holds for 'runnings'. The ordinary conception is that if we run for a certain time, then we are continuously running during this time. Closer analysis might however give rise to the opinion that running should be regarded as a series of recurrent deeds; see the argument three paragraphs ahead. If the term 'aspect' has not enough connotations for the reader, I have to refer to chapter 9, below, and the concept of'moment' which is there defined. 'Aspect' is here used as synonymous with 'moment'. For further problems connected with this view of mine, see section 6.3, below.

341

NOTES 23 Cf. Vendler, op. cit., p. 112. 24 I want to stress that the contrast being made is not a contrast between actions and functions on the one hand and events on the other. I have consciously avoided the term 'event' although it plays a very prominent role in discussions on action within analytic philosophy; see, for example, Myles Brand (1976) 'Particulars, events, and actions', in M. Brand and D. Walton (eds) Action Theory, Dordrecht, pp. 133-57. My reason for this is that 'event' has come to refer to events both as described in natural science and as described by common sense, and in the latter case events might behave exactly as actions and functions. 25 Anthony Giddens (1979) Central Problems in Social Theory, Berkeley and Los Angeles, p. 55. 26 There is a neat argument to the effect that an action is necessarily confined to a person's body. Suppose that in our example G. Princip dropped dead at exactly the moment he had fired the gun. In a sense he of course none the less would have shot the Archduke, but there could be no real action after his death. What happened in that case was u p to nature; cf. J . Hornsby (1980) Actions, London, p. 9. T h e argument is simple and persuasive, but I am not wholly convinced that the conclusion is true. What, for instance, is the difference between using a stick and only using one's arm to move something? T h e stick behaves like a part of the body. Of course the stick can get broken, i.e. the action is dependent on nature, but the arm can get broken just as well. However, I have here no intention of specifying the spatial limits of real actions. For the purposes of this book it is enough that there are limits. Even if the current discussion about actions and bodily movements should in the end yield the conclusion that actions are no more than (patterns of) bodily movements connected with mental acts a n d states, the relations of dependence and constituency described in this chapter will still of course obtain. 27 This distinction is similar to A. Goldman's between causal and conventional generation; cf. note 11, above. 28 In this context one should remember that we often lack concepts for the ultimate determinates. O n page 14 and 48 I pointed this out for colours and shapes, but it is true for a lot of actions as well. 6 P A T T E R N S , CHANGES, A N D P U R E G E S T A L T E N 1 For a discussion of this topic, see also A. Stroll (1979) 'Two conceptions of surfaces', Midwest Studies in Philosophy 4:279-91. 2 I confine the discussion to properties although it can be extended to substances. 7 S E L F - S U S T A I N I N G G E S T A L T E N A N D G E S T A L T E N CAUSA

SUI

1 Another way of phrasing the fact mentioned is to say that the lowerdimensional parts of a change are not themselves changes. In the same way as a three-dimensional whole contains parts which are not 342

NOTES

2 3

4 5

6 7

8 9

10

themselves three-dimensional (e.g. the ending surface), a change contains parts which are not changes. In particular, this means that the starting point and the ending point of changes are not themselves changes. Since I regard time as continuous I have to face the problem of how to handle the phrase 'lying next to each other'; on a continuous line a point has no point lying next to it because between any two points there are always an infinite number of other points. Obviously, we want to say that two temporal Gestalten (e.g. a change and the ensuing state) lie next to each other in time. They shall not overlap, but yet there shall be no temporal gap between them. T h e distinction between equi- and lower-dimensional parts affords us the solution. T h e ending of one temporal Gestalt can coincide with the starting of another temporal Gestalt without these Gestalten having any equidimmsional part in common; what they do have in common is one single lower-dimensional part. I confine the discussion to properties, but 'property' could well be exchanged for 'quality' in the definitions. Whether 'motion' ( = 'change of place') and 'be at a certain place' really refer to properties and not to relations between a thing and space is here unimportant. See, for example, A. Koyre (1968) Metaphysics and Measurement, London. For a short presentation of the scalar concept of velocity as it was developed at the universities in Oxford and Paris in the fourteenth century, see A.C. Crombie (1969) Augustine to Galileo vol. 2, Harmondsworth, pp. 102-7. Velocity was then regarded as an intensity (intensio) or a 'latitude of form' where the 'form' or extension (extensio) was the distance traversed. See S. Toulmin (1963) Foresight and Understanding, New York and Evenston, pp 46-53. I said above that when one 'goes down' one dimension, this corresponds to the division of a magnitude. When one 'goes up' one dimension, this operation corresponds mathematically to multiplication. This point also applies to the symbolism of the integration operation. In many introductions (especially older ones) to the infinitesimal calculus, integration is presented as a kind of addition; an addition of infinitely many and infinitely small units. In my view, if integration is to be explained by means of simpler operations, it should be explained as a synthesis of multiplication and addition. Mario Bunge (1959) Causality, Cambridge, Mass., pp. 108-11. See, for instance, G . H . von Wright (1971) Explanation and Understanding, Ithaca, N.Y., pp. 93f. and 116f. for the argument and for further references. I think most of the criticism of scholasticism in general, and Aristotelian substantial forms in particular, ever since the Renaissance has been based on a false opposition between quality and quantity or between substance and function; cf. E. Cassirer (1923) Substance and Function, Chicago and London. A lot of logical empiricists and affiliated 343

NOTES

thinkers have made a distinction between 'what-questions' and 'howquestions', where only the latter were regarded as scientifically legitimate; see, for example, Karl Popper (1961) The Poverty of Historicism, London, p. 29. The view I am criticizing has been given its most distinct formulation by the Finnish logical empiricist Eino Kaila, who distinguishes between an Aristotelian and a Galilean conception of knowledge. The former is concerned with 'predicative invariances', i.e. invariances within things (cf. the concept of velocity), and the latter is concerned with 'relational invariances', i.e. invariances between phenomena. Only the latter are really scientific. Kaila's view is set out most fully in (1939) Inhimillinen Tieto, mitä se ou ja mitä se ei ole (Human Knowledge. What It Is and What It Is Not), Helsinki, which, however, is only available in Finnish and Swedish. But it is also to be found in E. Kaila (1979) Reality and Experience: Four Philosophical Essays, Dordrecht, pp. 101-9. My views on 'relational invariances' are expounded in section 12.1, below. 11 The 'necessary connection' here spoken of is explained in detail in section 12.2 below. In the terminology used there, a state of mind is generically dependent upon some actions; these actions however are genetically independent but individually dependent upon that state of mind. 12 If there are such invariances they must obviously be mediated by variables taking into account different kinds of bodies and different kinds of situations. A sick man and a healthy man expresses happiness in different ways. 8 EXTERNAL, INTERNAL, AND GROUNDED RELATIONS 1 A.C. Ewing (1934) Idealism: A Critical Survey, London, ch. 4. 2 Bertrand Russell (1910) 'Some explanations in reply to Mr Bradley', MindN.S. 19:374. 3 See chapter 10 pp. 146-7, below. 4 Ewing, ibid., pp. 135-6. 5 See, for example, E.V. Ilyenkov (1977) Dialectical Logic, Moscow. The classic criticism of dialectical logic within analytical philosophy is Popper's chapter on 'What is dialectic?', in (1969) Conjectures and Refutations, London, pp. 312-35. In the East German dictionary (1971) Philosophisches Wörterbuch, Berlin, the explanation of the term 'dialectical logic' {Logik, dialektische) ends with the statement that there is still no satisfying exposition of marxist dialectical logic. 6 This and similar facts are stressed by K. Mulligan (1985) ' "Wie die Sachen sich zueinander Verhalten" inside and outside the Tractatus', Teoria 5:145-74; see especially point (ii) on p. 164. 7 David M. Armstrong (1978) Nominalism and Realism: Universals and Scientific Realism vol. 1, Cambridge, p. 110. 8 Peter Strawson (1964) Individuals, London, p. 167. 9 In anticipation, I might sketch my view by saying that laws in physics ordinarily talk explicitly about determinates but implicitly about determinables. 344

NOTES 10 The concept of'grounded relation' can be found in Anton Marty (1908) Untersuchungen zur Grundlegung der allgemeinen Grammatik und Sprachphilosophie, Halle, sections 70, 97, and 98 of 'Zweites Stück', and (1916) Raum und Zeit, Halle, sections 7, 12, 16, 17, and 27. Marty talks of begründete Relationen·, sometimes also of bedingte orfimdierteRelationen. But he never brings out the contrast between such relations on the one hand, and external and internal relations on the other hand. 11 Ferdinand de Saussure (1974) Course in General Linguistics, Glasgow, part 2, ch. 5.

9 EXISTENTIAL DEPENDENCE 1 David M. Armstrong (1978) A Theory of Universals: Unioersals and Scientific Realism vol. 2, Cambridge, p. 85. 2 ibid., p. 86. 3 The confused state in which the distinction between external and internal relations is in can be highlighted by noting how a present-day philosopher, Jon Elster, makes out the relation 'taller than' to be an example of an external relation! Elster defines external relations in a way which comes close to my definition of grounded relation. His definition (which is later on supplemented by some qualifications) in (1978) Logic and Society, Chichester, p. 23, reads as follows: Rab is an external relation if there exists predicates (not necessarily unique) F\ . . . F„ and G\ . . . Gm such that 'Rab' can be inferred from a truth function of'Fie', 'F?a',. . .'F„a', and 'G|A', 'G^b' . . . 'Gmb\ 4 ibid., p. 3 (my emphasis). 5 David M. Armstrong (1978) Nominalism and Realism: Universals and Scientific Realism vol. 1, Cambridge, pp. 109-10. 6 In the two books by Armstrong referred to, there is one reference to Husserl and the Logical Investigations but it does not concern the issue at hand. Of course, Husserl's thoughts have a prehistory in earlier philosophy. This is presented by B. Smith and K. Mulligan (1982) 'Pieces of a theory', in B. Smith (ed.) Parts and Moments: Studies in Logic and Formal Ontology, München, pp. 15-109. Smith and Mulligan have also noted the close affinity between some of Husserl's concepts and the concept of internal relation; ibid., p. 100 (n. 78). 7 In what follows, references to Husserl (1900-1) Logische Untersuchungen, Halle, will be to the 5 th edition Tübingen, 1968, vols 1 and 2 (parts 1 and 2), henceforth LU; and to J.N. Findlay's English translation (1970) Logical Investigations vols 1 and 2, New York, henceforth LI. 8 See, for example LU vol. 2, part 1, p. 242; LI vol. 2, p. 448. 9 LU vol. 2, part 1; LI vol. 2, part 3, sections 2-5 and 17. With regard to the translation of selbständig and unselbständig, I am now following K. Mulligan, P. Simons, and B. Smith; see their papers in B. Smith (ed.) (1982) Parts and Moments: Studies in Logic and Formal Ontology, München, especially p. 263.1 have previously used the translations 'self-sufficient' 345

NOTES

10

11 12 13 14 15 16 17 18

19 20

21 22 23 24 25

26 27 28 29

a n d 'non-selfsufficient', respectively in (1975) A Critique of Karl Popper's Methodology, Gothenburg, pp. 168-70. LU vol. 2, part 1, p. 261; LI vol. 2, p. 463. T h e quotation is not quite correct. I have exchanged an Μ for В in order to stress the similarity to the definitions of 'internal relation' given in the previous chapter. T h e title of chapter 2 of the 'Third investigation' is 'Thoughts towards a theory of the pure forms of wholes and parts'. P. Simons (1982) 'The formalisation ofHusserl's theory of wholes and parts', in B. Smith (ed.) Parts and Moments, München, section 3. See note 9, above. LU vol. 2, part 1, p. 252; LI vol. 2, p. 455. ibid. Some of the concepts Husserl lists as formal, I regard as material, b u t I shall not go into this problem here. See notes 9 above and 18 below. LU vol. 2, part 1, pp. 262-3; LI vol. 2, p. 464. K . Mulligan (1980) Representation and Ontology in Austro-German Philosophy, Manchester, pp. 166-83. B. Smith (1981) 'Logic, form and matter', Proceedings of the Aristotelian Society, Supplementary Volume 55:48. B. Smith and K. Mulligan (1983) 'Framework for formal ontology', Topoi 2:73-84. For arguments to the effect that there can be no negative universale, see Armstrong, A Theory of Universals, pp. 23-9. I am here thinking of books like B. Oilman (1976) Alienation: Marx's Conception of Man in Capitalist Society, Cambridge, and J . Israel (1979) The Language of Dialectics and the Dialectics of Language, Copenhagen. Both books defend ontological idealism. This fact, however, should not obscure the equally important fact that both books contain some good points; especially for those who have not thought about internal relations at all. For specific criticism directed at Israel's book, see B. Smith (1983) 'Reflections on dependence', in Forskningsrädsnämnden (ed.) An Inventory of Present Thinking about Parts and Wholes vol. 2, Stockholm, pp. 29-42. ibid., especially pp. 122-8. LU vol. 2, part 1, p. 267 (i.e. section 17); Z./vol. 2, p. 468. T h e distinction 'relative' and 'absolute' is introduced in section 13. For more details, see K. Mulligan, P. Simons, and B. Smith (1984) 'Truth-makers', Philosophy and Phenomenological Research 44:287-321. LU vol. 2, part 1, pp. 266-7; LI vol. 2, pp. 467-9. This is also noted by B. Smith and K. Mulligan, 'Pieces of a theory', op. cit., p. 49; and see Husserl LU vol. 2, part 1, pp. 277-82; LI vol. 2, p p . 476-80 to which they refer. ibid., p. 125. T h e idea of existential exclusion, as well as its importance, I owe to K. Mulligan, Cf. also Husserl LU vol. 2, part 1, p. 251; LI vol. 2, p. 455. I rely on the presentation of Hegel's distinction given by J . N . Findlay (1970) Hegel: A Re-examination, London, chapter 8:i, especially p. 225. For a presentation and an analysis of this fallacy, as well as arguments showing that its importance has been overlooked, see J . Elster (1978) 346

NOTES Logic and Society, Chichester, especially pp. 97-106. 30 The readers who find this meagre can consult Armstrong, A Theoiy of Universals, pp. 19-23, for arguments against disjunctive universals. 31 I think P. Simons conflates these two kinds of individual dependence, see above, note 12. 10 CONTAINER SPACE AND RELATIONAL SPACE 1 It seems to me, that it was not until the 1970s, through works by philosophers like H.M. Lacey, G. Nerlich, and L. Sklar that the distinction between relational conceptions, on the one hand, and relative and relativistic conceptions, on the other hand, became fully clear. I am heavily indebted to the three philosophers mentioned. The only thing I have found wanting is a sharp distinction between relative and relativistic spaces. See H.M. Lacey (1970) 'The scientific intelligibility of absolute space: a study of Newtonian argument', The British Journal for the Philosophy of Science 21:317^-2; G. Nerlich (1976) The Shape of Space, Cambridge; L. Sklar (1974) Space, Time, and Spacetime, Berkeley. The terminological mess to which these discussions as well as this book have contributed can be somewhat reduced if the following is borne in mind. What I call 'container space' Lacey calls 'absolute space', Nerlich merely 'space', and Sklar 'substantival space'. My 'relational space' is often called 'relational' or 'relationist'; Nerlich also talks of'quality spaces'. My distinction between non-relative (absolute) and relative spaces, which, by the way, I think conforms to Newton's use of'relative space', appears in Lacey, Nerlich, and Sklar only when they talk about absolute and relative motion. What Lacey calls 'relative space' is identical with my concept of 'relativistic space'. 2 I am even leaving general relativity out of account, except for some remarks in section 12.5, below. The possible extrapolations I am thinking of also include the new multi-dimensional conceptions of space in which space is ascribed up to ten dimensions; with time, eleven. For very brief presentations see B.S. De Witt (1985) 'Quantum gravity', Scientific American 249:104—15, and P. Davies, 'The eleven dimensions of reality', New Scientist, 9 February 1984, pp. 31-3.1 am not able to investigate these spaces myself. The question I would like to have answered is whether the multi-dimensional spaces proposed are relational spaces or container spaces or a mixture of both. In this chapter I stress the categorial differences between container space and relational space, but their mathematical representations can very well be woven together and in this sense create a mixture. 3 In order to see clearly in what sense container space has parts, one has to remember both how I use the concept of'part', and how it is related to 'actuality' and 'potentiality'. There are two kinds of parts, dependent and independent. The parts of space are independent, which means that they can exist without the whole space of which they for the moment are a part. This means that it is logically possible that space might shrink so that it encompasses no more than the part in question. 347

NOTES It must also be remembered that independent parts often are no more than potential parts (cf. p. 136). Parts of space are always potential. 4 Containment, necessary singularity, and independence have not always been thought to go together by philosophers who have reflected on space. According to Kant, space is a container and necessarily singular, but it is not independent. Space has, in his own terms, a 'transcendentaliideality'. This means that I partly agree with Kant, and partly disagree. Some of the agreements will crop up later in the chapter. The disagreements I shall not dwell upon at all, but I want to refer to Anton Marty (1916) Raum and Zeit, Halle. Marty extensively criticizes Kant while defending the container conception of space. On some points, however, I disagree with Marty too. In particular, I do not subscribe to his views that container space has to be unalterable and be without causal efficacy; for the latter point see section 12.5, below. 5 R. Harre (1972) The Philosophies of Science: An Introductory Survey, London, p. 111. 6 Cf. Nerlich, op. cit., chapter 1. 7 M. Bunge (1977) Treatise on Basic Philosophy vol. 3, Dordrecht, pp. 276fT., especially p. 284. 8 Bertrand Russell (1976) Human Knowledge: Its Scope and Limits, London, part 4, ch. 6, especially p. 297. 9 Sklar considers the idea that certain motions, e.g. absolute acceleration, are (one-placed) properties, but I do not really understand the argument; Sklar, op. cit., pp. 229fT. 10 L. Wittgenstein (1961) Tractatus Logico-Philosophicus, London, pp. 141, 143. 11 Nerlich, op. cit., ch. 2. 12 If there is anything in today's physics which could come close to being such a basic resting system it would, I guess, be the cosmic microwave background radiation brought into focus by the Big Bang theorists. This radiation has in every point of space the same intensity in all directions of space. See J . Silk (1980) The Big Bang, San Francisco, pp. 51-2, 75-9. 11 TENDENCY 1 By 'Newtonian mechanics' I shall here mean this kind of mechanics as it is presented in modern textbooks in physics. Newton himself had a lot of trouble with how to look at 'activity principles' in relation to his own theories; see E. McMullin (1978) Newton on Matter and Activity, Notre Dame and London. The main ideas in this section are also presented in my paper (1984) 'Är Newtons mekanik ännu inte filosofiskt forstidd?' (English: 'Is Newton's mechanics not yet philosophically understood?'), in S. Welin (ed.) Att firstä världen, Lund. 2 Bhaskar, of course, also has predecessors (cf. note 9 to chapter 1, above). One who seems to be forgotten, but undeservedly so, is A.P. Ushenko (1946) Power and Events: An Essay on Dynamics in Philosophy, Princeton.

348

NOTES 3 R. Bhaskar (1975) Α Realist Theory of Science, Leeds, p. 99. 4 See, for example, most of the papers in R. Tuomela (ed.) (1978) Dispontions, Dordrecht. 5 I am here thinking of R. Harre (1970) 'Powers', The British Journal for the Philosophy of Science 21:81-101; M. Fisk (1973) Nature and Necessity, Bloomington and London, ch. 10 (both are reprinted in R. Tuomela, op. cit.; and R. Bhaskar, op. cit.; pp. 229-38). Of course, Aristotle also grappled with the problem; see (1986) Aristotle vol. 7, Metaphysics, London, pp. 249-55, 429-33, 455-65. 6 For a recent paper in harmony with my views, though not explicitly talking of tendencies, which discusses something I have not touched upon at all: the connection between de re possibilities and underlying states, see K. Mulligan, 'Dispositions: their correlates and basis', in P. Simons (ed.) Essays on Meinong, Munich, forthcoming, 1989. 12 EFFICIENT CAUSALITY 1 A.P. Ushenko (1946) Power and Events, Princeton, pp. 80-115; M. Bunge (1959) Causality, Cambridge, Mass., pp. 57-38; R. Harre (1970) The Principles of Scientific Thinking, London, pp. 104—10; D.M. Armstrong (1983) What is a Law of Nature?, Cambridge, part 1. 2 The distinction between these two types of laws is clearly drawn in Bunge, op. cit., pp. 8-11, 255-62. It should be noted that Armstrong in his latest book comes to the conclusion that causal laws cannot be identified with laws of nature: op. cit., especially p. 95. Earlier, he thought that 'causal connection is reducible to nomic connection', (1978) A Theory of Univerals: Universals and Scientific Realism vol. 2, Cambridge, p. 149. 3 The distinction between 'classificatory', 'comparative', and 'quantitative concepts' which I use is in accordance with C. Hempel (1952) Fundamentals of Concept Formation in Empirical Science, Chicago and London, part 3. For a review of different kinds of scales, see B. Ellis, (1968) Basic Concepts of Measurement, Cambridge. 4 The quotation is from What is a Law of Nature?, p. 172. The 'others' alluded to are F. Dretske and M. Tooley; see Armstrong, op. cit., p. 85. What I am going to say has some obvious similarities with Armstrong's views, as they are put forward in the book just mentioned and in A Theory of Universals. Like Armstrong, I think (non-causal) laws refer to some kind of higher-order universal which relates other universals. However, the differences are equally as striking. Unlike Armstrong I regard grounded relations (= 'internal relation' in Armstrong) as universals, and this difference has far-reaching consequences. 5 I think this is the kernel of Bhaskar's proposed solution to the problem of induction; see (1975) A Realist Theory of Science, Leeds, pp. 215-28. Cf. also Armstrong, What is a Law of Nature?, pp. 103-4. 6 Armstrong has in What is a Law of Nature? a very short chapter on 'functional laws' (chapter 7). In this he suddenly and regretfully admits that there might be determine^/« universals. Such universals he regards 349

NOTES

7 8

9

10

11

12

13 14

15 16

(in contradistinction to me) as second-order universals. According to Armstrong, a functional dependence is a universal which relates determinables, and is of third order. From my point of view, Armstrong's theory lacks grounded relations. Neither determinates nor determinables are in themselves quantities, and Armstrong can only neglect this point because he never tries to analyse numerical magnitudes. I am speaking of this law as it appears in modern textbooks in physics; cf. note 1 to chapter 11, above. For a short description of the exception to this rule, i.e. the 'singularist position', see Armstrong, What is a Law of Nature?, pp. 95-6. Later on I shall make a distinction between 'individual dependence' and 'particular dependence'. T h e latter notion is, I think, almost identical with the idea of'singular causality'. I argue that particular dependence is an impossibility while individual dependence is indispensable to the concept of'efficient causality'. Actually, I think 'individual dependence' saves the grain of truth contained in the intuitions behind the singularist view on causality. I. Lakatos (1970) 'Falsification and the methodology of scientific research programmes', in I. Lakatos and A. Musgrave (eds) Criticism and the Growth of Knowledge, Cambridge, p. 102. T h e remarks made imply a very simple but devastating criticism of Popper's falsificationism; see I.Johansson, (1980) 'Ceterisparibus clauses, closure clauses and falsifiability', Zeitschrift fur allgemeine Wüsenschaftstheorie 10:16-22. T h e following presentation owes a great deal to J . Agassi (1968) The Continuing Revolution: A History of Physics from the Greeks to Einstein, New York, chs 11, 12, 15. T h e argument is to be found in R.J. Boscovich (1966) A Theory of Natural Philosophy, Cambridge, Mass., pp. 24ff.; it was first published in 1758. See A. d'Abro (1951) The Rise of the New Physics vol. 1, New York, p. 276. As I use 'mixture' there is a distinction to be drawn between a mixture a n d a blend; a blend is a mere juxtaposition. Corpuscularian ontologies allow only blends. Aristotle and the Aristotelian tradition has, however, grappled with the concept of'mixture'; see, for example, R. Sharvy (1983) 'Aristotle on mixtures', The Journal of Philosophy 80:439-57. I have chosen the term 'mixture' because there are affinities - though not identity - with Aristotle's term. G. Nerlich (1979) 'What can geometry explain?', The British Journal for the Philosophy of Science 30:69-83. Cf. M. Bunge (1977) Treatise on Basic Philosophy vol. 3, Dordrecht, p. 276; and L. Sklar (1974) Space, Time, and Spacetime, Berkeley, p. 164. 13 I N T E N T I O N A L I T Y

1 See note 9 to chapter 1, above.

350

NOTES 2 This whole chapter could well be dedicated to the phenomenological tradition, even if I am very critical towards several common positions within phenomenology. In order to avoid too many notes I shall therefore start by giving a general reference to the classic by Herbert Spiegelberg (1969) The Phenomenological Movement: A Historical Introduction 2 vols., The Hague. 3 This kind of pointing can in principle be momentary, which means that the subject category introduced here also allows momentary subjects. 4 I have chosen the word 'satisfaction' for several reasons. One of them is the fact that J . Searle's excellent book (1983) Intentionality, Cambridge, employs a similar notion ('conditions of satisfaction'). I think I am in agreement with most of the claims made by Searle, but there are some flaws in his book which I shall comment upon; see notes 6, 9, and 15, below. 5 The concepts of 'mere presentations' and 'non-positing acts' are taken from Husserl's Logische Untersuchingen, especially the fifth investigation; LU vol. 2, part 1, pp. 343-508; LI, vol. 2, pp. 503-69. Two central papers about fictional discourse are J . Searle (1979) 'The logical status of fictional discourse', in his Expression and Meaning, Cambridge, pp. 58-75, and R. Rorty (1982) 'Is there a problem about fictional discourse?', in his Consequences of Pragmatism, Brighton, pp. 110-38. 6 Husserl has a similar but not identical distinction between 'intuitive' and 'significant intentions'; LU vol. 2, part 2, pp. 53fT.; LI vol. 2, pp. 710fT. Searle makes a distinction between 'presentations' and 'representations' which is almost identical with mine. The difference is that I regard the distinction as mutually exclusive whereas Searle takes presentations to be a subclass of representations; op. cit., p. 46. Actually, in intent I think it is the same distinction. Searle merely conflates two uses of 'representations'. He sometimes uses the term as a determinable equivalent to my term 'intentionality', and sometimes he uses it as a determinate equivalent to my 'representational intentionality'. 7 Husserl discusses this topic extensively under the concept of 'fulfilment', especially in the sixth investigation; LU vol. 2, part 2, pp. 1-127; LI vol. 2, pp. 667-770. 8 LU vol. 2, part 2, pp. 115-27; LI vol. 2, pp. 760-70. 9 This view was first put forward by A. Marty (1908) Grundlegung zur allgemeinen Grammatik und Sprachphilosophie, Halle, pp. 408-15. It makes it possible to describe the mistake in Searle's notion 'conditions of satisfaction' (see above, note 4). He explicitly says that such conditions are internal to intentional phenomena (op. cit., p. 11), but he also talks of them as states of affairs in the world external to the intentional phenomena (op. cit., p. 13). What is actually internal are representations of conditions of satisfaction. He himself thinks this is a harmless process-product ambiguity (op. cit., p. 13), but his argument for this view is very brief and unconvincing. I think Searle's ambiguity reflects the difficulty of keeping intentionality as such distinct from its modality of satisfaction.

351

NOTES 10 In the vast literature on Husserl I would like to mention the book I mostly rely on: D. Haglund (1977) Perception, Time, and the Unity of Mind: Problems in Edmund Husserl's Philosophy, Gothenburg, mimeographed, Department of Philosophy. What to my mind is missing in most commentaries on Husserl is an appreciation of the role his immanent realism plays, at least in his earlier publications; cf. note 12 to chapter 1, above. 11 I sometimes wonder whether the later Husserl himself did not fall prey to the temptation to locate intentionality in a simpler part of our intentional acts. I am thinking of his concept of 'noema' which in his (1950) Ideen zu einer reinen Phänomenologie und phänomenologischen Philosophie, The Hague, came to replace some of the older concepts and alone represent the directedness typical of intentionality. 12 T o appreciate the discussion which follows, it can be noted that Armstrong, for one, explicitly argues that intentionality can be reduced to powers or tendencies; D.M. Armstrong and N. Malcolm (1984) Consciousness and Causality, Oxford, pp. 149-53. In a second step he wants to reduce tendencies to genuine properties and natural laws. I am opposed to both reductions. 13 A good discussion of such phenomena is J . Elster (1976) Ά note on hysteresis in the social sciences', Synthese 33:371-91. 14 T h e thesis that some intentional acts are relational in the sense that they are existentially dependent upon their intentional correlates, has also been put forward by B. Smith (1958) 'Acta cum fundamentis in re', Dialectica 38:157-78, and by K. Mulligan and B. Smith (1986) Ά relational theory of the act', Topoi 5:115—30. There are two differences between my views and the views in these two papers. First, in contrast with Mulligan and Smith I maintain that the acts in question contain the correlates they are dependent upon. Second, Mulligan and Smith employ a notion of de re necessity which, to my mind, conflates the notions of individual dependence and particular dependence which I worked out in chapter 9 (see above, pp. 182—4). Particular dependence, I argued, is an impossibility; in a (presentational) relational act the relation is that of individual dependence - which, of course, is a kind of de re necessity. 15 When Searle grapples with these problems he introduces a quite new notion of'causal self-referentiality'. He himself calls it 'a crucial addition to the conceptual apparatus of this book' (op. cit., p. 49), and it leads him (together with other things) to a new conception of causality in general (see chapter 4 of his book). In this conception he invokes the old concept ofinternal relation (op. cit., p. 121). He regards cause and effect as logically related. From a very broad perspective this conforms with my attempt to rehabilitate a modified notion ofinternal relation, but in detail there are a lot of differences. I shall not discuss those details here. Of course I claim that my analysis ofinternal relations, efficient causality, and intentionality brings out a lot of things which Searle has not seen, or at least not written about. 16 J . Searle (1982) 'Minds, brains, and programs' in D.C. Dennett and

352

NOTES D.R. Hofstadter (eds) The Mind's I, Harmondsworth, pp. 353-73. 17 One might perhaps also talk of a third tendency within phenomenology, a tendency which wants to place phenomenology within the framework of a realist metaphysics of an Aristotelian type. Roman Ingarden is one proponent of this view (see notes 5 and 6 of chapter 2 above), John Wild is another (see J . Wild (ed.) (1953) Return to Reason: Essays in Realistic Philosophy, Chicago). The latter even started an 'Association for realistic philosophy', op. cit., pp. 357-63. Kevin Mulligan, Barry Smith, and Peter Simons are philosophers who have defended positions of this sort (see above, notes 6, 9, 12, 18, 20, 23, 25, 26, 27, 31 of chapter 9, and note 14 above). My own philosophical outlook is of course very similar to this third tendency. 18 The only philosopher I know of who has stated this is Samuel Alexander, but only in a footnote; (1912) 'The method of metaphysics; and the categories', Mind N.S. 21:1-20, p. 3 n. 2. 14 NATURE: PARTS AND WHOLES W I T H O U T INTENTIONALITY 1 E. Nagel (1961) The Structure of Science, London, pp. 425-6. 2 I. Kant (1951) Critique ofJudgement, New York, part two, 'Critique of the teleological judgement'. 3 It is in my view very misleading to speak of feedback systems as systems with 'circular causality', as is often done in system theory. See, for example, Ludwig von Bertalanffy's classic (1968) General System Theory, New York, p. 162. 4 The terms are taken from R. Brown (1963) Explanation in Social Science, London, p. 123; but the conceptual distinction is of course quite common. 5 Polanyi's views, including his level ontology (cf. chapter 2 above), are forcefully defended in the philosophy of biology by M. Grene (1794) The Understanding of Nature: Essays in the Philosophy of Biology, Dordrecht, 'Boston Studies in the Philosophy of Science' vol. 66. 6 M. Polanyi (1964) Personal Knowledge, New York, p. 360. 7 ibid., p. 329. 15 MAN AND SOCIETY: NESTED INTENTIONALITY 1 The essential part of the argument is taken from D.-H. Ruben (1983) 'Social wholes and parts', Mind 92:219-38. See also D.-H. Ruben (1985) The Metaphysics of the Social World, London, ch. 2. 2 The term 'intentionally connected' may need some words of clarification. When a subject is connected in this way with an intentional correlate, there is of course a relation between the subject and the correlate. This is quite consistent with the view put forward in chapter 13 that intentionality cannot possibly be a relation. Intentional connections only exist when the modality of satisfaction has the value 'satisfied'. That a subject has a representational intentional connection 353

NOTES

3 4 5

6 7 8 9 10

11 12

13 14 15 16 17 18

19 20 21

22

23 24

with с means that с exists or has existed and that there is a grounded relation between с and a representational act of the subject. That a subject has a presentational intentional connection with с means that с exists or has existed and that there is a relation of containment between a presentational act of the subject and c. Noretta Koertge (1975) 'Popper's metaphysical research program for the human sciences', Inquiry 18:437-62. In particular, this is true of Max Weber with his famous concept of 'Ideal types'; see above, note 5 to chapter 5. This idea can probably be more tightly integrated with the rest of my category system than is the case in this chapter. I think the concept of 'contradiction' is closely connected with the concept of'existential exclusion' of chapter 9 (D9.5). Husserl has put forward the claim that 'contradictory contents' are merely special cases of a wider notion 'conflicting contents'; see LU vol. 2, part 2, pp. 102-15, LI vol. 2, pp. 749-59. J . - P . Sartre (1966) Being and Nothingness, New York, p. 302. ibid., p. 303. ibid., pp. 301-2. Cf. M. Furberg (1974) 'Friendship', Filosofiskt herbarium tillägnat Ivar Segelberg, Gothenburg, pp. 25—47. T h e same kind of infinite structures which I have just described can be found in another notation in S. Schiffer (1972) Meaning, Oxford, ch. 2, section 2, 'Mutual knowledge'. Grice first put forward his views in his paper (1957) 'Meaning', Philosophical Review 66:377-88. I am primarily thinking ofJ . Searle (1969) Speech Acts, Cambridge, pp. 42-50, and SchifTer, op. cit., pp. 17-30. Schiffer, op. cit., p. 11. R.D. Laing (1970) Knots, London, p. 1. ibid., p. 48. Cf. R.D. Laing, H. Phillipson, and A.R. Lee (1966) Interpersonal Perception: A Theory and a Method of Research, New York, p. 29. ibid., pp. 28-9. In what I am saying in this section I am heavily indebted to J . Elster (1978) Logic and Society: Contradictions and Possible Worlds, Chichester, ch. 4, and (1979) Ulysses and the Sirens: Studies in Rationality and Irrationality, Cambridge, ch. 4. K a n t uses these terms when stating his 'third antinomy of pure reason'. J o h n Donne, The Canonization. G. Bateson (1971) Steps to an Ecology of Mind, New York. Once more (cf. p. 148) we see how important the notion of existential exclusion (D9.5) is. It can also shed light on contradictions; see also note 27 to chapter 9, above. R.D. Laing, op. cit., p. 4. With regard to marxism, see note 7 to chapter 1 above; with regard to analytic philosophy, I want to say that I regard 'analytic metaphysics' as part of analytic philosophy.

354

NOTES 25 J . Elster, op. cit., especially Introduction and chs 4 and 5. Cf. also note 18 above. 26 ibid., pp. 68, 70. 27 ibid., pp. 97-106; cf. also p. 141 above. 28 K. Marx, Capital vol. 1, Harmondsworth: Pelican, 2nd edn, ch. 4. 29 ibid., pp. 113-18 and 130-2, respectively. 30 Here I mainly have in mind a group of thinkers in West Germany and Denmark who use to be labelled 'capital logicians' (Kapitallogiker). Their interpretations of Marx were quite influential during the 1970s in the two countries. T h e 'key text' is, I would say. H. Reichelt (1970) Zur Logischen Struktur des Kapitalbegriffs bei Karl Marx, Frankfurt am Main. Other writers in this tradition are H.-G. Backhaus, Η. Krahl, and A. Schmidt from Germany, and A. Lundkvist and H.-J. Schanz from Denmark. 31 Very relevant in this context is F. Rossi-Landi's attempt to view both language and material production in a unified way, i.e. to see the similarities and differences between the 'production and circulation of goods' and the 'production and circulation of sentences'; F. RossiLandi (1974) 'Linguistics and economics', Current Trends in Linguistics vol. 12, The Hague and Paris, part 3, pp. 1,787-2,017. 32 It ought to be pointed out (as an anonymous reader has stressed) that the symbolism used bypasses some relevant classical problems concerning the relationship between reason, passion, and the will. From the point of view of this theory of categories, there are at least three different kinds of relationship which may obtain between two intentional acts or between two psychological faculties. First, one act (or faculty) can be existentially dependent upon another; a desire to have something is for instance dependent on a belief that this something is missing. One can only (consciously) desire what one thinks is missing. Second, one act can have another as its intentional correlate; as when you have a (second-order) desire to get rid of another (first-order) desire. Third, one act (or faculty) may control another. The distinction between existential dependency and second-order intentionality can easily be explained if we regard believing (B) and desiring (D) as modal operators which operate on states of affairs (s). It then holds, quite generally: (i) Ds is one-sidedly existentially dependent upon B(-1 s). Ds, however, is not directed at B(-> s). It is directed at s. Let us compare this case with a second-order desire to desire that s would be the case: (ii) D(Ds). D(Ds) is directed at (Ds) but existentially dependent upon Β-, (Ds), since, once again, you can only desire a desire which you believe is missing. Both existential dependency and second-order intentionality differ from the third kind of relationship where one act (or faculty) controls or monitors another. The latter is assumed to be the case when the will is regarded as something which creates compromises between reason

355

NOTES a n d passion. In chapter 2 (p. 25-6) I warned against conflation of control hierarchies and existential hierarchies. Here, I want to repeat the warning and add that neither should there be conflation with orders of intentionality. When, in the following pages, I speak of a want existing for a reason, then the want is assumed to be controlled by this reason. 33 I find the following quotation from J . Elster symptomatic (op. cit., p. 157): J . F . Nash has suggested that the decision to cooperate should itself be seen as a move in a larger non-cooperative game, because the latter is somehow more fundamental. I believe that this view is indeed sound, and follows from the principle of methodological individualism, but I also agree with the critics who point out that it lacks persuasive power so long as it has not been spelled out in detail. 34 I wrote some pages ago (see above, p. 314) that this book is inspired by what I take to be problems within a marxist approach to philosophy. With reference to the theses now stated, I would like to quote a part of the first of Marx's 'Theses on Feuerbach' in K. M a r x and F. Engels (1969) Basic Writings on Politics and Philosophy, London, p. 283: T h e chief defect of all hitherto existing materialism - that of Feuerbach included - is that the thing (Gegenstand), reality, sensuousness, is conceived only in the form of the object (Objekt) or of contemplation (Anschauung), b u t not as h u m a n sensuous activity, practice, not subjectively. Hence it happened that the active side, in contradistinction to materialism, was developed by idealism - but only abstractly, since, of course, idealism does not know real sensuous activity as such. 35 O n this I would like to quote from Marx's third thesis on Feuerbach (op. cit., p. 284): T h e materialist doctrine that men are products of circumstances and upbringing, and that, therefore, changed men are products of other circumstances and changed upbringing, forgets that it is men that change circumstances, and that the educator himself needs educating. 36 R. Taylor has argued for the same point; (1969) ' T h o u g h t and purpose', Inquiry 12:154, 168-9. 37 ibid., p. 285. In view of notes 34 and 35 above, and what is to follow, it is worth quoting something Sidney Hook wrote in the middle thirties: Ί believe that Marx's critical theses on Feuerbach represent in nuce a turning point in the history of philosophy'; (1962) From Hegel to Marx: Studies in the Intellectual Development of Karl Marx, Michigan, p. 273. O n some points I disagree with Hook's interpretation of M a r x , b u t to my mind the book compares very well with later commentaries. All of them lack the category of intentionality. For a recent commentary, see W. Suchting (1979) ' M a r x ' s "Theses on Feuerbach": a new translation

356

NOTES

38 39 40 41 42

43 44

and notes towards a commentary', i n j . Mepham and D.-H. Ruben (eds) Issues in Marxist Philosophy vol. 2, Brighton, pp. 5-34. I shall end this note with another quotation, this time due to a marxist who does not regard the 'Theses' as belonging to the 'mature Marx': 'Those brief sparks, the Theses on Feuerbach, light up every philosopher who comes near them, but as is well known, a spark dazzles rather than illuminates'; L. Althusser (1969) For Marx, London, p. 36. R. Bhaskar (1979) The Possibility of Naturalism: A Philosophical Critique of the Contemporary Human Sciences, Brighton, p. 40. E. Dürkheim (1964) The Rules of Sociological Method, New York, ch. 1, especially pp. 3, 13. This remark is also made by R. Keat and J . Urry (1975) Social Theory as Science, London, p. 189. ibid., p. 42. ibid., p. 46. The sociologist Anthony Giddens has - simultaneously with Bhaskar - worked out a very similar model. Giddens calls it 'the theory of structuration'; see A. Giddens (1979) Central Problems in Social Theory, Berkeley and Los Angeles, and (1984) The Constitution of Society, Cambridge. I have chosen to take Bhaskar as my point of departure for the simple reason that he isolates the philosophical level. Giddens tries to create a new ontology and a new sociological grand theory in one move. He is as interesting as Bhaskar, but he will not be commented upon here. ibid., p. 46. Cf. K. Marx (1973) Grundrisse, Harmondsworth, p. 92: 'Hunger is hunger, but the hunger gratified by cooked meat eaten with a knife and fork is a different hunger from that which bolts down raw meat with the aid of hand, nail and tooth.' 16 E P I S T E M O L O G I C A L P O S I T I O N S

1 N. Hartmann (1975) New Ways of Ontology, Westport, p. 137. 2 Some of the arguments put forward in sections 16.1 and 16.2 below are taken from my paper (1987) 'Beyond objectivism and relativism', Radical Philosophy 47:13-17. Danish version (1986) Philosophia 15:248-61. 3 Lenin (1967) Materialism and Empirio-Criticism, Moscow. K. Popper (1966) 'Addenda', The Open Sotiety and its Enemies, vol. 2, London, 'Truth, rationality and the growth of scientific knowledge', and 'Addenda', in Conjectures and Refutations, London; (1972) Objective Knowledge, Oxford, chs 2, 9. 4 K. Popper, Open Society, p. 373. In order to understand the discussion in the 1970s of Popper's concept of'verisimilitude', one has to recognize that in addition to the concept of and the criteria for verisimilitude there might be a measure of the latter. To construct such a measure is the same as to construct a scale or a metric corresponding to an ordinary property. Even when the construction is successful one has not got criteria for the application of the measure. There has been a 357

NOTES

5 6 7 8

9 10 11

12

13

14 15 16 17 18

lively discussion of the measure of verisimilitude in several philosophical journals; see, for example, The British Journal for the Philosophy of Science 24 (1973) and later, Synthese 30, 38 (1975, 1979). I shall not comment upon their discussion, since I think that even if there is no measure we cannot do without truthlikeness or verisimilitude. Lenin, op. cit., p. 129. It should perhaps be said that Lenin takes over the distinction between absolute and relative truth from Engels. К . Popper, Conjectures, p. 229. T h e inference is taken from R. Rorty's paper (in Danish) (1986) 'Demarkationsproblemet nog engang', Philosophia 15:190-210. A. Collier (1978) 'In defence of epistemology', Radical Philosophy 20:8-21; quotation, p. 16. I share many of the epistemological positions put forward in this paper. Quine's classic paper is (1961) 'Two dogmas of empiricism', From a Logical Point of View, New York and Evanston, pp. 20-46. ibid., p. 27 (my emphasis). P. Strawson (1950) ' O n referring', Mind 59:320-44. As will become apparent, I shall not discuss any details in this or other writings of his. I refer to him merely as a representative of the view that it is not necessary to regard all statements as being either true or false. T h e second law is extensively discussed in chapter 12. With regard to the third law, one can argue in the following way. The force relation is founded upon some property of the bodies which it relates. Since this property inheres in both bodies, then by reason of symmetry if the one body affects the other with a force, then the other has to affect the first body. This argument does not say, as Newton's law does, that the forces are also of equal magnitude. But I d o not need that for the point I make in the book. I hope that a general reference to the kind of philosophy of science which is associated with Thomas K u h n and Paul Feyerabend will suffice in order to indicate what I mean by 'incommensurability'. The classic discussion of the case at hand, Newtonian mass versus Einsteinian mass, is Feyerabend (1968) 'How to be a good empiricist a plea for tolerance in matters epistemological', in P.H. Nidditch (ed.) (1968) The Philosophy of Science, Oxford, pp. 12-39. For an overview of modern philosophy of science, see A.F. Chalmers (1982) What is This Thing called Science?, Milton Keynes. I have written a similar overview myself, but in Swedish. See the first part of I. Johansson and S.-E. Liedman (1987) Positivism och Marxism, Stockholm, 2nd enlarged edn. For arguments to this effect, see G. Nerlich (1976) The Shape of Space, Cambridge. I am here in particular thinking of the so-called 'strong programme'; see D. Bloor (1976) Knowledge and Social Imagery, London. Chalmers, op. cit., especially ch. 3. See I.Johansson (1975) A Critique of Karl Popper's Methodology, Gothenburg, ch. 13, especially p. 127. J . Agassi has made a similar point when discussing Popper's criterion 358

NOTES of falsifiability; (1964) 'The nature of scientific problems and their roots in metaphysics', in M. Bunge (ed.) The Critical Approach to Science and Philosophy, New York, pp. 189-211, especially pp. 191-2. Cf. also my comment in (1975) A Critique of Karl Popper's Methodology, Gothenburg, p. 84. 19 J.J. Gibson (1966) The Senses Considered as Perceptual Systems, New York, and other writings. I have made some philosophical comments on this book in A Critique of Karl Popper's Methodology, ch. 17.

359

BIBLIOGRAPHY

d'Abro, A. (1951) The Rise of the New Physics 2 vols, New York: Dover. Agassi, J . (1964) 'The nature of scientific problems and their roots in metaphysics', M. Bunge (ed.) The Critical Approach to Science and Philosophy, New York: Free Press of Glencoe. (1968) The Continuing Revolution: A History of Physics from the Greeks to Einstein, New York: McGraw-Hill. Alexander, S. (1912) 'The method of metaphysics; and the categories', Mind N.S. 21:1-20. (1920) Space, Time and Deity 2 vols, London: MacMillan. Althusser, L. (1969) For Marx, London: Allen Lane. (1970) Reading Capital, London: New Left Books. Angelelli, I. (1967) Studies on Gottlob Frege and Traditional Philosophy, Dordrecht: Reidel. Anscombe, G.E.M. (1957) Intention, Oxford: Blackwell. Anscombe, G.E.M. and Geach, P.T. (1961) Three Philosophers, Oxford: Blackwell. Aristotle (1973) Aristotle vol. 1, London: Loeb Classical Library. (1980) Aristotle vol. 17, London: Loeb Classical Library. Armstrong, D.M. (1978) Nominalism and Realism: Universals and Scientific Realism vol. 1, Cambridge: Cambridge University Press. (1978) A Theory of Universals: Universals and Scientific Realism vol. 2, Cambridge: Cambridge University Press. (1983) What is a Law of Nature?, Cambridge: Cambridge University Press. Armstrong, D.M. and Malcolm, N. (1984) Consciousness and Causality, Oxford: Ayer, A.J. (1965) Philosophical Essays, London; MacMillan. Bateson, G. (1972) Steps to an Ecology of Mind, New York: Ballantine. Berger, P.L. and Luckmann, T. (1967) The Social Construction of Reality, New York: Doubleday. Bertalanfly, L. von (1968) General System Theory, New York: Braziller. Bhaskar, R. (1975) A Realist Theory of Science, Leeds: Leeds Books. (1979) The Possibility of Naturalism: A Philosophical Critique of the Contemporary Human Sciences, London: Harvester.

360

BIBLIOGRAPHY

Blanshard, В. (1973) Reason and Analysis, La Salle, 111.: Open Court. Bloor, D. (1976) Knowledge and Social Imagery, London: Routledge and Kegan Paul. Boscovich, R.J. (1966) A Theory ofNatural Philosophy, Cambridge, Mass.: MIT Press. Brand, M. (1976) 'Particulars, events, and actions', Action Theory, M. Brand and D. Walton (eds) Dordrecht: Reidel. Brodbeck, Μ. (1958) 'Methodological individualisms: definition and reduction', Philosophy of Science 25, 1-22. Brown, R. (1963) Explanation in Social Science, London: Routledge and Kegan Paul. Bunge, Μ. (1959) Causality, Cambridge, Mass.: Harvard University Press. (1967) Scientific Research I: The Search for System, Berlin: Springer. (1973) Method, Model and Matter, Dordrecht: Reidel. (1977, 1979) Treatise on Basic Philosophy vols 3, 4. Dordrecht: Reidel. Cassirer, E. (1923) Substance and Function, Chicago and London: Open Court, New York; Dover. Chalmers, A.F. (1982) What is this Thing called Science?, Milton Keynes: Open University Press. Chisholm, R. (1976) 'The agent as cause' in M. Brand and D. Walton (eds), Action Theory, Dordrecht: Reidel. Collier, A. (1978) 'In defence of epistemology', Radical Philosophy 20:8-21. Crombie, A.C. (1969) Augustine to Galileo vol. 2, Harmondsworth: Penguin. Danto, A. (1968) 'Basic actions', in A. White (ed.) The Philosophy of Action, London: Oxford University Press. Davies, P. (1984) 'The eleven dimensions of reality', New Scientist 9 February 1984. De Witt, B.S. (1983) 'Qantum gravity', Scientific American 249, no. 6. Dray, W. (1957) Laws and Explanations in History, Oxford: Oxford University Press. Dreyfus, H.L. (ed.) (1982) Husserl, Intentionality, and Cognitive Science, Cambridge, Mass.: MIT Press. Dürkheim, Ε. (1964) The Rules of Sociological Method, New York: Free Press. Ellis, B. (1968) Basic Concepts of Measurement, Cambridge: Cambridge University Press. Elster, J . (1976) Ά note on hysteresis in the social sciences', Synthese 33:371-91. (1978) Logic and Society: Contradictions and Possible Worlds, Chichester: Wiley. (1979) Ulysses and the Sirens: Studies in Rationality and Irrationality, Cambridge: Cambridge University Press. Ewing, A.C. (1934) Idealism: A Critical Survey, London: Methuen. Feyerabend, P. (1968) 'How to be a good empiricist - a plea for tolerance in matters epistemological', in P.H. Nidditch (ed.) Philosophy of Science, Oxford: Oxford University Press. Findlay, J.N. (1970) Hegel: A Re-examination, London: George Allen and Unwin. Fisk, M. (1973) Nature and Necessity: An Essay in Physical Ontology, 361

BIBLIOGRAPHY Bloomington and London: Indiana University Press. Furberg, M. (1974) 'Friendship', Filosofiskt herbarium tillägnat Ivar Segelberg, Gothenburg: University of Gothenburg. Gibson, J.J. (1966) The Senses Considered as Perceptual Systems, Boston: Houghton Mifflin. Gibson, Q . (1983) 'Tendencies', Philosophy of Science 50:296-308. Giddens, A. (1979) Central Problems in Social Theory, Berkeley and Los Angeles: University of California Press. (1984) The Constitution of Society, Cambridge: Polity Press. Goldman, A.I. (1970) A Theory of Human Action, Englewood Cliffs, N.J.: Prentice Hall. Grene, M. (1974) The Understanding of Nature: Essays in the Philosophy of Biology, Dordrecht: Reidel. Grice, H.P. (1957) 'Meaning', Philosophical Review 66:377-88. Haglund, D. (1977) Perception, Time, and the Unity of Mind: Problems in Edmund Husserl's Philosophy mimeographed, Gothenburg: University of Gothenburg. Harre, R. (1970) The Principles of Scientific Thinking, London: Macmillan. (1970) 'Powers', The British Journal for the Philosophy of Science 21:81-101. (1972) The Philosophies of Science: An Introductory Survey, London: Oxford University Press. Harre, R. and Madden, E.H. (1975) Causal Powers, Oxford: Blackwell. H a r t m a n n , Ν. (1975) New Ways of Ontology, Westport: Greenwood Press. Hegel, G.W.F. (1873) Hegel's Logic: Being Part One of the Encyclopedia of the Philosophical Sciences, trans. W. Wallace, London: Oxford University Press, 1975. Hempel, C. (1952) Fundamentals of Concept Formation in Empirical Science, Chicago and London: Chicago University Press. Hochberg, J . E . (1964) Perception, Englewood ClifTs, N.J.: Prentice-Hall. Hook, S. (1962) From Hegel to Marx: Studies in the Intellectual Development of Karl Marx, Michigan: Ann Arbor. Hornsby, J . (1980) Actions, London: Routledge & Kegan Paul. H u m e , D. (1964) A Treatise of Human Nature, London: Everymans Library. Husserl, E. (1900-1) Logische Untersuchungen 3 vols, 5th edn, Max Niemeyer: Tübingen, 1968; Logical Investigations 2 vols, New York: Humanities Press, 1970. (1950) Ideen zu tiner reinen Phänomenologie und phänomenologischen Philosophie erstes Buch, T h e Hague: Martinus NijhofT. Ilyenkow, E.V. (1977) Dialectical Logic, Moscow: Progress Publishers. Ingarden, R. (1964, 1965, 1974) Der Streit um der Existenz der Welt 3 vols, Tübingen: Max Niemeyer. Israel, J . (1979) The Language of Dialectics and the Dialectics of Language, Copenhagen: Munksgaard. Johansson, I. (1975) A Critique of Karl Popper's Methodology, Gothenburg: Scandinavian University Books. (1980) 'Ceteris paribus clauses, closure clauses and falsifiability', Zeitschrift fir allgemeine Wissenschaftstheorie 11:16-22. 362

BIBLIOGRAPHY (1984) 'Ivar Segelberg, Edmund Husserl och Tractatus', in Spräk, kunskap, medvetande: Festskrift tillägnad Ivar Segelberg pi hans 70-arsdag, Gothenburg: University of Gothenburg, pp. 109-20. (1984) 'Är Newtons mekanik ännu inte filosofiskt iorstidd?', in S. Welin (ed.), Att firstä världen, Lund: Doxa, 87-95. (1986) 'Levels of intension and theories of reference', Theoria 52:1-15. (1987) 'Beyond objectivism and relativism', Radical Philosophy 47:13—17; Danish version, Philosophia 15:248-61. 'Planck's constant and necessarily time-extended phenomena', forthcoming. Johansson, I. and Liedman, S.-E. (1987) Positivism och Marxism, Stockholm: Norstedts, 3rd enlarged edn. Kaila, Б. (1979) Reality and Experience: Four Philosophical Essays, Dordrecht: Reidel. Kant, I. (1929) A Critique of Pure Reason, trans. N. Kemp Smith, London: Macmillan, 1968. (1951) Critique of Judgement, New York: Hafner Press. Keat, R. and Urry, J . (1975) Social Theory as Science, London: Routledge & Kegan Paul. Koertge, N. (1975) 'Popper's metaphysical research program for the human sciences', Inquiry 18:437-62. Kosik, K. (1976) The Dialectics of the Concrete, Dordrecht: Reidel. Koyre, A. (1968) Metaphysics and Measurement, London: Chapman and Hall. Künne, W. (1982) 'Criteria of abstractness', in B. Smith (ed.) Parts and Moments: Studies in Logic and Formal Ontology, München: Philosophia Verlag. Lacey, H.M. (1970) 'The scientific intelligibility of absolute space: a study of Newtonian argument', The British Journal for the Philosophy of Science 21: 317-42. Laing, R.D. (1970) Knots, London: Tavistock. Laing, R.D., Phillipson, H., and Lee, A.R. (1966) Interpersonal Perception, New York: Perennial Library. Lakatos, I. (1970) 'Falsification and the methodology of research programmes', in I. Lakatos and A. Musgrave (eds) Criticism and the Growth of Knowledge, Cambridge: Cambridge University Press. Lenin (1967) Materialism and Empirico-Criticism, Moscow: Progress Pubishers. Leontjev, A.N. (1975) Probleme der Entwicklung des Psychischen, Berlin: Dietz. Liedman, S.-E. (1977) Motsatsemas spei: Friedrich Engels'flosofi och 1800-talets vetenskap vols 1, 2, Lund: Cavefors; Das Spiel der Gegensätze, Frankfurt am Main: Campus, 1986. Maclntyre, A.C. (1958) The Unconscious, London: Routledge & Kegan Paul. McMullin, E. (1978) Newton on Matter and Activity, Notre Dame and London: Notre Dame Press. Marty, Α. (1908) Untersuchungen zur Grundlegung der allgemeinen Grammatik und Sprachphilosophie, Halle: Max Niemeyer. (1916) Raum und Zeit, Halle: Max Niemeyer. 363

BIBLIOGRAPHY Marx, К. (1974) Grundrisse, Harmondsworth: Pelican. Capital vol. 1, Harmondsworth: Pelican. Marx, К. and Engels, F. (1969) Basic Writings on Politics and Philosophy, London: Fontana. Melden, A.I. (1961) Free Action, London: Routledge & Kegan Paul. Moore, G.E. (1953) Some Main Problems in Philosophy, London: Macmillan. Mulligan, K . (1980) Representation and Ontology in Austro-German Philosophy, dissertation, Manchester. (1985) ' " Wie die Sachen sich zueinander Verhalten" inside and outside the Tractatus', Teoria 5:145-74. (1989) 'Dispositions: their correlates and basis', in P. Simons (ed.), Essays on Meinong, Munich; Philosophia Verlag, forthcoming. Mulligan, K., Simons, P. and Smith, B. (1984) 'Truth-makers', Philosophy and Phenomenological Research 44:287-321. Mulligan, K. and Smith, B. (1986) Ά relational theory of the act', Topoi 5:115-30. Nagel, Ε. (1961) The Structure of Science, London: Routledge & Kegan Paul. Neriich, G. (1976) The Shape of Space, Cambridge: Cambridge University Press. (1979) 'What can geometry explain?', The British Journal for the Philosophy of Science 30:69-83. O'Connor, D.J. (1972) 'Substance and attribute', The Encyclopedia of Philosophy vol. 8, New York and London: Macmillan. Oilman, B. (1976) Alienation: Marx's Conception of Man in Capitalist Society Cambridge: Cambridge University Press. Philosophisches Wörterbuch (1971) Berlin: Das Europäische Buch. Piaget, J . (1977) Understanding Causality, New York: Norton. Polanyi, M . (1964) Personal Knowledge, New York: Harper. (1967) The Tacit Dimension, London: Routledge & Kegan Paul. Popper, K. (1961) The Poverty of Historicism, London: Routledge & Kegan Paul. (1966) The Open Sodety and its Enemies vol. 2, London: Routledge & Kegan Paul. (1969) Conjectures and Refutations, London: Routledge & Kegan Paul. (1972) Objective Knowledge, Oxford: Oxford University Press. Quine, W.V.O. (1961) From a Logical Point of View, New York: Harper & Row. (1964) Word & Object, Cambridge, Mass.: M I T Press. Reichelt, Η . (1970) Zur logischen Struktur des Kapitalbegriffs bei Karl Marx, Frankfurt am Main: Europäische Verlagsanstalt. Rorty, R. (1982) Consequences of Pragmatism, Brighton: Harvester. (1986) 'Demarkationsproblemet nok engang', Philosophia 15:190-210. Ross, D. (1968) Aristotle, London: Methuen. Rossi-Landi, F. (1974) 'Linguistics and economics', Current Trends in Linguistics vol. 12, T h e Hague and Paris: Mouton. Ruben, D . - H . (1983) 'Social wholes and parts', Mind 92:219-38. (1985) The Metaphysics of the Social World, London: Routledge. Russell, B. (1910) 'Some explanations in reply to M r Bradley', MindN.S. 19.

364

BIBLIOGRAPHY (1976) Human Knowledge: Its Scope and Limits, London: George Allen & Unwin. Ryle, G. (1963) The Concept of Mind, Harmondsworth: Penguin. Sartre, J . P. (1963) The Problem of Method, London: Methuen. (1966) Being and Nothingness, New York: Washington Square Press. (1976) Critique of Dialectical Reason, London: New Left Books. Saussure, F. de (1974) Course in General Linguistics, Glasgow: Collins. Schiffer, S. (1972) Meaning, Oxford: Oxford University Press. Schwarz, S.P. (ed.) (1977) Naming, Necessity, and Natural Kinds, Ithaca and London: Cornell University Press. Searle, J . (1969) Speech Acts, Cambridge: Cambridge University Press. (1979) Expression and Meaning, Cambridge: Cambridge University Press. (1979) 'The intentionality of intention and action', Inquiry 22:253-80. (1982) 'Minds, brains, and programs', in D.C. Dennett and D.R. Hofstadter (eds) The Mind's I, Harmondsworth: Penguin. (1983) Intentionality, Cambridge: Cambridge University Press. Sharvy, R. (1983) 'Aristotle on mixtures', The Journal of Philosophy 80:439-57. Silk, J . (1980) The Big Bang, San Francisco: W.H. Freeman & Company. Simons, P. (1982) 'The formalisation of Husserl's theory of wholes and parts', in B. Smith (ed.) Parts and Moments: Studies in Logic and Formal Ontology, München: Philosophia Verlag. Sklar L. (1974) Space, Time and Spacetime, Berkeley: California University Press. Smith, B. (1981) 'Logic, form and matter', Proceedings of the Aristotelian Society, supplementary volume 55:47-63. (1983) 'Reflections on dependence', in Forskningsrädsnämnden (ed.) An Inventory of Present Thinking about Parts and Wholes vol. 2 Stockholm: Forskningsrädsnämnden, pp. 29-42. (1983) 'Framework for formal ontology', Topoi 2:73-84. (1984) 'Acta cum fundaments in re', Dialectica 38:157-78. Smith, B. and Mulligan, K. (1982) 'Pieces of a theory', in B. Smith (ed.) Parts and Moments: Studies in Logic and Formal Ontology, München: Philosophia Verlag. Spiegelberg, Η. (1969) The Phenomenological Movement: A Historical Introduction 2 vols, The Hague: Martinus Nijhoff. Strawson, P.F. (1950) 'On referring', Mind 59:320-44. (1964) Individuals, London: Methuen University Paperbacks. Stroll, A. (1979) 'Two conceptions of surfaces', in Midwest Studies in Philosophy vol. 4, Minneapolis: University of Minnesota Press. Suchting, W. (1979) 'Marx's Theses on Feuerbach: a new translation and notes towards a commentary', i n j . Mepham and D.-H. Ruben (eds) Issues in Marxist Philosophy vol. 2, Brighton: Harvester. Taylor, C. (1964) The Explanation of Behaviour, London: Roudedge & Kegan Paul. (1977) Hegel, Cambridge: Cambridge University Press. Taylor, R. (1969) 'Thought and purpose', Inquiry 12:149-69.

365

BIBLIOGRAPHY Toulmin, S. (1963) Foresight and Understanding, New York and Evanston: Harper & Row. Tuomela, R. (ed.) (1978) Dispositions, Dordrecht: Reidel. Ushenko, A. P. (1946) Power and Events: An Essay on Dynamics in Philosophy, Princeton: Princeton University Press. Weber, M . (1962) Basic Concepts in Sociology, Secaucus, N.J.: Citadel Press. Weiss, D.D. (1984) 'Toward the vindication of Friedrich Engels', in R.S. Cohen and M.W. Wortofsky (eds) Methodology, Metaphysics and the History of Science, Dordrecht: Reidel, Boston Studies in the Philosophy of Science vol. 84. Vendler, Z. (1967) Linguistics in Philosophy, Ithaca, N.Y.: Cornell University Press. Wild, J . (ed.) (1953) Return to Reason, Essays in Realistic Philosophy, Chicago: Reguery. Wittgenstein, L. (1961) Tractatus Logico-Philosophicus, London: Routledge & Kegan Paul. Wright, G.H. von (1965) Logik, filosofi och spräk, Stockholm: Aldus. (1971) Explanation and Understanding, Ithaca, N.Y.: Cornell University Press.

366

INDEX

absolute authority 299-300, 301-2, 303 absolute motion 154-6 absolute space 145, 154 absolute truth 315, 317, 320 abstract parts 127; see moments abstract universal 140 abstraction levels 14 acceleration: absolute/relative 155, 157-8; Newton's second law 163, 166, 177-85; velocity and 103, 107-8, 166-8, 207 accidents 30, 31, 32, 39-40, 135 accomplishment terms 77-8,80-1, 237-9 accordion effect 68-9,70,71,83,239; inclusiveness in time 74, 75, 76, 78, 81 achievement terms 77, 79-80, 89, 237-9 act matter 207 act quality 207 action 2, 3, 10, 58-84; accordion effect 68, 74; appropriateness and redescription of 58-64 (see also methodological individualism); basic 68-74, 82-4; causa sui 68, 106-9; hierarchy 59-61,69,71,73; inclusiveness in time 74-81; patterns and levels 82-4; pure temporal Gestalt 94, 95, 98; spontaneous 72,97 (set also agents) action (cause): at a distance 187-90, 223; by contact 187-90, 219; by mixture 172,191-5,219 activity terms 77-8, 79,80 actual existence 27, 38

actual parts 90, 91, 136-7 addition lines 74—5 additive property 46, 47, 50, 52; vector addition 165-9 after-images 204 agents (and subjects) 262-7; interaction 270-9 aggregate properties 87-8,146-7,165, 228-9 Alexander, Samuel 28 ambition 275 analysis of the situation 263,265,267 analytic philosophy 42,110,124, 287-8,321 analytic truth 113,320-2,324,326, 327-8 Anaxagoras 42 Anglo-American philosophy 58,96 anomalies 329-32 anomoeomerous properties 42—3,53 anthropomorphization 162,169 anti-dependence 140,143 anti-nominalism 243 appropriateness (and redescription) 58-64,66,68,73 approximate truth 316;«« truthlikeness Aristotle 100,130,185,288; accident concept 30,31,32,135; categories 1,2,3-4,129,196,197; level ontology 26-7,30-1,38-9; physics of 9,193-5,207,318-19, 326; substantial forms 27,30-2,39, 41,108,169; tendency 161-2,169, 196,207 arithmetic 327 Armstrong, David M.: causality 171,

367

INDEX 175-6,177; determinate-determinable relation 15-16,17; homoeomerous properties 42,53; internal relation 117,124-6; universale 11-12,13,17,33,34, 147,175-7 artificial intelligence 215-16 aspects 34,37,78; see moments atomic actions 72; see basic actions atomic theory 39 atomism 110,258-60,297 atomistic ontology 6, 7-8 authority, absolute 299-300,301-2, 303 basic actions/functions 68-74,82-3 basic mass 83 Bateson, Gregory 284-5 behaviourism 7 being 4,129,130 beliefs 197,215 Berger, Peter 310,312,314 Berkeley, George 153 betweenness 149-50,152,153 Bhaskar, Roy 164,310,311-14 biological nature 307,309-10 Blanshard, Brand 17-18 Bosanquet, Bernard 110 Boscovich, R.J. 188,189-90,191,192 Boule-Charles' law 172-6,177,179, 181 Bradley, F.H. 110,137 Brentano, Franz 124,196,275 Bunge, Mario 28-9,101,152,171 'by-means-of relation 71-2 Cartesian systems 329 categories: contemporary problems 5-11; epistemological status 328-33; list 2,20-1,133; theories of 1-5,19-20 category-mistake 4—5,96 category anomaly 331-2 category systems 3-4,9,11,133-5, 137 category terms 331 causasui 95-109,137,161; actions 106-9; temporal Gestalten 97-105 causal relations 96-7,161,180-1,192 causality 3,4,107,108; efficient i « efficient causality 'center of change' 40

change(s): non-change and 93-9,101, 102,103; and pure Gestalten in time 93-4; and temporal Gestalten 96-7,98,101 character traits 106,107 chequered patterns 54 choice mechanism 67; preference system 62-3,265 Clarke, Samuel 155 classificatory concepts 174 climax (ofaction) 77, 78 closed loop 246-7 closed system 246-8 closure clause 266,267 cognitive general will 333 coherence model 72-3 collective dependence 141,143 Collier, A. 320 collision 188-90,191,192 colour 89-90; dependence 142-3; determinates 14-18,44-5,48,86, 125; distance 102-3, exclusion and inclusion 43-5,51-2,53,56; instances 121-2,136; pattern 86-7; space 150-1 commodity markets 62-4,265,266, 268 common sense 9-10 comparative concepts 174 compactness condition 219-20 compression 50-1,52 concept 4,23-4; internal relations 111,113,114-15,116,117, 122-3; nominalism 12 concrete dependence 140-3,183,211 concrete individual 39 concrete thing 137 concrete universal 140 'connection at a distance' 260,305 consciousness and intentionality 73, 214-16 constant conjunction 171,175, 177, 187 constant velocity 99,102 constitution relation 134,135 constructive interpretations 114 container space 11,145-9; as efficient cause 193-5; relational space and 152,153; relative space and 155,156,157; relativistic space and 159-60 contiguity problem 171,172,187-93 continuous flow view 82

368

INDEX contradiction, nested 283-6; problems of dialectics 287-92,294 contradictory intentionality 267-70 contrary opposites 165,168 control hierarchies 25-6; see also chapter 15n32 conventional general terms 14,15, 17-19,321 conventionality 308 conventionalization 18-19 converse relations 131 corpuscles 140,190,191,193,195,220 correspondence theory of truthlikeness 315-20 cosmology 23 Coulomb's law 180 count nouns 42 counterfinality 289 'criterion philosophies' 317 currency 311,313 cybernetics 253 Danto, ArthurC. 68-70,71,72,83 Darwin, Charles 320 date (category) 2 deeds 77,78,80,81,106,237; recurrent 79 bis Democritus 259 density 53,56,86; mass and 45-7, 50-1,90,230 deontic logic 116 dependence set existential dependence dependent existence 31-2,34 dependent parts 127,130,135,136-7; see moments Descartes, Rene 24,188,275,320 description of the situation 263,265-7 descriptions, unfinished 237-9 determinable-determinate distinction 14-21,35,45,119,140, 142,331 determinate properties 44—5,48; definition 53-4 dialectical materialism 28 dialectics, Marxist 287-97 differential calculus 8-9,101 directedness 196,197,204,211,333; and tendency 207-8,210 direction (motion) 98,99,100 disjunctions 143,242-3,330 disposition, potentiality and 169-70; tendency and 169-70 distance relation 102-3,175,179

double-bind phenomenon 284—5 double reference 175 Dray, William 58-9,61 dreams 221 drives 169,209-10,314 Dürkheim, Emile 310,311 'effect sense of function' 253-4 efficient causality 171,283-4; action 187-93; analysis 192; existential dependence and 177-85; non-causal laws 172-7; space as efficient cause 193-5; universal· ty and 185—7 Einstein, Albert 38,145-6,159, 160, 194,325 elasticity concept 189 electromagnetic radiation 18-19 element-set relation 258 Elster, J o n 288-90,191-2 emergent qualities 24,28 emotional states 106-9,262,273,274 empirically criticizable metaphysics 328-33 empty space 153-4 'enantiomorphic things' 157,159,195 energy 27,38-9,217,219; transport thesis 222,224 Engels, Freidrich 28 Enlightenment 7,9 epistemology 29,120,211-12,304; metaphysical systems 328-33; synthetic a priori 320-8; truthlikeness 315-20 epoche 216,217-18 equi-dimensional parts 89,90,91,100, 229-32 equivalence principle (in physics) 38 essence 4,118-19; lawsof 127-8;of man 306-7,309-10 essential properties 39-40,118-19 Euclidean geometry 193-4,320,326, 328-9 evolution 23 Ε wing, A.C. 111,113,114,117,128 exchange value 292-6 exclusion 42-3; existential 138-44,148 exclusive properties 43-50; semi 53-5 exclusive qualities see exclusive properties, exclusive substances exclusive substances 55-6

369

INDEX exclusiveness: in space 232; in time 79,232 existence: actual 27,38; dependent 31-2,34; independent 31-2,34,40; potential 27,38-9 existential dependence (or: dependence) 19,21,25-6,148,205, 206; Armstrong 124r-6; definition of 130,138,139; efficient causality and 177-85; Husserl 126-30, 138-40; see abo collective, concrete, existential exclusion, individual, mutual, one-sided; cf. also independence existential exclusion 138-44, 148 existential independence see independence existentialism 7,263 extension 12,18,47,142,151-2,153, 154; in space/time 45,78,870 extensive magnitudes 50-2 external relations 20,121,122; causality and 178-9; 180; definitions 110-12,125; existential dependence 124-6,131,132;noncausal laws 173,175-6; space 152-3,154,160 fallacy of composition 289-90,311 false statements 322-6passim falsificationism 185 fantasy circles 281 feedback mechanism 249,250-1, 252-3 fictional intentionality 20,199-201, 204-5,212 first order properties 54—5,57,82, 86-7 first substances 30-1,32,39,41 fitness description 251 fitness theorem 252 force: Newton's second law 177-85; partial/resultant 162-3,166,167-8; relation 180-5,191-2 form 26-7 formal logic 113-17,128-9,130 formal ontology 128-9,130,133 forms ofintuition 2-3 foundation relation 128,134,135 four-dimensional space 160 freewill 306 freedom-causality 284

friendship 275-8,298 function 25,58-84; accordion effect 76; basic 68-74; and function substratum 242-4,246,250,252; structure and 244-6; hierarchy 66, 70-1; pure temporal Gestalt 94,95, 98;redescriptionof 66 functional relation 179 functional wholes: normative principles and 255-7; and their parts 239-54 functions 64-7; accordion effect 76, 239; basic 68-74; hierarchy 66, 70-1 fundamental state of affairs 31-4,41 Galilean physics 318-19 Galileo 179 game theory 304 gas law 172-6,177,179,181 Geisteswissenschaften 81 general wiU 298-303,307-8,333 generic dependence 139,143-4,182, 183,184-5,187-8 generic independence 144,182,183, 185,211 geometry 193-4,230-1,320,326-9 Gestalten, pure 85,87,259; actions causa sui 106-9; patterns and 90-2; temporal 95-6; temporal (causa sui) 97-105; in time, changes and 93-4 Gibson,J.J. 332 Giddens, Anthony 82 glass-door effect 267-8,280 Grice, H.P. 278,280 grounded relations: category 20; causality and 172,173-5,178,179; definitions 110,117-22,125; existential dependence 125,131, 132; force relation 180; intentionality and 205,206,212; relational space 149,150-1,152, 153 H a r r i , Rom 149,152,171 Hartmann, Nicolai 28,315 'heed concepts' 60,80 Hegel, J.W.F. 4,7,25,27-8,133, 140-1,287,288 Hegelian intuition 183 Hegelianism 110,111,133,292,293-4 'Heraclitan non-change' 102 hermeneutics 7,263 heterogeneity 42-3

370

INDEX higher actions 60-1,71 higher functions 66,71 Hjelmslev, Louis 23 Hobbes, Thomas 301 holism 6-8,110, 111,258-60,297 homoeomerous properties 42-3,53 homogeneity 42-3 'homomorphic things' 157 how-questions 60-1,63-4,65-6,70 Hume, David 2,175-6 Humean intuition 183 Husserl, Edmund 34; existential dependence 124,126-21,135, 137-40; intentional· ty 204,207, 216-18,288 Huygens, Christian 191 hysteresis phenomena 209,214,215, 223 ideal type 264,295-6,302,303 idealism 110,111,126,132-3,330 idealistic level ontology 22 identity-in-difference 16-17,142 immanent correlates 212 immanent realism 11,13,44,139,147, 184,316 impact relation 188-90,191,192 impossibility theorem 252,268-70 impossible figure 283,284 inclinations 209 inclusion 42-3,239 inclusive properties 43-50,229; semi53-5 inclusive qualities see inclusive properties, inclusive substances inclusive substances 55-6 inclusiveness: in space 74,81,232; in time 56-7,74-81,219,232 incommensurability thesis 318,325 independence: absolute/relative 134-5, 149; existential 131,134-5,143,144, 149; see also generic and total independent existence 31-2,34,40 independent parts 127,130,135, 136-7 indifference curves 62-3 individual accident 30,31,32 individual dependence 143,182-3, 184-5,187-8,211-12 individual property 30 individuation principle 156 induction problem 176 inertia 20,96-7,98,101-2,103,193-4

inertial system 158-60,172 inferences 113 infinitesimal calculus 99-100,104-5 Ingarden, Roman 28 inner perceptions 225 instance of a state of affairs 33-4 instances of universale 13 integral calculus 8-9,101 integration operation 101,103,105 intelligence 215-16 intensive/non-intensive magnitudes 50-2 intended causal consequences 71 intentional: act 205-7,209-14,222; correlate 197-206,209-14,219,223, 262; intention 214; mirror 276-7; object 197-8,211,223-4,226 intentionality 10,20,35,58,140; consciousness and 214—16; contradictory 267-70; at a distance 199,214,218-26/юшт; fictional 199-201,204-5,212; introduction to 196-9; mixed 204-5; naive realism in 216-26; nested see nested intentionality; other categories and 205-14; parts/ wholes without 227-57; presentational see presentational intentionality; real 197,199-201; representational see representational intentionality intentions 107-8; in action 72-4,82, 83-4; conscious 73,214-16 internal relations 110-12; existential dependence and 124-6,127,128, 130,132-3,138; formal logic and 113—17; grounded relations and 117-22; world and language 122-3 intuition, forms of 2-3 invariability 186 invariant relations 108,109,159-60 irreducibility, ofintentionality 206 irreducibly substantival universale 33, 36 irreductive materialism 221,22,305, 314; Aristotle 26-7; definitions 22-3; ontological levels 20,23-6,135,191,224; postscholastic level ontologies 27-9; substratum and 37-8 Johnson, W.E.

371

35

INDEX K a n t , Immanuel 2-4, 7,50-1,156, 248,263,284,320,327-8 kinematics 233-4 knowledge, and norms 255,256; sociology of 332-3 Koertge, Noretta 263 Köhler, W. 87 K u h n , Thomas 318,326,329 Laing, R.D. 280,281 Lakatos, I. 185-6 language 296; monomania 218; natural 321 'Lebenswelt' 216,217 Leibniz,G. 8-9,27,145-6,150,155, 224,261 Lenin, V.l. 316-17 level see ontological level Leviathan-structure 298,301,303 Lewin, L. 87 linguistic: level 122-3; phenomena 197,201,206; sign 23, 24,307-9,311,313,314; sociological idealism 133 Linnaeus, Carolus 280 logic 327-8; formal 113-17,128-9, 130; material 115,116,128-9;ofthe situation 61,64 logical connection argument 107,116 logical constants 113-14,115,116,130 logical impossibilities 116-21,128-9, 139-41,182 logical positivism 7 logical predicate 35,37 logical subjects 35-6,37,40-1 logical truth 114 Lorentz transformations 159 love 283-5,288 lower-dimensional parts 88-90,91,92, 137,229-32 lower-dimensional vector 100 lower actions 60-1,69,71,73 lower functions 66,71 Mach, Ernst 145-6,159 machines 228,254,255,256-7,296, 297,304 Maclntyre, Alasdair 72 macroagents 297-304,333 magnetic field 170,184-5,209 man: essence of 306-7,309-10; nature, society and 6,7,9,304-14 marginal utility 61

Marx, Karl 5,25,274,306-7,309 Marxist dialectics and intentionality 287-97,314 Marxist praxis-philosophy 263 mass 43,49,52,53,56,74; density and 45-7,50-1,90,230; determinates 46,83,85,87,125; as energy 27,38;equi-dimensional parts 229-30; existential dependence 136-7; Newton's second law 177-85; nouns 42 masses in masses 85 mass nouns 42 master-slave dialectic 287 material logic 115, 116, 128-9 material ontology 129,130,133 material substratum 206 materialism 132; intentionality and 216-17,218-20,221,224; irreductive see irreductive materialism materialistic level ontology 22, 28 mathematics 327-8 matter 26-7 meanings (analysis) 278-9 means-ends relation 248-9 mechanics 236 mechanisms 227, 257; and their parts 232-6 Meinong, A. 87 Melden, A.I. 59, 61 membership relation 260-1 memory 197, 199, 201, 209, 210, 215 mental subjectivity 217, 220-2,224 mere presentation 201 Merleau-Ponty, Maurice 197, 218 metaphysical systems 328-33 methodological individualism: action 59, 61-2, 63, 66-7, 68, 73; intentionality 263, 267, 280, 281 metric (space) 145, 157-8, 159, 160 microagents 297, 298-9, 302-4 Mill, John Stuart 185 mind-body problem 305 mind-independence 90 Minkowski 160 misintentionality 267-70; nested 279-83 misperception 200, 202, 215, 280 mixed intentionality 204-5 modal logic 115, 116 modal operator 141, 143, 182, 183, 330

372

INDEX modality 3; satisfaction 199-203, 205-6, 211, 267, 316-18, 322, 333 mode 135 Model I (society/man) 310-11, 314 Model II (society/man) 310-11, 314 Model I I I (society/man) 310, 311-12, 314 Model IV (society/man) 313-14 modern physics 9, 1 0 - l l , 27, 56, 98, 100, 108

momentary instant 10; see also temporal point moments 127, 130, 135-6, 166-7; part as 258-9 'monadological wholes' 261 monadology 217, 218, 224 monads 149 money 292-7 Moore, George Ε. 110, 111 motion (movement) 254; absolute 154-6; actions causa sui 106-7, 108, 109; as illusion 8-9; mechanisms 227, 232-6; relative 154-6, 158, 159; temporal Gestalten 98-9, 100-2, 103-5 Mulligan, K. 128, 129, 130, 138 mutual existential dependence 110, 130-8, 143, 173, 179, 181 mutuality 173, 177, 183, 205; feedback mechanism and 251; means-ends 248-9 Nagel, Т . 245-6 naive realism 9-10, 272; defence of 216-26,332 natural-causality 284 natural kinds 36 natural level 24, 37 natural science 81,216-17,218 nature (parts and wholes) 227-57 nature and society 5-6, 7, 9; man and 304-14 Naturwissneschaften 81 necessary singularity 149 necessity 115-16 negation 116,120 neo-Kantianism 7,263 Nerlich, G. 156, 157, 193 nested cognitions 333 nested contradictions 283-7 nested intentionality 270-9, 333; abstract structure 270-9; agents and

subjects 262-7; Marxist dialectics and 287-97; macroagents 297-304, 333; misintentionality 267-70; nested contradictions 283-7; nested misintentionality 279-83; parts 258-61; summary 304-14 Newton, Isaac 8-9, 145-6, 155, 159, 188; second law 177-85, 186-7, 189; tendency 162-5, 166, 167 Newtonian mechanics 43, 56, 96, 101, 107, 154-5, 157-8, 162, 167, 182, 187, 189, 193, 325 Newtonian physics 10,11,107,109, 162, 194, 207 nomic necessitation 175-6,177 nominalism 11,12,13,147,175,184 non-actions 69,83 non-causallaws 172-7,186 non-change 93-9, 101, 102, 103 non-intensive magnitudes 50-2 non-positing acts 201 non-quantifiable property 47-50, 51-2 non-relational ties 118 non-relative space 145-6,158-9,160 non-relativistic space 145-6 normative principles 228, 255-7 noumenal world 2, 4 numerical identity 13 numerical law 172, 173-5, 176-7, 179 object, intentional 197-8, 211, 223, 224, 226 objectivation 312 O'Connor, C.J. 39 one-sided existential dependence 130-8, 139, 143 ontological atomism 6, 7 ontological contexts 223-4 ontological inferences/constants 130 ontological level: Aristotle 26-7; category of 23-6; definitions 22-3; existential dependence 134-5, 137; historical examples 26-9; patterns and 82-4; post-scholastic 27-9 ontology: -centred philosophy 315; contemporary problems 5-11; determinable-determinate distinction 14-21; universal in re 11-14 ontology categories (theories) 1-5; epistemological status 328-33 open system 247-8

373

INDEX 'open texture' 122 'operational principles' 255-7 organism-level 24—5 organisms 228, 254-6, 297, 304 'overlapping' relation 152 pains 204,225 paradigm 326, 327, 331 'paradigm case argument' 141 paradigmatic relations 123 part-actions 238-9 part-functions 241, 243 part-motions 99 part-shapes 238-9,243 part-whole relations 10-11, 134, 139, 258-61; aggregates 228-9; exclusive/ inclusive properties 46-7, 49-50; functional wholes 239-57; Husserl on 126-30, 135, 138; lowerdimensional and equi-dimensional parts 229-32; mechanisms and their parts 232-6; spatial 24,46; unfinished descriptions 237-9 partial accelerations 168 partial forces 162-3 partial tendencies 163-6, 169, 208 participants 260-1; subjects and agents 262-7 'particularism' 11 particularity 11-12, 13, 30-1, 121, 137, 147, 184 partitioning (shortening) 48-50, 53, 55, 74-6, 85-6, 88 parts 34,44; dependent/independent 127, 130, 135, 136-7; equi-dimensional 89, 90, 91, 100, 229-32; four concepts 258-61; lower-dimensional 88-90, 229-32; potential 90, 91, 136-7; potentially parts 137; spatial; 53, 56, 90,91, 136-7 passivity (category) 2, 3 patterns 259; colour and density 45-6, 54, 55, 56; levels and 82-4; lower-dimensional parts 88-90; pure Gestalten and 90-2; as second-order properties 85-8; smallest possible 72-3 penetrables 190,191,220 perceivability 153—4 perception 201-5,210-13,217-25, 289,332-3

perpetuum mobile 209,214 phenomenology 197-8,216-18 philosophy and common sense 9-10 physicochemical descriptions 242-4, 248 physics: Aristotelian 9, 193-5, 207, 318-19, 326; modern 9, 10-11, 27, 56, 98, 100, 108; Newtonian 10, 11, 107, 109, 162,. 194, 207 Piaget, Jean 9 picture-contradiction 270 pictures 197, 199, 201, 206 'pieces' 127, 130, 131-2, 135, 136-7 place (category) 2 planning model (action) 72, 73 Plato 11,22 Platonic realism 184 plural self-dependence 138-9,148 plural self-excluding 148 pluralism 319 plurality-universal 138-9, 148 point-formed motion 232-3, 235-6 'pointing' 165, 169, 198, 199, 219, 260 Polanyi, Michael 28, 73, 228, 255-7 Popper, Karl 185, 316-17, 318 positivism 7 possibility 115-16, 150; logical impossibility 116-21,128-9, 139-41, 182 post-hypnotic suggestion 214 post-scholastic level ontologies 27-9 posture (category) 2, 3 potential existence 27, 38-9 potential parts 90,91,136-7 potentiality and disposition 169-70 predicate concepts 32, 33 predicate logic 113 preference logic 116, 323 preference system 62-3, 265 presentational intentionality 20, 201-4, 205, 208-145; naive realism 218, 220, 221-3, 225; nested intentionality 261, 269-70, 272, 283-4,307 pride 273, 275 prima materia 27, 38-9 'primary existents' 304 prior intentions 72, 73, 82, 83, 306 proper names 37 property: concepts 14-15, exclusive/ inclusive 43-50, 136-7, 227-39 passim, 259,260;-instances 13,31, 32, 33-4, 44, 54, 118, 121; quality

374

INDEX and 20, 31, 34-6, 42; semiexclusive/semi-inclusive 53-5; substance and 32-4, 35-6, 133-6; state of affairs and 20, 31, 34-6, 136; substratum and 36-41; universal and individual 30-2; variation 91,92 provability 114 'psychic work' 72 psychoanalysis 215 pupil-teacher relation 111, 113, 114, 115, 116, 117, 123, 274 'purpose sense'of function 253-4

representational intentionality 20, 201-6, 208, 210-12, 225; nested 261, 269-70, 272, 283-4 'resemblance' 12, 121-2 rest mass 43 resultant forces 162-3 resultant tendencies 163-6, 169, 208 Romantic philosophers 7 rotational motion 233 Rousseau, Jean-Jacques 298,302,333 'rules of lightness' 228,256 Russell, Bertrand 110, 11-12, 119, 152 Ryle, Gilbert 4-5,7,60,76

qualitative identity 13,122 qualities 2, 3, 19, 20, 31-2, 34-6, 42; inclusiveness in time 56-7, 136-7, 227-39 passim, 259, 260; instances 13, 31, 121;«« also property, state of affairs, substance 'quality-instance point of view' 13 quantifiable property 46, 47, 49, 51-2, 98 quantitative concept 174-5, 177-9 quantity 2, 3, 46, 47 quantum mechanics 10-11, 56, 326-7 Quine, Willard V. 320-2, 326

Sartre, Jean-Paul 270-2,273,274, 278, 280, 283, 285, 288 satisfaction modality 199-203, 205-6, 211,267,316-18 Saussure, Ferdinand de 23, 123 scalar addition 168 scalar quantities 98, 99 scholasticism 108 Schutz, Alfred 218 science 9-10; truth and 315-16,318, 326, 327-8, 330-1, 332 Scotus, Duns 126 Searle, John R. 69,70,71-2,81,83, 215 second order properties 54, 56, 82; patterns as 85-8 second substances 30-1,32,39 self-consciousness 226 self-movement 101-2 self-sustaining Gestalten 96, 97-8 semi-exclusive properties 53-5, 85 semi-inclusive properties 53-5, 85 'sensory content' 207 sentential logic 113 shame 271-2,273,274-5 shape 43,47,50-3,56; determinates 19, 48-9, 74-6, 78, 85-6; equi-dimensional parts 229-30; -instance 91, 136; size and 117-18,126,127,131 shapes in shapes 82, 85-6 similarity 121-2 Simons, P. 128, 129, 134, 138-9 simple ideas 68 simple parts 207 singularity-universal 148 size 117-18,126,127,131 smallest basic action 84 Smith, B. 128,129,130,138

rational action 263-5, 267 rationalism 328-9, 332 real intentionality 20, 197, 199-201 realism: immanent 11, 13, 44, 139, 147, 184, 316; naive 9-10, 216-26, 272, 332 realist general terms 13-14, 15, 18-19 reality 4, 104-5, 118 reciprocal action 248-9 reciprocity 3,4 redescriptions 58-64,66,68, 73 Reductive Principle 125 referring expressions 323-4, 325 refrigerator (study) 65-8, 70, 239-44, 246-8, 256 relation 2, 3, 4; see also converse, external, grounded, internal relational space 145-6, 149-54 relative motion 154-6, 158, 159 relative space 145-6, 154-7, 159 relative truth 316 relativistic space 145-6, 157-60 relativity theory 10, 11, 27, 56, 145, 193, 195, 325; general 195; special 158, 159-60

375

INDEX social nature 307, 309-10 social relations 306, 307, 309 social roles 274 socialization 312 societal level 24—5 societal matter 312-14 society and nature 5-6, 7, 9; man and 304-14 sociologist! 328,332-3 souls, theory of 26 space: absolute 145, 154; accomplishments in 78; container see container space; as efficient cause 193-5; exclusiveness in 43, 45, 56; inclusiveness in 74,78,81; metric 145, 157-8, 159, 160; relational 145-6,49-54; relative 145-6, 154-7, 159; relatives tic 145-6, 157-60; spherical 194-5; -time 20, 109, 134-5, 160, 195, 304 spatial: change 78; distance 210-11; 'hops' 223, 224, 225, 262; patterns 87,92 spatio-temporal part 258, 259-60 species 127 species-genus relation 35 spherical space 194—5 Spinoza, Benedict de 275 Spinozist intuition 183 spontaneity 20, 96-8, 102,. 103, 264, 274,283-4 spontaneous actions 72, 97; see also agents state 2, 3, 106-7, 109; see also change; non-change state of affairs: existential dependence 133-5, 136, 137; fundamental 31-2, 34, 41; intentionality in 35, 197-8, 202-3, 206-7, 209, 216, 222, 224-5; lowerdimensional parts 88-90, 91, 92, 137; qualities and 31-2, 34-6; and substance 32-5, 133-4; substance and property 20, 35-6, 136; substratum as 27, 36-41, 134—6 state terms 77, 79, 80-1 statement 201,202 steam engine 249-52 Strawson, Peter 118, 323, 325-6 stretching 50-1,52 structuralism 7 structure 25; function and 244-6

subactions 74-5,81 subject: agents and 262-7; category 7-8,9-10; concepts 32; intentional 197-9, 210, 212, 221-6; interaction 270-9; -object distinction 3-4;-pole 221-3, 225-6; -predicate logic 323 subliminal perception 214-15 suboptimality 289 substance 2, 3, 4, 19, 20; -accident distinction 30,31; concepts 32-3; exclusive/inclusive 55-6, 136-7, 227-39 passim; 259, 260; -instance 31, 33-4, 55, 56, 118; notions of 39-41; pattern 87; property and 32-4, 36-5, 133-4, 136; qualities and 20, 31, 34-6, 42; state of affairs and 32-5, 133-4, 136; and substratum 27,32,36-41,84, 134-6 substantial forms 2 7 , 3 0 - 2 , 3 9 , 4 1 , 108, 169 substratum 27, 32, 36-41, 84, 118, 134-6 subsumption relation 35, 36 superposition principles 162-3, 167-8, 186, 191 surface-problem 88 synonymy relation 321-2,327-8 syntagmatic relations 123 synthetic a posteriori 320-1, 323-8 synthetic a priori 320-8 synthetic truths 113 systems theory 228, 244, 246-8, 253 Taylor, Charles 59, 61 technology 255 teleological explanations 64, 67-8, 255-6 temperature 47 temporal extension 90 temporal Gestalten 95-6; causa sui 97-105; spontaneity and inertia 96-7 temporal patterns 87-8, 93-4, 95 temporal point 90, 93; see also exclusiveness in time temporal vector addition 165-9 temporally exclusive: action 78, 79, 80; universals 81 tendency 10, 20, 205, 229, 252, 254; Aristotle 161-2, 169, 196, 207; efficient causality 186-7,191,195;

376

INDEX intentionality and 207-9, 210, 213, 214-15, 219; Newton 162-5, 166, 167; potentiality and disposition 169-70; vector addition 165-9; velocity and 166-8, 170, 207 'tendential intention' 214 theory anomalies 331 thermodynamics 56 things (first/second order) 87, 146-7, 219 'this-operator' 143, 183-4 three-dimensional shapes 232, 233, 236, 237 three-dimensional space 88-90, 160 time 3, 11, 43, 45; exclusiveness in 79; inclusiveness in 56-7, 74-81; pure Gestalten, changes and 93-4 total independence 182 totalitarianism 298, 300 totality 3, 4, 5, 6-7 transcendent correlates 212 transcendent realism 11,12 transcendental ego 156, 218 transcendentalism 328,332 transformational model 312-13 translational motion 233 truth 114, 115-16, 200; absolute 316, 317, 320; synthetic 113, 320-8 truth-telling 141-2 truth-value 323,326 truthlikeness 315-20, 332, 333 two-dimensional space 88-9

ultimate actions 71, 83 ultimate functions 71 unconscious intentionality 214-16 understanding (categories) 3 uniform motion 154—5,166 unifying category 261 unity 129, 130

universal: -particular distinction 31, 137, 147-8; property 30-2; quantifier 141, 143; terms 37 universality 185-7 universale in re 11-14, 19, 31, 127, 142-3; see also immanent realism use value 292-6 passim Ushenko, A.P. 171 utility maximization 61-4 utterances 201-2, 203, 278-9 vector addition 165-9 vector concept 103-4,107,108 vectorial quantities 98,99,100,101, 103 velocity 96, 98, 100-1, 105-8, 172, 232-3; of colour change 102-3, 104; constant 99, 102; tendency and 166-7, 168, 170, 207 Vendler, Zeno 77-8, 79 veridical perception 210-11, 212-13 verisimilitude 200, 316; see also truthlikeness void 148 volume 19, 43-5, 49, 51, 52-3, 87; equi-dimensional parts 229-30 volume-determinates 44,46 Weber, Max 310 what-questions 59-60,66 wholes (functional): normative principles and 255-7; and their parts 239-54; see also part-whole relations why-questions 59-61,63-64,65-6,67, 70 'will of all' 298,299,303 wishes 169, 196, 199, 208 Wittgenstein, L. 6, 156-7 world, language and 122-3 Zeno' paradoxes

377

8, 190

Appendix 1: An aphoristic summary of Ontological Investigations. This appendix is a translation of a paper written for the Swedish journal Ord&Bild (2/1986). It contains a summary of my views in Ontological Investigations spiced with some of my political-philosophical views (5.41-5.51 and 6.41-6.45). The translation contains one change. In aphorisms 5.5 and 5.51, 'socialism' has been substituted for the original 'communism'. This mirrors the fact that I no longer believe that planned economies are, ceteris paribus, more efficient than market economies. The socialism spoken of is a society that manages to combine socio-political equality with the market mechanism. In both style and structure, but definitely not in content, I have tried to imitate Ludwig Wittgenstein's Tractatus Logico-Philosophicus. The main aphorisms 1-7 are meant to be directly comparable. Tractatus covers, somewhat consecutively, basic ontology, philosophy of language-andlogic, and a philosophy of the mystical. Naturally, I stick to the first, but exchange the other two for philosophy of intentionality and political philosophy, respectively.

Tractate on Reality Ingvar Johansson 1 1.1 1.2 1.21 1.3 1.31 1.4 1.41 1.42 1.5

Reality is all that is the case. The limit of reality is space-time. Space and time are also the foundation of reality. Facts exist in space and time. Both the limit and the foundation of reality belong to reality. Reality is not limited by anything outside reality. Even empty space can belong to reality. Where there is empty space there is no thing, but not nothing. Empty space is something, but not some thing. We know that reality comprises more than empty space.

2 2.1 2.11 2.12 2.2 2.21 2.22

2.3 2.31 2.32

2.33 2.4 2.41 2.42 2.43

2.44

2.5 2.51 2.52 2.53 2.54

It is the case that facts are complex. Many facts have spatial parts. What is spatially indivisible makes up a simple fact. If there are atoms they constitute simple facts. Even simple facts are complex. Atoms have to have at least two aspects, shape and volume. Pure sensations - if there are any - also have at least two aspects. Visual sensations have colour and shape; auditory sensations have pitch and volume. Some facts are necessarily extended in three spatial dimensions; neither more nor less. A sphere is necessarily extended in three dimensions, a circle in two. What is three-dimensional can be bounded by something which is two-dimensional. A sphere is bounded by a two-dimensional surface. What is two-dimensional is to the three-dimensional what a point is to a line. Some facts are necessarily extended in time, too. Some facts cannot be punctual in time. Actions are necessarily extended in time. You can neither make a revolution nor make love nor play football in a momentary point of time. There is not even time to take a shower. You might say that the photo finish shows who is the winner, and that therefore the winning is an action which takes place at a momentary point of time. I say: The one who has not run before the photo finish cannot win. Properties like colour and shape can exist at a momentary point of time. Mass, electric charge, force, electromagnetic field strength, and energy can exist at a momentary point of time. Velocity and acceleration can be punctual in time. Physicists think that all the facts of reality can be punctual in time. They are dazzled by differential equations with time derivatives.

380

2.6 2.61

2.62 2.63 2.7 3 3.1 3.11

3.12 3.2 3.21 3.22

3.23 3.24 3.25 3.3 3.31 3.32 3.4

Movements (changes of place) are necessarily extended in time. The integral over a velocity function v(t) from time point ti to time point t2 - ί v(t) dt - has to be zero if t2 equals ti. Change of place presupposes time extension between ti and t2. A movement can have but not be a velocity. What is necessarily extended can contain but not be a punctuality. What is extended (the movement) is not an epiphenomenon to the punctual (the velocity). Both actions and movements belong to reality. Some facts are directed towards other facts. Such directedness is called intentionality. Intentionality and consciousness are often assumed to be one and the same thing. To be conscious is to be conscious of something. Consciousness reaches out towards something; it has a direction towards it. If the unconscious belongs to reality, the area of the intentional is larger than that of consciousness. Intentional phenomena are facts of a special kind. Nature is the non-intentional part of reality. A billiard-ball can affect another billiard-ball, but the one cannot have intentionality towards the other. Direction of movement has nothing to do with intentionality. When I look at the billiard-ball, I am by means of intentionality directed towards the ball. When I think of the billiard-ball, I am by means of intentionality directed towards the ball. A billiard-ball can neither see nor think. Whether intentionality befalls anything more than human beings is to reality a contingency. Pain can only exist as an object of intentionality. Where we draw the boundary of intentionality is for animals a very essential matter. Intentionality can be either presentational or representational.

381

3.41

3.42

3.5 3.51 3.52 3.53 3.54 3.55

3.56 3.57 3.6 3.61 3.62 4 4.1 4.2 4.21

To see or to hear something means being presented for this something. What is being presented lays claim to existence in the intentional act itself (the perception). To think of something means representing this something. What is represented does not lay claim to existence in the intentional act itself (the thinking). Intentional acts can either reach what they are directed towards or fail in this. Veridical perceptions, true statements, and wishes which are fulfilled are intentional acts that reach their goals. Hallucinations, false statements, and wishes which are not fulfilled are intentional acts that do not reach their goals. Presentational intentionality that reaches its goal has an immanent object. Representational intentionality that reaches its goal has a transcendent object. Presentational and representational intentionality that do not reach their goals lack objects. They are, so to speak, pure directedness; like an arrow that points without pointing at something existing. In such cases there are within reality only the pointing and outside reality nothing. A hallucination has no corresponding fact, but the hallucination itself is a fact. A false statement has no corresponding fact, but the statement itself is a fact. Meaning has to belong to reality but cannot belong to nature . In nature everything is as it is, and everything happens as it does happen. Meaning is only lodged in intentionality. Intentionality forms the meaningful man. Mind is constituted by intentionality. Man is constituted by body and mind. The body is a necessary condition for the mind, but the body is not the mind.

382

4.22 4.23 4.24 4.25 4.3 4.31 4.32 4.33 4.34 4.4

4.41 4.5 4.51

4.52 4.53 4.54 4.55

4.56 4.57

Intentionality exists in our bodies in the same way as the meaning of words exists in the vibrations of the air. Mind and body make up a unity. They are fused. In this fusion we are living. Or, better phrased: We are such fusions. Mind and body are both different, fused, and partly shaped by each other. Mind and soul are one and the same thing. The soul belongs to reality. When the soul dies it becomes nothing. It falls outside reality. When the body dies its parts are spread in space. The fundamental parts belong to reality for ever. The parts of your body are immortal, but you are not. You can believe in the soul without becoming an idealist, but you cannot believe in the soul's immortality without becoming an idealist. You can become a non-reductive materialist. Emotions are fusions of the rest or restlessness of the body and the directedness of the mind. When you are glad you are glad about something. When you feel sorry you feel sorry for something. When you are kind you are kind to someone. When you are angry you are angry at someone. The directedness of our emotions shows itself in language in the form of prepositions. Emotions, just like perceptions and thoughts, have intentionality. Happiness makes it hard to stand still. If emotions had legs they would have one in the body and the other in the soul - and the emotions would not be able to walk unless the legs were connected. The exceptions prove the rule. Life anxiety and general euphoria have a directedness, too. They are directed at everything and nothing.

383

4.6 4.61 4.62

Mind can only discover itself indirectly; like an eye it needs a mirror in order to see itself. Therefore, the mind easily appears mystical. Hereby, to the natural scientist, the most ordinary easily becomes the most mystical.

4.63 4.64

What is ordinary cannot be mystical in principle. What is large-scale mystical - transcendent reality beyond

4.65

space and time - does not exist, and, consequently, is no reality. Not even a mystical reality. What is small-scale mystical - mystical for the time being - exists.

4.66 4.67

Solar eclipses were but are not mystical. The tidal waves were but are not mystical. Breeding was but is not mystical. What is mystical for the time being is not mystical in principle.

4.68

The soul is mystical for the time being.

5

Social reality is a function of nature and nested intentionality.

5.1

Nested intentionality is intentionality directed towards another intentionality which is directed towards the first intentionality.

5.11

When you look someone in the eyes, you do see this other, but you see that he/she sees you, too. When the other sees you, he/she does not only see you but also your seeing o f him/her. My seeing o f you encompasses your seeing o f me, and vice versa. We are woven together by presentational intentionality.

5.12 5.13

In spite of the separation of our bodies we make up a unity. A unity constituted by the peculiarity of intentionality.

5.14

Both psychologically strong persons and psychologically weak ones form parts o f such unities. Some emotions are essentially social.

5.2 5.21

Pride is primarily proudness of oneself before others.

5.22

Pride presupposes that one is directed towards others who are directed towards oneself. Pride presupposes nested intentionality.

5.23

Shame is primarily shame of oneself before others. Only secondarily it can become shame of oneself before one's super-ego.

384

5.24

5.25 5.26 5.27 5.28 5.29 5.3 5.31 5.32

5.33 5.34 5.35 5.36 5.4 5.41 5.42 5.43

By means of nested intentionality we jump into each others' lives in the way two opposed mirrors are reflected into each other. The image in the one mirror involves essentially the other mirror, which, in turn, essentially involves the first one. In love and friendship we mirror ourselves in one another. This goes for hatred and enmity as well. In the latter cases we are nested by intentionality into a division; in the former cases we are nested into a communion. Even a nestedness into a communion is a nestedness of two intentionalites. The unity does not cancel the two-ness. Even in love you are two. The unity the mystics are longing for does not exist. Social reality contains considerably more than social emotions. Reality contains bodily powers which are connected to souls; also, it contains bodily powers which are connected to nested souls. These powers can work upon nature, and they can take already worked upon nature to further the appropriation of nature. Nature becomes both an object of labour and a means of production. Speech acts in language have a structure. They conform to a grammar. The grammar belongs to social reality. In labour, as it is in language. The grammar of labour is called relations of production. Human activity can take on many forms. Labour is only one of them. Labour is either necessary labour, surplus labour, or both at once. Necessary labour is the activity required to reproduce the labouring people at a given level of living. Surplus labour is the activity required to make new investments and to sustain the luxury consumption of the upper class.

385

5.44

5.45 5.46 5.47

5.48

5.49 5.5 5.51 5.6 5.61 5.62

5.63 5.64 5.65 5.66 5.7

In class societies, the upper class has power over the surplus labour of the lower class. There is exploitation. The lower class is not allowed to perform its necessary labour unless it also performs surplus labour for the upper class. The upper class robs the lower class of some of their time on Earth. Exploitation is only one of a multitude of forms for human oppression. The lower class is often forced to perform even the necessary labour in ways which are far from free. The lower class is oppressed also when it only reproduces itself. Even absence of exploitation can be a form of oppression. For almost everyone it holds true: better to be exploited than to be a beggar. For many it holds true: better to be exploited than to be unemployed. The world of the oppressor is a different one from that of the oppressed. Socialism is a grammar without oppression and exploitation. Socialism is like Esperanto: a possible structure searching for a people. Structures come and go. Social structures are, simultaneously, both condition and outcome for certain forms of human activity. If our language had no structure, we would not be able to understand one another. You would not even understand what you yourself are saying. If no one talked there would be no language structure. In labour, as it is in language. Structures do not walk with legs of their own. Structures are changed by the people inhabiting them. The structural determination of social reality differs from the determination of nature. It is of the human essence to have two essences. First, a biological nature, second a social nature.

386

5.71 5.72

5.73 5.74

5.75 5.76

5.77

6. 6.1 6.12 6.13 6.2 6.21

6.22 6.23

Your social nature is biologically founded, but what is social is not biological. A biological nature is inherent in each single individual. A social nature, on the other hand, is constituted by relations of nested intentionality. Almost all human beings have a biological nature containing general capabilities to become socially programmed. A man speaking Chinese and a man speaking Swedish have the same biologically determined general capability for language, but without the social specification the capability will remain a potentiality. In our drives, as it is in language. Everybody has a general drive for food. Hunger is hunger, but hunger which has to be gratified by cooked meat eaten with a knife and fork is a different hunger from that which will be gratified by raw meat torn apart by hand, nail, and tooth. Our social nature, just like our biological nature, offers resistance to changes. It is hard to eat raw mice. You do not change your sexuality by sheer will-power. The general form of intentionality is S —» O. This is also the general form of unreality. When a subject S looks at or thinks of an object O, this subject is directed towards the object. Reality includes the form S O. A subject can also be directed towards another subject, S| —> S2. The basic form of nested intentionality is Si (S2 —> Si). The general form of intentionality is S —> О even when О does not exist. Hallucinations, mirages, fantasies and false beliefs have no O, i.e., there is no O, but the corresponding intentional acts have the form s-»o. Unreality is brought into reality. Our world may differ from the world, and your world may differ from mine.

387

6.24

At the railway station: You can see that the train next to you begins to move, although, in reality, it is your own train which is leaving. You are intentionally directed towards something partly unreal.

6.3 6.31

Reality sometimes creates an objectively given unreality. The mirage in the desert is not due to any fault in your subjective perceptual system. The mirage is determined by the refractions o f light.

6.32

At dawn, Earth turns around so that your meridian begins to enjoy the light. But in your eyes the sun rises above an immobile horizon.

6.33 6.4 6.41

Perceptions o f nature can contain an unreality. Social reality often contains, as an essential part, an unreality. Any God belongs to unreality.

6.42

God has been in the habit of saying that those of lower rank in society ought to serve those of higher rank, and women serve men.

6.43

This real unreality has been justification both for class society and patriarchy. Modern capitalist societies - as well as existing planned economies - conjure up illusions with similar effects.

6.44 6.45

As the sun actually rises in our natural perception, the wage labourer is actually in our social perception being paid for the whole working-day.

6.5

Unreality determined by reality can be studied. It belongs to reality.

7

We can speak about reality, some choose to pass over it in silence.

388

DETERMINABLES AS UNIVERSALSI

According to immanent realism, there are universale in the spatiotemporal world quite independently of language and the mind. The existence of these universale, furthermore, is not dependent upon there being Platonic universale existing outside the spatiotemporal world. In this paper I will try to show that immanent realism holds not only for many determinate universale, but for some determinable universale ae well. In other words, there are ontological determinables as well as conceptual determinables. 1. Determinables and Determinates The terms 'determinables' and 'determinates' were made important in philosophy by the Cambridge philosopher W. Б. Johnson at the beginning of this century,2 even though the senses of the terms have a history dating back at least to Aristotle. Franz Brentano lectured about similar ideas in Vienna in 1890-91, but his lectures were not available in print until 1982.3 What Johnson was to call 'determinables', Brentano called 'logical parts'.4 Johnson's distinction between determinables and determinates is applicable both to universal concepts and to universale in re. Examples abound. The concept scarlet is a determinate of the concept red, and the latter concept, in turn, is a determinate of the concept colored).5 Conversely, the concept colored) is a determinable for the concept red, which, in turn, is a determinable for the concept scarlet. Two main features of the distinction between conceptual determinables and determinates are: (a) if a determinate concept (e.g., red) can be truly predicated of something, then at least one determinable concept (e.g., colored) must necessarily also be predicable of exactly the same thing; (b) if a determinable concept (e.g., colored) can be truly predicated of something, then there must be some (though unspecified) determinate concept (e.g., red) that is also truly predicable of exactly the same thing. "Determinables as Universal!" by Ingvar Johansson, The Monist, vol. 83, no. 1, pp. 101-121. Copyright О 2000, THE MONIST, La Salle, Illinois 61301.

389

INGVAR JOHANSSON

A determinable concept subsumes its determinate concepts. The determinable-determinate relation is a relation which connects two universale. It is not a relation between a universal and a particular falling under this universal. If a certain particular instantiates a certain determinate, it instantiates the corresponding determinable too. The two features of the determinable-determinate relation mentioned, (a) and (b) above, characterize the genus-species distinction too. However, Johnson himself wanted to keep the determinable-determinate relation distinct from the genus-species relation. Also, he added a third characteristic, namely (c) that two determinates (e.g., yellow and red) of the same determinable (colored) cannot exist in the same particular at the same time. These latter views of Johnson will be discussed in the last section of this paper. As a foil and a point of departure for my arguments, I will use the views of David Armstrong.6 When Armstrong, with his two-volume book Universals and Scientific Realism (1978), managed to put Aristotelian immanent realism on the modern philosophical agenda, he also discussed the distinction between determinables and determinates.7 No determinable concept can, he claimed in that book, refer to a real determinable universal in re, since there are no such (ontological) determinable universals. All conceptual determinables should, he then said, be conceived of as classes of determinate universals. His immanent realism was in this sense a very restricted realism. Armstrong soon, however, changed his mind. Already in What Is a Law of Nature? (1983) he said that there probably are some ontological determinables, too.8 He pointed out that the functional laws of physics, if realistically conceived, seem to require that there are ontological determinables. And this position is endorsed again in Armstrong's latest book, A World of States of Affairs (1997).» It should be noted, though, that in Armstrong's view it is only numerical functional laws which require ontological determinables: "It is a plausible conjecture that all ontological determinables are quantities In my view, there is much more to be said in favor of the spatiotemporal existence of determinable universals, and there is no reason to restrict a belief in their existence to quantities. When the distinction between determinables and determinates is applied to universals in re, the main features of the distinction can be described as in the following quotation from Armstrong: 390

DETERMINABLES AS UNIVERSALS (a) If a particular has a determinate property, then it is entailed that the particular has the determinable property. Necessarily, if a thing is triangular, it has a shape. Necessarily, if a thing is red, it has a colour, (b) If a particular falls under a determinable, then it is entailed that it has one of die corresponding determinate properties, although it is not entailed which. Necessarily, if a thing has a shape, it has a particular shape. Necessarily, if a thing is coloured, it has a particular colour.11

In the next section I shall present my basic reasons for being an immanent realist with regard to determinates. The remaining sections deal with determinables. Those readers who will not be convinced by my very brief argumentation for immanent realism can, e.g., consult Armstrong's works for detailed arguments against various forms of nominalism and conceptualism.12 2. Immanent Realism Before considering universale in re, i.e., wholly mind-independent and language-independent universale, I will consider universale which are mind-dependent but language-independent, namely universals in perceptual objects. Obviously, there are, as a matter of fact, determinate universals or tropes in perceptual objects. Most people have sometimes perceived a determinate color hue at two different places at one and the same time. And the same is true of perceived shapes. Two of our most common everyday perceptual properties can be at several places in our perceptual space at one and the same time. This means that they are either universals or tropes. From a realist point of view, they obviously conform to the usual characterization of universals: a universal is an entity which can be at two places simultaneously. A universal can have instances. According to immanent realism, a universal is identical in and with all its instances. As far as I can see, all forms of nominalism (including conceptualism)13 rejecting tropes must, confronted with the existence of perceived colors and shapes, make the claim that universal linguistic concepts have unconsciously structured our perceptions. In my view, such variants of nominalism implicitly claim that the basic non-linguistic particulars of the world are not structured. Such forms of nominalism, pursued to the (bitter?) end, must hold that not even a color sensation can be ascribed a qualitative identity of its own. Such a sensation must then be a particular 391

INGVAR JOHANSSON

which, in spite of the fact that it can be structured by a linguistic universal, has no qualitative identity of its own. However, the idea of an aggregate of propertyless particulars seems to be a contradiction in terms. What, in such an aggregate, makes one particular distinct from another? How can two spatiotemporal entities be spatially distinct from each other if both lack a shape? If they do not have a shape they do not seem to have a border, and the idea of a finite spatiotemporal particular without a border seems contradictory. So much for trope-less nominalisms. But what about the view that tropes exist? What about the view that in each perceived color spot and in each perceived shape there is a trope, i.e., a nonrecurring determinate color and a nonrecurring instance or "moment" of a determinate shape, respectively? Let us think of two exactly similar color spots, i.e., the spots have the same color hue with the same degree of intensity and of saturation. According to the trope analysis, there are then primarily two exactly similar tropes but no universal. Because of the existence of the resemblance relation we can, secondarily, construct a universal concept, but there are not in our perceptual objects any universals. The difference between immanent realism and trope nominalism can, I think, be traced back to different views on the relationship between the relation of exact resemblance and (generic or) qualitative identity of properties. Trope theorists have to claim that exact resemblance is logically prior to qualitative identity of properties,14 whereas immanent realists have to make the opposite (and true) claim: qualitative identity of properties is logically prior to the relation of exact resemblance. When, according to immanent realism, we perceive two instances of the same lowest specific determinate color, then we have two numerically distinct but qualitatively identical color spots. Secondarily, and due to the qualitative identity of the properties, there is also a relation of exact resemblance. This relation supervenes "with increase in being"15 upon the colors of the relata, and it is not identical with the qualitative identity of the two color instances in question. The difference between the qualitative identity of a property and the relation of exact resemblance might be subtle, but the difference is there.16 Let me explain. Assume that we perceive a determinate volume relation, e.g., being larger than, between A and B. The relation can be described by: "The volume of A is larger than that of B"; but this description loses some of 392

DETERMINABLES AS UNIVERSALS

the perceived determinateness. In my view, the perceived determinate relation is an ontological universal, although it has to be grounded in two ontological property universale, i.e., in the perceived volumes of the relata A and B. Similarly, the relation of exact resemblance has to be grounded in some properties of the relata. Resemblance, necessarily, has to supervene upon something. If A and В were the only entities in the universe, and A were larger than B, then this relational fact would be dependent upon the fact that A has one determinate volume and that В has another one. If A were to pass out of existence, then the relation being larger than would also pass out of existence, but B's volume (the property) would exist nonetheless. Conversely, if В were to pass out of existence, then the relation would again pass out of existence, but A's volume would still exist. Quite generally, there is an asymmetry between relations of this kind and the connected (= subvenient) monadic properties. The properties can exist, with their qualitative identities, without the relation, but the relation (with its qualitative identity!) can not possibly exist without the properties. The kind of ontological asymmetry that exists between the relation being larger than and the property instances it relates, seems to be necessary also between the relation of exact resemblance and the property instances it relates. As far as I can see, a trope theorist cannot ascribe properties to a trope in a one-trope-world without taking recourse to counterfactual resemblances. But such a move makes no sense to me. Counterfactuals can be used to explicate concepts, but they are unable to perform any ontological work. By definition, a counterfactual describes something which does not exist. Therefore, in my view, the existence of the relation exact resemblance cannot possibly explain the existence of any property universal, whereas qualitative identity of properties can explain the existence of the relation of exact resemblance. The structure of this brief argument of mine for immanent realism— with regard to determinates—can be represented as follows: a) Either immanent realism or trope nominalism is true of perceived determinate colors. b) Of these two views, the one which is true of determinate perceived colors is also true of the determinates of perceived shapes and volumes. 393

INGVAR JOHANSSON

c) That view which is true of determinate perceived shapes and volumes, ought to be true as well of the mind-independent non-perceived determinates of shapes and volumes. d) Qualitative identity of properties is logically prior to the relation of exact resemblance; therefore, immanent realism makes more sense than trope nominalism. e) Immanent realism is true in relation to both perceived and mind-independent determinate universale. I will now regard immanent realism in relation to determinates as proven true, but discuss it in relation to determinables. (A trope philosopher can of course in a corresponding way try to prove that there are trope-determinables.)17 My conclusion will be: f) Immanent realism is true in relation to some perceived determinable universale, as well as in relation to some wholly mind-independent determinable universals, i.e., there are ontological determinables. 3. The Gap Argument Between the lowest determinate (mind-dependent phenomenal) colors, between the lowest determinate (both phenomenal and mind-independent) shapes, and between the lowest determinate (both phenomenal and mind-independent) volumes, there are relations of resemblance. For instance, relations like a is more like b than c. With the help of such resemblance relations, determinate volumes can in thought be ordered on a line where each volume-determinate is ascribed a punctual position and a real number. Phenomenal colors cannot, like some of their underlying causes, the wavelengths of electromagnetic radiation, be quantified. Nonetheless, they can be ordered. In the so-called color spindle (see Figure 1) each color is ascribed one single point. Each point in this "color space" represents a specific color-hue (the periphery of the circle in the Figure), a specific color-intensity (the vertical line), and a specific colorsaturation (the radii of the circle). The resemblance relations among shapes, to take a third example, are even more complex. Even though some subclasses of shapes can be ordered on a line (for instance, ellipses according to their eccentricity), nobody seems as yet to have managed to 394

DETERMINABLES AS UNIVERSALS

construct a "shape space" in which all possible shape-determinates can be ascribed a point. But, perhaps, some day it will be done.18 White

Figure 1 Now, first of all, I want to remove a possible misunderstanding. I am going to argue that volume, color, and shape are ontological determinables, but that does not mean that their determinates have to be thought of as points on a line, where the line represents the determinable. Volumes may be thought of in this way, but shapes cannot (at least today). Trivially, two arbitrarily chosen volume-determinates can always, in thought, be continuously connected by a chain of intermediary determinates. The color spindle shows that the same goes for colors, too. With respect to colors there are even infinitely many such chains between any two points. Shapes can today neither be quantified nor ordered in a "shape space." Nonetheless, even two arbitrarily chosen shape-determinates can always, in thought, be continuously connected by intermediate determinates (see Figure 2). One small reservation, however, has to be added. The expression 'continuous connection' is not meant in a strict mathematical sense. According to geometrical topology, a square, to take merely one example, cannot possibly be turned into a circle unless three mathematical points are taken away.19 But such infinitesimally small lacunas and

Figure 2 395

INGVAR JOHANSSON

Despite all their differences in quantifiability and orderability, the determinates of volume, color, and shape, have one feature in common. It can be expressed as follows: two different determinates of the same determinable can always be continuously connected. Can, similarly, two determinables be continuously connected by a chain of intermediaries? If such a connection is possible, it would of course also connect some corresponding determinates. But this seems to be impossible. No color-determinate resembles more closely a shape-determinate or volume-determinate than any other. There is, so to speak a necessary lack of resemblance between the determinables under discussion. We simply cannot conceive of something which continuously connects them with each other, and this non-conceivability seems to be grounded in the entities themselves rather than in a weakness in our faculty of imagination. In what follows, I will sometimes talk of this peculiar fact as the gap between determinables. "Lack of resemblance," in the sense now spoken of, must not be conflated with "being very dissimilar," Two shapes which are very dissimilar can nonetheless be continuously connected by means of other determinates. Resemblance admits degrees, and being very dissimilar is merely "long-distance resemblance." Lack of resemblance is something else. All determinate colors seem to differ from all determinate shapes in exactly the same way. The simplest explanation of this lack of resemblance is that all color-determinates have something in common, namely the ontological determinable of color. All the shape-determinates have something else in common, namely the ontological determinable of shape; and similarly for volumes. The determinables, in turn, are simply different. Usually, conceptual determinables can be ordered into levels like the concepts scarlet, red, and coloredj.20 My claim about ontological determinables is by no means meant to imply that there are ontological universale which correspond to the conceptual determinables of every such level. On the contrary: where there is no gap there are no separate ontological determinables. In my view, the concepts volume, color, and shape can be used to refer to ontological determinables since there are gaps between them (and there are corresponding ontological determinates), but the concepts of red, orange, and yellow cannot refer to ontological determinables since there are no gaps between the concepts. The limits of the latter concepts are conventional, and the extension of each concept is merely a class of lowest ontological color-determinates. 396

DETERMIN AB LES AS UNIVERSALS

None of these determinable concepts can be used to refer to an ontological determinable.21 The remark just made applies, with some reservations, to shapes and volumes, too. Most ordinary shape concepts are, like the ordinary color concepts, only conceptual determinables, i.e., they have as their extension merely a class of different shape-determinates. Some shape concepts, however, are special in the sense that they have exactly one determinate in their extension. Examples are the concepts perfectly circular and perfectly square-shaped. Tbey can only be used to refer to one lowest ontological determinate each. Such concepts ought to be called 'lowest possible conceptual determinates'. The ordinary concepts round and square, however, have as their extensions classes of lowest ontological shape-determinates which contain more than one member. Everyday volume concepts, like large and small, behave of course like the ordinary concepts round and square. Their extensions are multimembered classes of ontological determinates. Note, however, that every determinate quantitative concept, e.g., the concept 5,013 cm3, should, like the concept perfectly round, be regarded as a conceptual lowest determinate. It can only be used to refer to one ontological determinate. Such a quantitative concept is of course relational, whereas the monadic property it may refer to is non-relational. But there is no contradiction here. The explicit referent of a relational concept can very well be a non-relational entity. If ontological determinables are accepted as that which explains the gap or lack of resemblance between classes of determinates, then they can also fulfill other explanatory functions. Prima facie, resemblance is always resemblance in a certain respect. Two things cannot just be similar or dissimilar. They have to be similar or dissimilar with respect to volume, color, shape, or some other such property. The hypothesis of ontological determinables explains in a simple way why resemblance is always resemblance in a certain respect. (This is also Armstrong's view.)22 4. The Argument from Physical Magnitudes Before I put forward my next argument in favor of ontological determinables, I will briefly present Armstrong's argument, since, like mine, his argument is centered around the magnitudes of physical laws. As I said in Section 1, Armstrong thinks that the existence of functional laws 397

INGVAR JOHANSSON

indicates the existence of ontological determinables. The simplest functional laws have the form Q = f(P), where Q and Ρ are determinable universale. A real example from the history of science would be Boyle's law for gases: ρ · ν = constant (at constant temperature); ρ is the pressure and ν the volume.23 Armstrong's argument for the view that a law like Boyle's, if true, implies the spatiotemporal existence of the determinable universale pressure and volume, relies on a very specific assumption. He "assume[s] the truth of what may be called Actualism."24 According to this kind of Actualism, firstly, dispositions and powers cannot be real; and, secondly, a genuine law cannot be uninstantiated, i.e., it cannot be about the merely physically possible. In Armstrong's (and my) view, true functional laws represent necessitation relations between universals. Given this assumption, one way to analyze Boyle's law, ρ · ν = constant, is to view it as a mere shorthand for an infinite conjunction of laws. Each such law would then relate two determinate universals to each other: px if and only if v,, p2 iff v2, p3 iff v3, рл iff v4, and so on ad infinitum. Surely, a lot of values of ρ and ν will never correspond to any determinate universals in re, and this creates a dilemma for the Actualist Armstrong: either all functional laws must be given an instrumentalist interpretation, or some functional laws must have a reference which makes the whole law refer to something actual. Since, obviously, Armstrong wants to avoid the first horn of the dilemma, i.e., dismiss all possible functional laws of the natural sciences as being a priori instrumentalist, he must find some way in which functional laws can fit in with his Actualism. His solution is to regard the determinable concepts of true functional laws as referring to ontological determinables. What determinables there are in the world can, according to Armstrong, only be decided a posteriori. Armstrong's dilemma is not a problem for immanent realism in general. In my opinion, we can speak and think of non-existent (= non-instantiated) universals, just as we can speak and think of non-existing entities like unicorns. Boyle's law would then (even if true) not be only about really existing determinate universals. It would imply a lot of counterfactual truths about all the non-instantiated relevant determinate universals as well. If the (assumed) non-existent pi9 would become instantiated, then necessarily v39 would have to become instantiated, too. 398

DETERMINABLES AS UNIVERSALS

Since I think that Actualism is false, Armstrong has in my view reached a true conclusion with the help of a false premise. Even Armstrong himself, it should be noted, finds Actualism "most difficult and uncertain,"25 but nonetheless he believes in it. There are, however, other arguments in favor of the existence of determinables which appeal to physical magnitudes and functional laws. Let us first look at additive magnitudes. A numerically determinate physical magnitude can be regarded as a determinate number associated with a physical dimension. Now, it only makes sense to add (or subtract) magnitudes which have the same dimension. 5 cm3 + 3 cm3 = 8 cm3, and 5 kg + 3 kg = 8 kg, but 5 cm3 + 3 kg is not equal to 8 of any physically meaningful magnitude. One determinate volume can be added to another determinate volume, and one determinate mass can be added to another determinate mass, but it makes no sense to add a volume to a mass. In physics it is taken for granted that it only makes sense to add quantities which have the same dimension.26 The best explanation of this (often tacit) assumption, is that each dimension differs from all other dimensions by gaps of the kind described in the preceding section. And, therefore, when the determinates of such a dimension are given a realist interpretation the dimension (variable) itself represents an ontological determinable universal. In the Gap Argument it was assumed from the start that all the determinables discussed (phenomenal color, shape, and volume) have ontological determinates. In the Argument from Physical Magnitudes this assumption has to be relaxed. Obviously, a lot of variables in physics have to be given an instrumentalist interpretation. Even though, from a physical point of view, magnitudes with different dimensions can never be added or subtracted, they can be multiplied (as in Boyle's law) and divided by each other. Does, perhaps, this fact tell against the view that some physical dimensions are best interpreted as representing ontological determinables? I think not. Rather, it supports this view, as I will try to show by some brief remarks on (scalar) multiplication. As already said, a numerically determinate magnitude consists of a determinate number associated with a physical dimension. This means that we are performing two operations simultaneously when we are multiplying physical magnitudes. We are not only making a purely mathematical multiplication of the numbers at hand, we are also multiplying the corre399

INGVAR JOHANSSON

sponding dimensions. If a pressure-determinate, e.g., 6 N/m2 (Newton per square meter) is multiplied by a pure dimensionless number, say 7, we get another pressure, 42 N/m2. This is equivalent to the addition 6 N/m2 + 6 N/m2 + 6 N/m2 + 6 N/m2 + 6 N/m2 + 6 N/m2 + 6 N/m2. However, if we multiply a pressure of 6 N/m2 with a volume of 7 m3 we get the number 42 associated with the new dimension Nm (Newton times meter). If we multiply 1 N/m2 with 1 m3 there is (since 1 times 1 equals 1) no purely arithmetical difference between the magnitude of the product and the magnitudes which are being multiplied, but nonetheless there is an important difference in physical dimension. N/m2 times m3 equals Nm. Two things should now be noted. First, it holds true in general, that a product must have another physical dimension than those of the multiplicands. Second, the multiplicands can always be represented by means of an x-axis and an orthogonally opposed y-axis in an abstract space. The second point requires further comment. When one class of determinates (e.g., determinate x-lengths or determinate pressures) is represented as being orthogonally opposed to another class of determinates (e.g., ylengths or volumes, respectively), then all the determinates of the first class are represented as being different in exactly the same way from all the determinates of the second class. The simplest explanation of the fact that orthogonal representations make sense and works, seems to be the assumption that there is a necessary lack of resemblance between the determinates of each class. Since a determinable D, (e.g., x-length or pressure) is simply different from another determinable D2 (e.g., y-length or volume, respectively), all the determinates of Dt must differ from all the determinates of D2 in the same way. The two remarks made apply also when the multiplicands seem to have the same physical dimension, as, e.g., in 6 m times 7 m. In spite of the fact that both these multiplicands have the same dimension, meter, the product has another dimension, m2. If all the magnitudes are realistically interpreted, the multiplicands refer to length-determinates in real space, whereas their product refers to an area-determinate. Furthermore, the two length-determinates, 6 m and 7 m, have to be regarded as belonging to two different dimensions which are orthogonal to each other in real space, an x-dimension and a y-dimension. The multiplication makes no physical sense if both magnitudes are regarded as belonging to, e.g., the xdimension. This, in turn, means that these two length-determinates, in 400

DETERMINABLES AS UNIVERSALS

spite of appearances, belong to different physical determinables. From a realist physical point of view, the multiplication above is best represented as 6 пц · 7 my = 42 m2. Real space is three-dimensional, and each such dimension constitutes an ontological determinable in the sense under discussion. Obviously, some (and obviously not all) physical dimensions ought to be given a realist interpretation. Therefore, the reflections above indicate that if the existence of ontological determinables is denied then addition and multiplication in physics become a mystery. 5. The Argument from Patterns The gap argument makes no difference between quantitative and non-quantitative determinables, but the argument from physical magnitudes applies only to quantities. Like the gap argument, the argument from patterns, which now follows, is independent of the quantifiäbility of the determinables discussed. The most pervasive and conspicuous patterns in our everyday world are color patterns. Let us make a partial analysis of them.27 A color pattern is more than an aggregate of colors situated close to each other. It consists of determinate colors with determinate shapes, i.e., the components of a color pattern are spatial unities of determinates of colors and shapes. Any finite uni-colored color spot must have a border. If the border is sharp the border has a determinate shape, if the border is blurred it has a blurred shape, but there is nonetheless a border and a shape. Every kind of property pattern has components which are properties-with-shapes, and the components of color patterns are colors-with-shapes. This means that the components of a color pattern are also spatial unities of two different determinables, color and shape.28 I will now put forward a reductio ad absurdum argument. Assume that the concepts of color and shape do not reflect any ontological determinables. This assumption entails that each component of a color pattern is a unity of two determinates which belong to two conventionally created classes of determinates. Since these classes are merely man-made constructions, it must be possible to think of components of color patterns in which a color-determinate is united, not with a shape-determinate, but with a determinate universal u which belongs to another class. All we have to do is to delimit the class of shape-determinates in a new way 401

INGVAR JOHANSSON

which makes и a member, too. But such unities, colors spatially fused only with non-shapes, seem impossible to think of as components of any color pattern. Therefore, the conceptual determinables color and shape cannot delimit only classes of determinates. They must reflect ontological determinables too. 6. The Observability of Determinables In this section I will try to counter a prima facie objection to my belief in ontological determinables.29 Above, I have explicitly or implicitly made the following three claims: (i) There are determinate universale in perceptual objects. (ii) A determinate and its determinable are instantiated in exactly the same spatiotemporal region, therefore there are also determinable universale in perceptual objects. (iii) Determinable universale of determinate universale in perceptual objects cannot possibly be unobservables. My arguments commit me to the view that whoever perceives a determinate universal perceives its determinable as well. If no determinables are observable, I have to reject my realism with regard to determinables. The question I have to face can also be put like this: we perceive determinates of volumes, colors, and shapes, but do we really perceive the corresponding determinables? It is an established truth in perceptual psychology that every perception contains a foreground-background duality. This duality is a spatial duality, and can equally well be called a center-periphery duality. Every perception has a spatial center which "stands out" against a surrounding spatial background. There are, however, some other dualities in our perceptions which very well can be called foreground-background dualities too. When we perceive a material thing, we primarily see one of its sides, let us call it the front. But in some sense we also immediately see that the thing has a back and an inside. The thing is immediately given as a whole thing. It makes a lot of difference to see a facade as a fa;ade of a real house or as a mere (movie) fa9ade. In a perception of a material thing, the 402

DETERMINABLES AS UNIVERSALS

front is a directly seen foreground which stands out against a background of indirectly seen sides. That which is "really seen," the front, is the foreground, and that which is "seen but not really seen" constitutes a kind of background. A third kind of foreground-background duality exists in Gestalt qualities. When we perceive a Gestalt, say a face, we also perceive—in the center and in the front of the face—a lot of details (the mouth, the nose, the eyes, etc.), but we do not attend to these details. Such nonattended details form still another kind of background, a background for which the Gestalt to which we attend is the foreground. There are various kinds of foreground-background distinctions. Is there a specific such distinction when we perceive determinates? When we perceive a color pattern, we primarily perceive (attend to) a Gestalt (the pattern) against a background of (non-attended) details (the color spots). But I think that, in fact, we also see that all the different color determinates have something in common, namely that each of them is, precisely, a color. There is, as a kind of background, a strictly identical something throughout the whole pattern: the color determinable. The color determinable is really perceived, but like a lot of other perceptual features it is perceived indirectly. In one way, however, the perception of determinables may differ from the other kinds of background-perceptions hinted at. What is periphery in one perception can become center in another perception. Similarly, what is backside in one perception can become front in another perception, and what is first a non-attended detail can later become a Gestalt of its own. But it seems very hard, if not impossible, to make the color determinable into a foreground and push its determinates into the background. This difference, though, has a natural explanation. Determinables can be said to be more abstract than their (concrete) determinates. In themselves, they seem to be more "thin" than their "thick" determinates. Necessarily, a determinable is less vivid than its determinates. This fact explains why it is so hard to attend directly to a determinable property of a perceived object. However, the background of a perception is as much in the perceptual act as the foreground is. A background can only—as background—be indirectly observed, but the fact that determinables are indirectly observable supplies all that is needed in order to refute the view that ontological determinables are not observable at all. 403

INGVAR JOHANSSON

7. Laws for Ontological Determinables The Armstrong of Universals and Scientific Realism accepted only lowest determinates as universals in re. Today, Armstrong accepts ontological determinables, but he does it hesitatingly, and he restricts the realm of ontological determinables to quantitative ones. I hope this paper shows that there really are determinable universals of various kinds, quantifiable as well as non-quantifiable. Some, but not all, conceptual determinables, reflect the existence of real spatiotemporally existing determinables. An ontological determinable is strictly the same in all its determinates. Therefore, there is in the world identity in difference. There is nothing fundamentally wrong with this old notion. There are not only ontological determinables, there are also interesting necessary relations connected with such determinables. We have stumbled upon three such laws or principles; The Law of Addition and Subtraction: Only determinates of the same determinable can be added and subtracted in a physically meaningful way. The Principle of Determinable Dependence: Some ontological determinables cannot possibly exist unless another determinable exists in the same spatiotemporal region (e.g., colors require shapes). The Principle of Determinate Exclusion: For some ontological determinables (e.g., color) it is true that two determinates cannot possibly exist in the same spatiotemporal region. If there are ontological determinables, then, there is of course no philosophical reason to try to replace the corresponding concepts with nominalistic constructions. There are not even pragmatic reasons. To speak of determinables is simpler and easier than to speak of classes of determinates. 8. Determinables as Properties, Relations, and Genera In the earlier sections I have restricted my discussion of the determinable-determinate relation to monadic properties. This is in conformity with tradition. In fact, however, I do not think that there are any good reasons for such a restriction. Obviously, just as there are determinate-determinable series between property concepts (e.g., "scarlet—>red—>color(ed)" 404

DETERMINABLES AS UNIVERSALS

and "square—>quadrilateral—>shape"), so there are such series between relation concepts (e.g., "a little longer than—Honger than in general—^length relation" and "much brighter than—brighter than in general intensity relation"). Relational concepts, just like property concepts, can be fitted into determinable-determinate trees. The series above can be prolonged. On the one hand we get the two series "scarlet—>red—>color—>property" and "square -^quadrilateral—^shape—^property", which both culminate in property, and on the other hand we get the two series "a little longer than—longer than in general—^length relation—^relation" and "much brighter than—brighter than in general—^intensity relation -^relation", which both culminate in relation. In this way, one can talk of a conceptual property-determinable as well as of a conceptual relation-determinable, both of which are determinables on a level above the determinables earlier discussed. If, as I think, relations cannot be reduced to properties, then the conceptual distinction between properties and relations reflects the existence of two ontological determinables. The existence of an ontological property-determinable and an ontological relation-determinable affects what is now and then called The Principle of Determinables. Here is Johnson's formulation of the principle: if any determinate adjective characterises a given substantive, then it is impossible that any other determinate under the same determinable should characterise the same substantive.30 A thing cannot simultaneously be both red and blue in one and the same part. Neither can it be both square and round or large and small. Such impossibilities are easily formulated by means of the determinabledeterminate distinction. Johnson, however, generalized too quickly. The Principle of Determinables is true for a lot of determinables (e.g., color, shape, and volume), but not for all determinables. If the concept property is a determinable which reflects an ontological determinable, we have an obvious counterexample which falsifies Johnson's principle. Surely, a thing can at one and the same time have several determinates of the property-determinable. Even more, things must have at least two such determinates since volume and shape are determinates of the property-determinable. In my view, Johnson's Principle of Determinables should be substituted by the The Principle of Determinate Exclusion of Section 7: for 405

INGVAR JOHANSSON

some ontological determinables it is true that two determinates cannot possibly exist in the same spatiotemporal region. The falsity of the principle of determinables is one of my reasons for identifying the determinable-determinate relation with what Searle has called the specifier relation, i.e., the conditions (a) and (b) of Section l. 31 My second reason for this identification is that there are similarities between the genus-species relation and the classical determinable-determinate relation which neither Johnson nor Searle have noted. Of course, the genus-species relation differs from determinable-determinate relations like color-scarlet and shape-square in the way stressed by Johnson and Searle. As Searle says: In short, a species is a conjunction of two logically independent properties— the genus and the differentia. But a determinate is not a conjunction of its determinable and some other property independent of the determinable. A determinate is, so to speak, an area marked off within a determinable without outside help.32

If the species man is defined as a rational animal and it is logically possible that there are non-animal rational beings (i.e., rationality is logically independent of animality), then the species man is "marked off' (to use Searle's expressions) from its genus animal "with outside help" from the concept rationality. There is, however, no problem in regarding the genusspecies relation as a determinable-determinate relation. All we have to do is to allow the determinable-determinate relation to be applicable to itself. There are then different determinates of the determinable determinabledeterminate relation. One such determinate is the genus-species relation, another such determinate is the determinable-determinaterelationfor properties, and a third such determinate is the determinable-determinate relation for relations. Just like all other determinates, these three determinates differ at the same time as they have something in common. In order to understand the similarity between ontological species and ontological determinates, one should note that when a species is an ontological universal, it is a complex unity of the genus-universal and the speciesuniversal. On the purely conceptual level the species-concept (e.g., man) can be described as a conjunction of two logically independent concepts, rationality and animality, but in re it is different. The genus-universal 0animality) and the species-universal (rationality) are spatiotemporally fused. They are merely aspects, not concrete parts of an actual particular 406

DETERMINABLES AS UNIVERSALS

of the species in question. The concepts man, rational, and animal, can be held (as now) spatially apart, but this is impossible when the corresponding universale exist together in re. A species is, rather, a Gestalt and as much a unity in its own right as its genus and its differentia specifica are. Therefore, the relation of a species to its genus is very similar to the relation between an ontological property-determinate and its ontological determinable. Kevin Mulligan has maintained that there are at least three types of non-conceptual relations of determination.33 Apart from genus-species trees (which relate to nouns) and Johnson's determinable-determinate trees (which relate to adjectives), there are also specification trees for actions (which relate to verbs). The intuitive appeal of this suggestion also underlines the fact that the relation between the classical determinable-determinate distinction and its own determinable is in need of investigation. Whether or not this relation should be called a specifier relation, or, as I have argued, a determinable-determinate relation, is of minor importance. Ingvar Johansson Umeä University, Sweden NOTES 1. I want to thank Kevin Mulligan, Geneva, for many discussions related to this paper. I have also benefited from comments given by members of the philosophy seminar in Umeä and by D. Bonevac, F. Correia, U. Meixner, and R. Poli. 2. W. E. Johnson, Logic (Cambridge: Cambridge University Press, vol. 1, 1921; vol. 2,1922; vol. 3,1924). 3. P. Brentano, Deskriptive Psychologie (Hamburg: Felix Meiner, 1982, eds. R. M. Chisholm and W. Baumgartner), Part 1, ch. II. 4. For a short overview of Brentano's philosophy and his concept of logical part, see K. Mulligan and B. Smith, "Franz Brentano on the Ontology of mind," Philosophy and Phenomenological Research XLV (1985), 327-44. 5. I have written 'colored)' in order to give associations both to the series 'being scarlet—being red—being colored! and the series 'scarlet is a kind of red which is a kind of color'·, a third corresponding series is 'having scarletness—having redness—having color'. The conceptual and grammatical differences between these series are unimportant in relation to the issue at hand. 6. The Italian philosopher Roberto Poli has written a paper in which Johnson's distinction, as well as its reception and later history, is presented. Therefore, there is no need 407

INGVAR JOHANSSON

for me to provide a history of the distinction. See Poli, "W. E. Johnson's Determinable-Determinate Opposition and His Theory of Abstraction,' in F. Coniglione, R. Poli and R. RoUinger (eds.), Abstraction and Idealization: Historical Studies (special issue of Poznan Studies in the Philosophy of the Sciences and the Humanities), forthcoming, 2000. 7. D. Armstrong, Universals and Scientific Realism (vols. I & Π) (Cambridge: Cambridge University Press, 1978), in particular vol. 2, pp. 111-13 and 116-20. 8. D. Armstrong, What Is a Law of Nature? (Cambridge: Cambridge University Press, 1983), ch. 7. 9. D. Armstrong, A World of States of Affairs (Cambridge: Cambridge University Press, 1997), chs. 13,16. 10. Ibid., p. 247. 11. D. Armstrong, Universals and Scientific Realism, vol. Π, p. 112. 12. A short introduction is Armstrong's Universals. An Opinionated Introduction (Boulder, CO: Westview Press, 1989), but the classic is his Universals and Scientific Realism. For criticism of trope nominalism see also H. Hochberg, "A Refutation of Moderate Nominalism," Australasian Journal of Philosophy 66 (1988), 188-207, and 'Troubles with Tropes," Philosophical Studies 67 (1992), 193-95. 13. Examples of such nominalisms are, in Armstrong's terminology, predicate nominalism, concept nominalism, class nominalism, and resemblance nominalisms. See Universals and Scientific Realism, vol. 1, Part Π. One may, however, speak about trope nominalism as a resemblance nominalism, too; see the next note. 14. I am confining the argument to what I take to be the most plausible kind of trope theory, "Resemblance classes of tropes," to use Armstrong's formulation; see his Universals. An Opinionated Introduction p. 17f. 15. I do not know who has invented this term, but I have taken it from Armstrong; see e.g., A World of States ofAffairs, p. 141. When, once upon a time, the term 'supervenience' was introduced by G. E. Moore and R. M. Hare, it was tied to non-reductive views; supervenience then entailed "increase in being." 16. This is not Armstrong's view. According to him, exact resemblance is the same as qualitative identity. See Universals and Scientific Realism, vol. II, p. 110. 17. K. Mulligan has entertained the idea that a trope theory of substances may require determinable tropes. See Mulligan, "Internal Relations," in B. Garrett & P. Menzies (eds.), Working Papers in Philosophy, 2, RSSS, Australian National University, Canberra, Proceedings of the 1992 Canberra metaphysics conference, pp. 1-22 (1993), esp. p. 13. Some trope theorists, e.g., Ivar Segelberg, have taken it for granted that there are both trope determinates and trope determinables, whereas others, e.g., John Bacon, explicitly have left the question open. See H. Hochberg (ed.), Ivar Segelberg. Three Essays in Phenomenology and Ontology, Stockholm: Thales (Library of Theoria), 1999, "Properties" ch. II. 1 and ch. IV.3 (about quality relations); and J. Bacon, Universals and Property Instances (Oxford: Blackwell, 199S), pp. 18 and 21. 18. See, e.g., J. J. Koenderink, Solid Shapes (Cambridge, MA: M.I.T. Press, 1990); see in particular chs. 6.8 ("The Local Shape Index") and 9.8 ("Shape Language"), but notice also 1.4 ("What Not to Expect"). 19. In topological terms, a square and a circle are homeomoiphic but not diffeomorphic. 20. The concept of level was not stressed by Johnson himself, but J. Searle, in an early non-ontological paper, highlighted this feature. See Searle, "Determinables and the Notion of Resemblance," Proceedings of the Aristotelian Society, Suppl. vol. 33 (1959), 141-58.

408

DETERMINABLES AS UNIVERSALS

21. Searle was not, in the paper mentioned in the preceding note, interested in the distinction between conceptual and ontological determinables which I am focusing attention on. Nonetheless, he made a distinction between absolute and non-absolute determinables which seems to run parallel to the distinction between ontological and conceptual determinables. He even said: "The notion of an absolute determinable is relevant to the traditional problem of categories: every predicate carries with it the notion of a kind or category of entities of which it can be sensibly denied or affirmed." Ibid., p. ISO. The same views can be found in his article "Determinables and Determinates" in Encyclopedia of Philosophy (New York: Macmillan, 1967). 22. Universals and Scientific Realism, vol. I, p. 57. 23. Armstrong's examples are Newton's Second Law, F = m a, and the Law of Gravitation F = (m, • m2)/ta. They have the form Q = f(P,N) and Q = f(P,N,M), respectively. 24. What Is a Law of Nature?, p. 9. 26. There are also interesting ontological things to be said about the distinction between additive and non-additive magnitudes; see my paper "Physical Addition," R. Poli and P. Simons (eds.), Formal Ontology (Amsterdam: Kluwer, 1996), pp. 277-88. 27. I have given a more thorough analysis in the paper "Pattern as an Ontological Category," N. Guarino (ed.). Formal Ontology in Information Systems (Amsterdam: I.O.S. Press, 1998), pp. 86-94. 28. I could equally well have said that the components are unities of spatial extension, color, and shape. 29. The problem is similar to, but not identical with, the "epistemological difficulty" discussed by Armstrong in Universals and Scientific Realism, vol. II, pp. 98-99. 30. W. E. Johnson, Logic I, p. 181. 31. See J. Searle, "Determinables and the Notion of Resemblance," Proceedings of the Aristotelian Society, Suppl. vol. 33 (1959), 141-58, esp. p. 145. Since I am primarily interested in ontological determinables, I have avoided the kind of problems which arise when one tries to state general conditions which are both sufficient and necessary for a determinable-determinate relation to hold between two concepts. To take an example, is red a determinate to any of the disjunctive concepts red or angry and colored or angry? Since I do not think there are any disjunctive ontological universals, I have no corresponding problem. Such problems, however, seem to be capable of being solved, as J. Searle (in the paper mentioned) and J. Woods ("On Species and Determinates," Nods 1, pp. 243-54) have shown. 32. Searle, Ibid., p. 143. 33. K. Mulligan, "Internal Relations," in B. Ganett & P. Menzies (eds.). Working Papers in Philosophy, 2, RSSS, Australian National University, Canberra, Proceedings of the 1992 Canberra metaphysics conference, pp. 1-22 (1993), esp. p. 8.

409

Appendix 3: (Talk given at the workshop "Reference Ontologies vs. Application Ontologies" at the German Conference on Artificial Intelligence, September 16, 2003.)

Ontologies and Concepts. Two Proposals. Ingvar Johansson 0 Introduction This workshop is meant to contain some interaction between computer science and philosophy. I am a philosopher, and I have got some space to try to make visible some views of my philosophical-ontological realism. According to my experience, I am sorry to say, people who work mainly with data assembled in interviews - be they anthropologists, sociologists, cultural-studies-people, or knowledge engineers - tend to become nonrealists with respect both to social reality and to mind-independent nature. Even more sadly, their bad philosophical ontology seems now and then to have bad repercussions on their pure scientific or practical work, too. In order to fuel discussions around these issues, I will now put forward a little realist pamphlet. 1 Only Fictions are Fictions Once upon a time there was a knowledge engineer called KE. He was given two tasks: 1. to construct a taxonomy of all the existing mammals on earth, and 2. to construct another taxonomy of all the humanoids in Star Wars. How did he proceed? In the case of the earthly mammals, KE started to read relevant zoological literature as well as to interview taxonomically interested zoologists, i.e., the domain experts at hand. Since KE was a very good knowledge engineer, he very quickly understood the contemporary scientific mammal classification and turned it into a good computer system. In the other case, KE started immediately to see the Star War movies, read about them, and to check on internet what the domain experts now at hand, the real Star War fans, had to say. Immediately, he found on

411

internet a list of the humanoids that occur in Star Wars. The Abyssin species was said to be "a swarthy-faced humanoid with a single slitpupilled eye, a thirsty representative of the cyclopean species;" the Aqualish species was "a bellicose humanoid species with some superficial properties both aquatic and arachnid; they have hair-lined faces with large, glassy eyes and mouths dominated with inwardly-pointing tusks;" and so on. After some work, as expected, KE had classified the humanoids into classes and families and put his taxonomy on internet. So far, so good. And no essential difference between the two cases. However, after a while KE found one indeterminacy in each taxonomy. In both cases, he went through all his material once again, but both the indeterminacies remained. What to do? His reputation as a good knowledge engineer seemed to be at stake. With respect to the fictional humanoids of Star Wars, KE's problem came to end when he happened to read Roman Ingarden's The Literary Work of Art (Evanston, 111.: Northwestern University Press, 1973). He then realized that fictional entities necessarily have some "loci of indeterminacy." What is not described by the creators of a fictional entity is simply indeterminate. And that was exactly the case with the problematic humanoid. Nothing could and nothing should here be done by a good knowledge engineer. With respect to the mammals, KE found quite another solution. He asked a zoologist to make some completely new observations related to the indeterminacy at hand. Since the mammals to be classified were really existing species, KE realized that there are in principle an infinite number of possible new observations that can be made. As things turned out, the new observations did in fact solve his indeterminacy problem. KE was quite happy that he had not tried to use the same methodology to solve both his problems. When a knowledge engineer meets domain experts that are experts in relation to some areas or aspects of the real world, he has, whether he is aware of it or not, a choice. Either: he behaves as if what the experts tell him is all there is to say, that is, he treats them as authors of a story about fictional objects that very well may contain indeterminacies and even inconsistencies. Or: he takes a realist perspective and keeps an eye to the possibility of interaction with the world itself; if the domain experts contradict each other, then he asks them to reconsider their views.

412

First proposal: Fictions apart, the default position for every ontology creator should be the realist position; concepts are primarily links to the world. In this respect, there is no difference between application ontologies and reference ontologies. 2 Concepts are Tools When one is using glasses, microscopes, or telescopes, one sees through them in order to look at something in the world. But something similar is true also of the tactile sense in relation to ordinary tools. When one is real good at using a tool, it disappears from the foreground of consciousness, i.e., one is no longer intentionally directly directed at it. Instead, one is directed at that in the world which the tool affects. For instance, when I write with a pen, I am hardly aware of the pen, only of what I am writing. I do not feel my fingers holding and touching the pen; if I feel anything at all, it is rather the pen pressing against the paper I am writing on. The same phenomenon can be observed even in relation to machines. When I am driving my car, normally, I do not feel my fingers pressing the steering wheels or my feet pressing the pedals or the floor. I am only directed at the road and the traffic. Speech acts are tools for many things. They can be used to convey information, give orders, have discussions, make emotional outbursts, etc. In spite of the fact that speech acts do not have the same robust material character as ordinary tools have, they have several features in common with such tools. Think of conveying information in a language in which one is both making and understanding speech acts spontaneously. In the assertive speech acts at hand, one is directly directed at the world, not at concepts, statements, or propositions; nor, by the way, at grammar and dialects. But, of course, we can also be directed at concepts and conceptual strings. So we are when we are looking in a dictionary, in a phrase book, or are browsing on WordNet. When one is not using glasses, pens, or cars, one can look at them and investigate them as objects of their own in the world. One can then try to find out their physical properties and internal structures. In the world of practice, normally, one observes "tools as objects" only when the tools in question are no longer functioning properly and, therefore, are in need of repairing. Something similar holds true of concepts. When our speech acts 413

function well, we do not bother about the concepts, but when they don't, as when we are learning a new language, we often direct ourselves directly at concepts in dictionaries. All ontologies make use of concepts, but only some ontologies (e.g., WordNet) are also ontologies for concepts. When, by means of concepts, we are directed at something in the world, then the concepts in question may very well both (i) select aspects, (ii) select granularity levels, and (iii) create boundaries, without thereby (iv) creating aspects, granularity levels, and what is bounded. Think of the concept "mountain." It selects a geographical aspect of the world, it focuses at a certain granularity level, and it is connected with a vague conventionally drawn boundary line that tells where mountains end. Nonetheless, it seems preposterous to say that the concept "mountain" creates the mountains. Ontologists who are reductive naturalists (and claim that only the fundamental entities of physics do really exist) or epistemological Kantians (and claim that we cannot lay claim to know anything about a mind-independent world) cannot be taken seriously until they have discussed the view of concepts just presented. Second proposal: Fictions apart, concepts should be looked upon as tools which can be used to direct us towards something that exists independently of these concepts. In this respect, there is no difference between application ontologies and reference ontologies.

414

REPRINT PHILOSOPHY Modern Classics of Analytical Philosophy edited by

Rafael Httntelmann · Erwin Tegtmeier · Küthe Trettin

Gustav Bergmann

г-

'S It

и

8

is —

·*

§·.§

I

ε-

-

--Si — О —' >-. Sj ^ « e s j= < "8 i С й '5 С = л Ш ι_ 3 · o.'3 5 a 3 § ai β ε

Collected Works Three Volumes Edlted and Introduced by Erwin Tegtmeier

Gustav Bergmann

Collected Works Vol. Ill

Vol. I:

Realism

Vol. II:

A Critique of Brentano and Meinong

Vol. Ill:

Selected Papers I

pp. 350, onto* vertag 2003 Hardcover С 99,00

Selected Papers II

pp. 371, onto* vertag 2003 Hardcover € 99,00

Realism. A Critique of Brentano and Meinong pp. 467, ontos vertag 2004 Hardcover € 99,00

fäniüü

i

vertag ?

Three Volumes In Subscription: pp. 1188, €201,00

Bergmann was bom In Vienna (Austria) In 1906. After finishing the Gymnasium he registered at the University of Vienna. Before he took a Ph.D. In mathematics wttti a minor In philosophy In 1928, he had already been Invited along with his Gymnasium classmate Kurt G6del to Join the Vienna Circle, where he was especially Influenced by Schlick, Walsmann and Carnap. In 1929-30 Bergmann taught mathematics at a Realschule In Vienna and In the following year he went to Berlin to work as assistant of Einstein together with W. Mayer his dissertation director. Discouraged by the discrimination against Jews at German and Austrian universities Bergmann returned to Vienna to study law. He took a JD and went Into a firm of corporation lawyers. When Nazi Germany annexed Austria In 1938 Bergmann emigrated to the United States with financial assistance from Circle Member Otto Neurath. In 1939 he obtained an appointment at the University of Iowa as an assistant to the psychologist Kurt Lewln. He was employed to develop a mathematical rep r* Mittat Ion of Lewin's psychological field theory. In 1940 Bergmann became assistant professor at the Department of Philosophy, and In 1950 tail professor of philosophy and psychology. During the 1960s and 1970s Bergmann had a major Impact on contemporary philosophy and contemporary Issues In the philosophy and methodology of psychology, attracting brilliant students who went on to teach In philosophy and psychology departments of leading universities of the United States. Bergmann and his philosophy students and followers were sometimes referred to as "The Iowa School" or "The Iowa Realists". It brought national and International status to a small philosophy department of a middle-sized mldwestem university. From 1967-68 he served as president of the American Philosophical Association (Western Division) and In 1972 he was awarded the first named professorship In the College of Liberal Arts at the University of Iowa as Carver Professor. Bergmann retired In 1974 and died In 1987.