Representation And Objects of Thought in Medieval Philosophy (Ashgate Studies in Medieval Philosophy)

Citation preview

REPRESENTATION AND OBJECTS OF THOUGHT IN MEDIEVAL PHILOSOPHY The notions of mental representation and intentionality are central to contemporary philosophy of mind and it is usually assumed that these notions, if not originated, at least were made essential to the philosophy of mind by Descartes in the seventeenth century. The authors in this book challenge this assumption and show that the history of these ideas can be traced back to the medieval period. In bringing out the contrasts and similarities between early modern and medieval discussions of mental representation the authors conclude that there is no clear dividing line between western late medieval and early modern philosophy; that they in fact represent one continuous tradition in the philosophy of mind.

ASHGATE STUDIES IN MEDIEVAL PHILOSOPHY Series Editors John Marenbon, Trinity College, Cambridge, UK Scott MacDonald, Cornell University, USA Christopher J. Martin, University of Auckland, New Zealand Simo Knuuttila, Academy of Finland and the University of Helsinki, Finland

The study of medieval philosophy is flourishing as never before. Historically precise and philosophically informed research is opening up this large but still relatively unknown part of philosophy’s past, revealing – in many cases for the first time – the nature of medieval thinkers’ arguments and the significance of their philosophical achievements. Ashgate Studies in Medieval Philosophy presents some of the best of this new work, both from established figures and younger scholars. Chronologically, the series stretches from c.600 to c.1500 and forward to the scholastic philosophers of sixteenth and early seventeenth century Spain and Portugal. The series encompasses both the Western Latin tradition, and the Byzantine, Jewish and Islamic traditions. Authors all share a commitment both to historical accuracy and to careful analysis of arguments of a kind which makes them comprehensible to modern readers, especially those with philosophical interests. Other titles in the series: Ockham on Concepts Claude Panaccio ISBN 0 7546 3228 8 Medieval Modal Systems Problems and Concepts Paul Thom ISBN 0 7546 0833 6 Theology at Paris, 1316–1345 Peter Auriol and the Problem of Divine Foreknowledge and Future Contingents Chris Schabel ISBN 0 7546 0204 4

Representation and Objects of Thought in Medieval Philosophy

Edited by HENRIK LAGERLUND University of Cambridge, UK

© Henrik Lagerlund 2007 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise without the prior permission of the publisher. Henrik Lagerlund has asserted his moral right under the Copyright, Designs and Patents Act, 1988, to be identified as the editor of this work. Published by Ashgate Publishing Limited Gower House Croft Road Aldershot Hampshire GU11 3HR England

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

Ashgate website: http://www.ashgate.com British Library Cataloguing in Publication Data Representation and objects of thought in medieval philosophy. – (Ashgate studies in medieval philosophy) 1. Philosophy, Medieval 2. Representation (Philosophy) 3. Object (Philosophy) I. Lagerlund, Henrik 189 Library of Congress Cataloging-in-Publication Data Representation and objects of thought in medieval philosophy / edited by Henrik Lagerlund. p. cm.—(Ashgate studies in medieval philosophy) Includes index. ISBN 0-7546-5126-6 (hardcover : alk. paper) 1. Philosophy, Medieval. 2. Philosophy of mind—History—To 1500. 3. Mental representation—History—To 1500. I. Lagerlund, Henrik. II. Series. B738.S68R47 2006 121'.4—dc22 2006008832 ISBN 978-0-7546-5126-0

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

Contents List of Contributors Foreword I

Introduction Henrik Lagerlund

II

The Terminological and Conceptual Roots of Representation in the Soul in Late Ancient and Medieval Philosophy Henrik Lagerlund

vii ix 1

11

III

Abstract Truth in Thomas Aquinas Robert Pasnau

33

IV

Representation in Scholastic Epistemology Martin Tweedale

63

V

Rethinking Representation in the Middle Ages: A Vade-Mecum to Medieval Theories of Mental Representation Peter King

81

VI

William Ockham and Mental Language Mikko Yrjönsuuri

101

VII

The Matter of Thought Calvin G. Normore

117

VIII

Objective Being in Descartes: That Which We Know or That By Which We Know? Deborah Brown

Index of Names

135

155

This page intentionally left blank

List of Contributors Deborah Brown is Lecturer in Philosophy at the University of Queensland. She has published several articles on medieval and Cartesian philosophy of mind and she has recently finished a monograph on Descartes’s theory of the passions. Peter King is Professor of Philosophy at the University of Toronto. He works primarily on medieval philosophy and has published numerous books and articles. Henrik Lagerlund is Associate Professor of Philosophy at Uppsala University and Research Associate at the University of Cambridge. He has published several books and articles on medieval philosophy. Calvin G. Normore is Professor of Philosophy at the University of California, Los Angeles. He works on medieval and early modern philosophy and has published numerous articles. Robert Pasnau is Associate Professor of Philosophy at the University of Colorado at Bolder. He has published numerous books and articles on medieval philosophy. Martin Tweedale is Emeritus Professor of Philosophy at the University of Alberta. He works primarily on medieval philosophy and has published several books and articles Mikko Yrjönsuuri is a Senior Assistant at the University of Jyväskylä and a Researcher at the Academy of Finland. He has published extensively on medieval and early modern philosophy.

This page intentionally left blank

Foreword Not more than ten years ago it would have been very odd to refer to ‘pre-modern’ philosophy of mind. The view that philosophy of mind began in the seventeenth century foremost with René Descartes was and perhaps still is the common opinion, but this view of the history of philosophy is on its way to being completely revised. Today the philosophy of mind of the Middle Ages is one of the most fruitful areas of research. This book brings together some of the leading scholars in the English speaking world in this field on the common topic of representation and objects of thought. It becomes very clear in their discussions that the sophistication of the medieval discussions of cognition and mental representation equals and in some respects even surpass contemporary philosophical discussions of the same issues. By digging deep into the treasures of medieval philosophy we are likely to learn a lot of new philosophy. Henrik Lagerlund Cambridge

This page intentionally left blank

Chapter I

Introduction Henrik Lagerlund

Every science and all areas of philosophy develop or change through the creation and change of their concepts and their terminology. A specific science does not get established until it has developed a proper terminology and conceptual framework for itself. These concepts are then gradually formed as the science develops and sometimes they are dropped and new are created. There are lots to be said in favor of the notion that the formation of conceptual frameworks is what science and philosophy are all about. If we adopt such a view of the development of science and philosophy, interesting perspectives open up for the historian of science and philosophy. The task of the historian becomes to trace the creation and change of concepts that are the basis of the science and philosophy of a certain time. But it also, I think, presents the historian with an important task, which has implications for contemporary science and philosophy as well, namely it gives the historian an opportunity and a possibility to clarify concepts used in the science and philosophy of today, which might help considerably in enhancing our understanding of the things we are now doing. Historical research of concept formation is probably more useful in philosophy than in science, since the concepts of philosophy are for the most part less exact and developed than the concepts of science. The formation of concepts seems also to be slightly different in science than in philosophy, but on the other hand philosophical conceptions of the world are often at a more fundamental level than those of the exact sciences. In philosophy, it seems to me, however, that a historical perspective of the basic concepts of some specific area of philosophy might help clarify a number of contemporary debates. One such area is philosophical psychology or philosophy of mind. This is an area of philosophy that for a long time has been dominated and even at times paralyzed by its history. René Descartes is often claimed to have been the father of modern philosophy and particularly of the modern conception of the mind. Contemporary philosophy of mind has never been able to step out from under the shadow of Descartes. However, several recent studies of Descartes and of the later Middle Ages have first of all radically reread Descartes but also re-evaluated his originality in light of the medieval background. (See, for example, Alanen 2003.) It seems that the more in depth we get to know the philosophical psychology of the later Middle Ages the more complex the development of the modern conception of mind becomes. By

2

Representation and Objects of Thought in Medieval Philosophy

rethinking the background of our contemporary philosophy of mind, the history of philosophy might help rethink philosophy of mind. If we take to heart the view of the development of philosophical conceptions expressed above, there are many things that speak for the importance of the twelfth and thirteenth centuries in the history of philosophy and science. During these centuries Aristotle, Avicenna and Averroes along with a number of other Greek and Arabic thinkers were translated into Latin and mixed with the already existing Western Latin tradition. The vast amount of new texts that pored over the Latin thinkers of the Middle Ages created conflicts, of course, but most of all it created new concepts and new ways of conceptualizing the world. What we know as modern thought developed out of this period. This seems to be true of philosophical psychology in particular. Already in 1976 did the renaissance scholar F.E. Cranz argue for the importance of the medieval De anima commentaries for the understanding of the development of modern thinking (Cranz 1976). He argued that in Aristotle there is no distinction between sense experience and the thing that is sensed, or between thinking and what is thought of, that is, no clear distinction between subject and object. He argued further that the whole Ancient tradition followed Aristotle in this sense and that it is not until the later parts of the Middle Ages and in the De anima commentaries that a modern way of viewing the world appears. At the same time as these commentators on one level followed Aristotle they broke fundamentally with the Aristotelian worldview. In the De anima commentaries an old terminology was used, but it was developed to express the dualistic worldview. Some of these terms were ‘species’, ‘intentio’, ‘spirituale’, ‘similitudo’, ‘imago’ and ‘repraesentatio’. Later research has shown that the Ancient tradition is more complex than what Cranz assumed, but I, however, think he was on to something when he claimed that the terminology and ultimately the conception of psychology got radically reworked in the later Middle Ages. The terms he mentioned are all part of the standard terminology of modern philosophical psychology. We should perhaps add ‘idea’ and ‘mens’ (‘mind’) to his list, but the main thing is that all these terms, the ones I added included, were introduced or redefined in the twelfth or the thirteenth centuries. Which means that by the end of the thirteenth and the beginning of the fourteenth centuries the terminology and the concepts on which so-called modern philosophical psychology or philosophy of mind are based were already in place. These concepts were then developed and remodelled to create different theories of thought and mind, but the foundation was already there. Some of these concepts were simply taken over from Ancient or Arabic philosophy by translation into Latin, but some were created by the translators in translating some Ancient or Arabic text, and yet others were taken from the existing tradition and applied in new ways. One of these almost magical words in contemporary philosophy is ‘intentionality’. The Austrian philosopher Franz Brentano introduced it into the contemporary debate in the late nineteenth century. He used it to designate a property he thought would demarcate the ‘mental’ from the ‘non mental’. It was the property of ‘in-existence’, that is, the mental could, crudely put, have things existing in it that are otherwise

Introduction

3

external to it. In more contemporary jargon it means that our thoughts have the property of being about things other than itself. Brentano, however, states explicitly that he is only borrowing or re-actualizing a term used frequently by scholastic philosophers. The Latin term he was thinking of was ‘intentio’. It is in turn a translation of the Arabic terms ‘ma’na’ and ‘ma’qul’. These terms were used by, for example, Al-farabi and Avicenna. In scholastic philosophy, a distinction was usually made between a first intention, which is used to designate the intellects relation to an object immediately before it that is external to it, and a second intention, which is used in the same way but where the object is internal, that is, a first intention. In this respect late medieval philosophy simply adopted a terminology already existing in Arabic philosophy in much the same way contemporary philosophy has done. (See also the discussion of Augustine’s use of this notion in Caston 2001.) Another term that is of vital importance for philosophical psychology as it has developed from Descartes onward is ‘representation’ or ‘mental representation’. It was also introduced into medieval philosophy in the twelfth century but not merely as a translation of an Arabic term. It was created or forged by a process of translation and philosophical discussion. The Latin term is, of course, the noun ‘repraesentatio’ or the verb ‘repraesentare’, and as is shown in Lagerlund’s chapter in this book the translators of Avicenna’s al-Shifâ’ took a whole group of related Arabic terms and translated them with either the Latin noun or verb for representation. Avicenna’s theory of the soul thus became in the Latin tradition a theory that relies heavily on representations in the soul and on a notion that external sensations are being represented to the intellect through the internal sense. This notion was then developed by scholastic thinkers influenced by Avicenna into a notion of mental representation, that is, a notion of concepts and meaning bearing signs in the intellect. The concept of internal representation and mental representation was thus created in the twelfth and thirteenth centuries. It also has a very interesting history in the Middle Ages as several papers in this book will show. A third term that has been of enormous importance to modern philosophy is ‘idea’. What would the philosophy of John Locke be without it? Some attempts have been made to write the history of the term, but lots still remain to be done. (See, for example, Ariew 1999.) Obviously, the Latin term ‘idea’ was used in relation to the Platonic ideas and through Augustine this became associated with ideas or archetypes in the divine mind, but the same term was also used in theories about human cognition long before Descartes and Locke. In the early thirteenth century, Richard Rufus of Cornwall claimed that the Platonic ideas are the objects of human cognition. Rufus is a very unorthodox Platonist and explains that the ideas or exemplars in Plato are like Aristotelian universals inhering in external object. All ideas are innate and there is no other connection between the ideas inhering in the external objects and the ones in our intellect except through God. Even though he talks about the process of abstraction in intellectual cognition, abstraction is just a triggering mechanism that actualizes our innate ideas. (See Wood 1997 and Normore in this book for a more careful discussion about Rufus use

4

Representation and Objects of Thought in Medieval Philosophy

of ideas.) The important thing for my argument here is, of course, that in Rufus we, perhaps for the first time, finds a thinker using the terminology of ‘idea’ to talk about objects in the human intellect. Even though medieval thinkers tended to reserve the term ‘idea’ for the exemplars in the mind of God there are others that like Rufus used the term in relation to their discussion of human cognition. Durandus of St Pourcain and Peter of Ailly in the fourteenth century are such thinkers, and it seems to be through them that the usage spread into the sixteenth century (see Lagerlund 2003 and Normore’s chapter in this book). The history of the term ‘idea’ in the Middle Ages has not been properly studied yet, and medieval philosophical psychology has only begun to be studied, but already now we see that the relation of the so-called medieval and early modern periods are much more complex than previously thought. This book is primarily about one of the notions mentioned above, namely ‘representation’ or even more narrow ‘mental representation’. It tries to bring out the complexity and sophistication of the medieval and early modern discussions of this notion by discussing both the metaphysical and the epistemological problems related to the notion of representation in the soul or mind. There are, however, a number of confusions about the term ‘representation’ as it is used by historians and philosophers. Let me therefore try to sort out some of this confusion and also to bring the topic of this book more into context. It is often assumed at least by many historians of philosophy that as soon as one starts talking about representational ideas or representation in the soul then one has committed oneself to some position in epistemology. This assumption is foremost made in relation to Descartes – although less nowadays than before. Anthony Kenny, Richard Rorty and Barry Stroud are good examples. They all interpret Descartes as unable to draw a distinction between epistemology and philosophy of mind. In his book on Aquinas’s philosophy of mind, Anthony Kenny argues that between Descartes and Wittgenstein a proper analytic philosophy of mind did not really exist. His claim is that Descartes muddled the analytic discussion of the operations of the human mind with a discussion of epistemology. It took Wittgenstein and to some extent Gilbert Ryle to free philosophy of mind from epistemology. He writes in Aquinas on Mind: Among philosophers in the Anglo-American tradition there has grown up, in the years since the Second World War, a branch of philosophy, a philosophical discipline, which is sometimes called philosophical psychology and sometimes philosophy of mind. The existence of the subject as a separate discipline in recent times was due primarily to the influence of Wittgenstein and secondarily to that of Ryle. In other philosophical traditions since the Renaissance it is not so easy to identify, as a specific area of philosophical study, the field which bears the name ‘philosophy of mind’. This is because since the time of Descartes the philosophical study of the operations of the human mind has taken place in the context of epistemology. Epistemology, as I have said, is the discipline which is concerned above all with the justification of our cognition, the vindication of claims to knowledge, the quest for reliable methods of achieving truth. Epistemology, as contrasted

Introduction

5

with philosophy of mind, is a normative rather than descriptive or analytic branch of philosophy (18–19).

According to Kenny, medieval philosophy included what he calls an analytic or descriptive philosophical psychology. He is primarily thinking about the tradition of commentaries on Aristotle’s De anima. I am not going to take a stand on Kenny’s historical claims, but I would like to stress the distinction he puts forward to clarify some often repeated misconceptions about representation. I, however, would like to rephrase what he calls the normative in terms of the epistemological problem of representation and the descriptive in terms of the metaphysical (or ontological) problem of representation. The epistemological problems of representation are well known and it has motivated whole philosophical traditions in modern philosophy. Over the years philosophers have taken different stands on these problems, for example, if one is a realist one can be a direct realist or a representational realist and then there are also a variety of idealist positions. The literature on Descartes is a good example of how difficult and problematic the discussions of the epistemological problem can become. Barry Stroud has in The Significance of Philosophical Scepticism expressed in a very clear manner the traditional reading of Descartes. He writes: We are confined at best to what Descartes calls ‘ideas’ of things around us, representations of things or states of affairs which, for all we know, might or might not have something corresponding to them in reality. We are in a sense imprisoned within those representations, at least with respect to our knowledge. Any attempt to go beyond them to try to tell whether the world really is as they represent it to be can yield only more representations, … […] This can seem to leave us in the position of finding a barrier between ourselves and the world around us. There would then be a veil of sensory experience or sensory objects which we could not penetrate but which would be no reliable guide to the world beyond the veil. If we were in such a position, I think it is quite clear that we could not know what is going on beyond the veil. […] I have described Descartes’s sceptical conclusion as implying that we are permanently sealed off from a world we can never reach (32–3).

On his reading Descartes is a representational realist while others have argued for other readings (see Brown’s chapter in this book). It does not even occur to Stroud that the distinction I drew above might also apply to Descartes. Another philosopher that is even more reluctant to think that such a distinction might apply to Descartes is Rorty. In Philosophy and the Mirror of Nature, he writes: In Descartes’s conception – the one which became the basis for ‘modern’ epistemology – it is representations which are in the ‘mind’. The inner Eye surveys these representations hoping to find some mark which will testify to their fidelity (45).

Rorty has his own agenda and reading Descartes as he does fits this agenda perfectly. I, however, think that the picture is not as simple as the one Rorty presents and the starting point of modern philosophy (if one can locate at all a starting point) is not as confused as he wants us to believe. By ignoring the distinction between the

6

Representation and Objects of Thought in Medieval Philosophy

epistemological and the metaphysical (ontological) problems of representations he largely misunderstands early modern philosophy. An aim of the present book is to considerably complicate the origins of modern philosophy. In late medieval philosophical psychology, the distinction between the epistemological and metaphysical side of internal representation is explicitly stated and all of the major thinkers seem to have been aware of it. In Summa Theologiae, Thomas Aquinas for example writes: The intelligible species is to the intellect what the sensible species is to the senses. But the sensible species is not what is perceived, but rather that by which the sense perceives. Therefore the intelligible species is not what is actually understood, but that by which the intellect understands (I, q. 85, art. 2, resp.).

Following Aquinas one must uphold a distinction between (i) the representation being what is understood and (ii) the representation being that by which something is understood. Taking a representation in the soul in the sense of (i) entails a commitment to representational realism (if one is a realist that is) on the epistemological problem while (ii) does not. (ii) might instead be read as a commitment to direct realism (if one is a realist that is), but it does not have to – it could be taken as a purely descriptive statement about representations. It is important to keep these distinctions in mind when approaching the history of this particular notion, and as will be clear from this book medieval and early modern philosophical psychology has been sensitive to them all along. The first chapter in this volume by Henrik Lagerlund tries to trace the history of the term ‘representation’ in Latin philosophy from Ancient to late Scholastic philosophy. He argues that it is not until the Latin translation of Avicenna in the twelfth century that this term was associated with the internal senses or the human soul. He goes through a number of uses of the term by late Ancient and early medieval thinkers from Quintilian to Anselm and Abelard, but argues that none of them are systematically applied to the soul and they seem to have had no influence on the development of a notion of internal representation or representation in the soul. In the Latin translation of the part of Avicenna’s Shifâ that came to be known as his De anima, the terminology of ‘representation’ is, however, used quite frequently. The translators translate a number of different terms in the Arabic with the Latin term for representation. Lagerlund argues on the basis of this that they are in this way forming the notion of internal representation. After Avicenna is translated the whole Scholastic tradition takes on this use of representation. Lagerlund also argues that there is no attempt in the Latin text of Avicenna’s De anima to use this terminology in relation to the intellect or the mind. This kind of extension of Avicenna’s terminology can, however, be found in the thirteenth century; for example in the works of Thomas Aquinas. Our contemporary notion of mental representation meaning a mental sign or idea standing in for its object in our mind without relying on a notion of resemblance is first developed in the early fourteenth century, according to Lagerlund.

Introduction

7

In his chapter ‘Abstract Truth in Thomas Aquinas’, Robert Pasnau asks what the abstracted objects of the intellect are, according to Aquinas, and, given that all things sensed are particular, how can these abstracted universal objects yield true beliefs about reality. The proper objects of the intellect are the natures or quiddities of things, argues Aquinas, but these natures only exist in particular things in reality external to the mind and never by themselves. In a sense, the objects of the intellect hence do not exist, Pasnau explains. They need to be abstracted or derived by the intellect, but the method of abstraction developed by Aquinas is not, according to Pasnau, suited to give an account of how these natures are structured. On Pasnau’s reading of Aquinas, he is a nominalist in the contemporary sense that everything existing is particular. He then needs to give an account of how there can be abstract truths in such a reality, that is, he needs to give an account of how reality and concepts are related. He gives this account in terms of formal agreement between the object in reality and the object in the intellect. In his chapter, Pasnau elaborates this theory in detail. In his contribution to this book ‘Representation in Scholastic Epistemology’, Martin Tweedale studies Ancient and the Scholastic approaches to intentional existence. He traces the discussion of this issue in Aristotle, Alexander of Aphrodisias, Themistius, Avicenna and Averroes, and explains how their discussions were taken up in the thirteenth and fourteenth centuries by Albert the Great and Aquinas. The species theory of cognition defended foremost by Aquinas was heavily debated in the late thirteenth century and the peculiarities of that theory eventually lead John Duns Scotus to introduce the distinction between subjective and objective being or existence. The latter parts of Tweedale’s chapter are devoted to a treatment of William of Alnwick’s attempts to get around some of the problems introduced by Scotus’s theory. The objects of thought or knowledge that Scotus had called ens objectivum are themselves things with some kind of separate existence. Such entities lead immediately in the early fourteenth century to skeptical problems about what it is that we know. Alnwick tries very hard to explain what these objects of thought are and whether they have a separate existence or not. Peter King presents an overview of medieval discussions of mental representation in his chapter ‘Rethinking Representation in the Middle Ages: A Vade-Mecum to Medieval Theories of Mental Representation’. He lists four ways to account for mental representation that he has found in the Middle Ages. These are: (R1) The mental representation and the represented item have the same form. (R2) The mental representation resembles, or is a likeness, of the represented item. (R3) The mental representation is caused by the represented item. (R4) The mental representation signifies the represented item. He traces these four to different medieval thinkers and ends up in Ockham’s critical stance towards representation in general.

8

Representation and Objects of Thought in Medieval Philosophy

Mikko Yrjönsuuri argues in his chapter ‘William Ockham and Mental Language’ that Ockham, despite what some has recently argued, puts forward a theory of mental language according to which it is an ideal language. Yrjönsuuri argues for his view by comparing Ockham’s project with Chomsky’s Cartesian linguistics. His conclusion is that although Ockham never quite finished his project, it is very different from the seventeenth century linguistic theories, particularly since Ockham’s mental language is supposed to mirror the metaphysical structure of the world while the later linguistic theories thinks that it is the mind’s structure that it mirrored by the grammar. In his contribution ‘The Matter of Thought’, Calvin Normore wants to relate the medieval and early modern discussions of the immateriality of the intellect with discussions about the ontological status of objects of thought. He starts by outlining two contemporary and medieval theories of what it means to think about something, namely the relational and the adverbial theories. Having done this he goes on to relate the medieval discussions about these theories to Aristotle’s argument that the mind cannot be of a certain sort, since it would then not be able to think about all possible things. This assertion was one of the main arguments for the immateriality of the intellect, and it is this argument, according to Normore, that shaped much of the discussion about theories of mental content. On the traditional reading of Descartes, he is presented as a representational realist, which means rather paradoxically that he introduced just after he had secured a firm foundation for knowledge a ‘veil of ideas’ to hide it behind. In her chapter ‘Objective Being in Descartes: That Which We Know or That By Which We Know?’, Deborah Brown is firmly rejecting this traditional reading of Descartes. The main question she asks is whether Descartes’s use of objective being or existence entails a representational realism. She starts by examining the medieval discussions of objective being before entering into a treatment of Descartes’s own usage of this notion. Her conclusion is simply that there is nothing in this notion or in Descartes’s use of it that entails a realism which is representational. Descartes develops a representational theory of mind, but not a representational realism, according to Brown. Bibliography Alanen, L. (2003), Descartes’s Concept of Mind, Cambridge, MA: Harvard University Press. Aquinas, Thomas, Summa Theologiae, ed. P. Caramello, Rome: Maretti, 1950–53. Ariew, R. (1999), Descartes and the Last Scholastics, Itaca: Cornell University Press. Caston, V. (2001), ‘Connecting Traditions: Augustine and the Greeks on Intentionality’, in D. Perler (ed.), Ancient and Medieval Theories of Intentionality, Leiden: Brill, 23–48.

Introduction

9

Cranz, F.E. (1976), ‘The Renaissance Reading of the De anima’, in XVIe Colloque International de Tours: Platon et Aristote a la Renaissance, Paris: Libraire Philosophique J. Vrin. Kenny, A. (1993), Aquinas on Mind, London: Routledge. Lagerlund, H. (2003), ‘Representations, Concepts and Words: Peter of Ailly on Semantics and Psychology’, in Proceedings of the Society for Medieval Logic and Metaphysics, Vol. 3, 15–36 Rorty, R. (1979), Philosophy and the Mirror of Nature, Princeton: Princeton University Press. Stroud, B. (1984), The Significance of Philosophical Scepticism, Oxford: Clarendon Press. Wood, R. (1997), ‘Richard Rufus and the Classical Tradition: A Medieval Defence of Plato’, in L.G. Benakis (ed.), Néoplatonisme et Philosophie Médiévale, Louvain: Brepols, 229–51.

This page intentionally left blank

Chapter II

The Terminological and Conceptual Roots of Representation in the Soul in Late Ancient and Medieval Philosophy* Henrik Lagerlund

1 Introduction The concept of mental representation lies at the heart of contemporary philosophy of mind. It is often claimed that the problem of intentionality is a problem of mental representation, since mental states have content due to their representational nature.1 Ever since Franz Brentano introduced the concept of intentionality as the mark of the mental it has been known that it was a scholastic concept that he revived. Later research has shown that the scholastic concept ‘intentio’ derives from Arabic philosophy and is a translation of the Arabic words ‘ma’na’ and ‘ma’qul’ as used by Al-farabi and Avicenna. They in turn claim to have translated Aristotle’s Greek word ‘noêma’ as he uses it in the beginning of the De interpretatione. For Al-farabi ‘intentio’ is that which is immediately before the mind, whether the object of the intention is outside the mind (in which case it is a first intention) or itself an intention (in which case it is a second intention). This distinction became absolutely central to scholastic philosophy. Note that for them and later scholastic philosophy it is primarily concepts that have ‘intentio’.2 Medieval discussions of intentionality have lately been given extensive treatments by philosophers,3 but have drawn attention from scholars for quite some time.4 The related concept of representation has not received the same attention and * Originally, I read an early version of this chapter as a paper presented at the SIEPM world congress in Porto 2002, but I have also presented drafts at ‘La Sapienza’ in Rome, in Melbourne and Brisbane, and of course at the research seminar in Uppsala. I am greatful to all the participants at these presentations, but foremost I would like to thank Alfonso Maierù for his helpful suggestions. 1 See Stalnaker (1984), 6. 2 See Gyekye (1971). See also Caston (2001) for a discussion of the notion of intentionality in Augustine. 3 See for example Perler (2001) and Perler (2002). 4 See for example Knudsen (1982).

12

Representation and Objects of Thought in Medieval Philosophy

virtually nothing is known about its origin. Scholastic philosophers, of course, also held that concepts have a representational nature,5 but not only concepts represent and exhibit intentionality for them. Many claimed that the senses, memory, imagination all represent things to the mind or the soul. It was usually claimed that they do so by being images of the things they represent, and they are thus like or similar to the things they are images of.6 They thus use representation in a much broader sense than we are nowadays used to in philosophy of mind. I will therefore often talk in this chapter about internal representation as representation in the soul, as opposed to external representation, like for example a blueprint, picture or map. The distinction between concepts as representations and images as representations, that is, so-called iconic representations, is also going to be important for what is to follow. When Descartes in the Third Meditation explains that ideas represent reality in different ways, that is, that some ideas have more objective reality that others,7 the

5 See for example the following passage from John Buridan, Commentary on Aristotle’s De anima (Second lecture), I, q. 5: ‘Termini autem vocales et scripti debent ordinari secundum exigentiam mentalium, quia non formantur nisi ad repraesentandum mentales.’ See also the following passage from John Major’s Sentence commentary: ‘Ad secundo concedo quod dus conceptus absoluti sufficienter absolute representat utrumquae extremorum. Hoc quodlibet extremorum in ordine ad se capio absolutum impresentiarum ut distinguitur contra relativum et non contra conotativum, sed illi duo conceptus nullo modo relative et comparative representant, cum unus illorum conceptum, puta a, representet sortem, et non representat eum in habitudine ad plato, et eodemmodo b respectu platonis. Necesse est dare unum alium conceptuum ab utrumquae istorum distinctum qui sortem in habitudine ad platonem representat, propter varios modos habendi rerum necesse est dare varios conceptus.’ (John Major, In secundum librum Sententiarum, dist. 38, q. 2, fol. cxlviiira.) 6 See for example John Buridan’s De anima commentary: ‘Quantum ad primum sciendum est quod idem est species, idolum, imago, similitudo; et ergo imaginandum est quod species intelligibilis est quaedam qualitas naturaliter repraesentativa ipsius obiecti, recte sicut imago, quae vulgariter dicitur esse in speculo, est repraesentativa rei obiectae speculo; sed sic directe in proposito: species intelligibilis est quaedam imago repraesentativa rei quae obiicitur intellectui.’ (John Buridan, Le Traité de l’Âme de Jean Buridan (De prima lectura), 457.) See also Tweedale (1989) and King in the present volume. 7 ‘Sed alia quædam adhuc via mihi occurrit ad inquirendum an res aliquæ, ex iis quarum ideæ in me sunt, extre me existant. Nempe, quatenus ideæ istæ cogitandi quidam modi tantum sunt, non agnosco ullam inter ipsas inæquilitatem, et omnes a me eodem modo procedere videntur; sed, quatenus una unam rem, alia aliam repræsentat, patet easdem esse ab invicem valde diversas. Nam proculdubio illæ quæ substantias mihi exhibent, majus aliquid sunt, atque, ut ita loquar, plus realitatis objectivæ in se continent, quam illæ quæ tantum modos, sive accidentia, repræsentant; et rursus illa per quam summum aliquem Deum, æternum, infinitum, omniscium, omnipotentem, rerumque omnium, quæ præter ipsum sunt, creatorem intelligo, plus profecto realitatis objectivæ in se habet, quam illæ per quas finitæ substantiæ exhibentur.’ (René Descartes, Oeuvres complètes de Descartes, VII, 40.)

Terminological and Conceptual Roots

13

usage of representations in the mind or soul is already a part of a well-established terminology and an important part of his theory of mind.8 The specific questions that this paper sets out to find an answer to are therefore: when does the term get the meaning here discussed and when does it become a central part of philosophy of mind or philosophical psychology? An important aspect of the history I am trying to sketch here is terminology, but it seems to me that the concept of representation as it is introduced into the history of philosophy is intrinsically tied to the term ‘representation’, and a necessary part of the history of the concept is therefore the history of the term. Part of my task of answering the above questions will hence be tackled by tracing the usage of the Latin terms ‘repraesentare’ and ‘repraesentatio’ in late ancient and medieval philosophy. It will then become evident that the usage of representations in the soul become a part of philosophical psychology at about the same time as the concept of ‘intentio’ is introduced, namely the answer to the two questions just posed is the Latin translation of the works of Avicenna. By using these Latin terms to translate several Arabic terms used by Avicenna the translators are forming or creating the concept of internal and mental representation. Interestingly, however, it is primarily imagination that Avicenna talks about as being representational in nature and not concepts. Thinking about concepts as representations comes into philosophy in a slightly different way and much later. 2 The Ancient Background The English words ‘representation’ and ‘to represent’ derive via Old French from the Latin words ‘repraesentatio’ and ‘repraesentare’, but these are by no means commonly used words in classical Latin.9 The Oxford Latin Dictionary gives primarily three meanings to these words. They can mean either (i) a payment in ready money, or (ii) an act of bringing something before the mind, or (iii) an image or a representation in art. For obvious reasons I am most interested in (ii) and (iii), and they involve the idea of re-presenting something previously absent as present, that is, making something present again.10 Quintilian is one of the few that uses the word in an interesting way. In his Institutio oratoria, he writes: 8 For a discussion of the background to Descartes usage of representational ideas see Nuchelmans (1983) and Ariew (1999). 9 A search in the CD-Rom BTL reviles that ‘repraesentatio’ is used nine times and ‘repraesentare’ 32 times. 10 This paper will not deal with the political connotations of the term representation, that is, talk about politicians as representatives and representative governments and so on. It has been dealt with in other studies and the terms seem not to have had this meaning in ancient times. It has been claimed that these concepts were not introduced until the thirteenth and fourteenth centuries when one started to talk about persons participating in church councils as representatives. See Pitkin (1967), 2–3, and Quillet (1971).

14

Representation and Objects of Thought in Medieval Philosophy onsequently we must place among ornaments that enargeia which I mentioned in the rules which I laid down for the statement of facts, because vivid illustration, or, as some prefer to call it, representation, is something more than mere clearness, since the latter merely lets itself be seen, whereas the former thrusts itself upon our notice.11

Quintilian uses the noun ‘repraesentatio’ in the sense of (ii) as something that clearly represents itself to the mind. Vivid illustrations or representations are important tools in good rhetoric, according to Quintilian. A person blessed with the ability to present a situation or action as if it were real – to create a representation so powerful and persuasive that the audience cannot but be convinced by it – is a powerful orator. The representation is then ‘self-evident’ or enargeia. It is literally like painting an unusually clear and convincing picture. This seems to be the closest to something like an internal representation an ancient author got, and Quintilian’s vivid illustrations are in a sense extremely rich and complicated representations. The orator describes a situation with words and tries to effect other person’s imagination with these words in order to create an inner picture, that is, the orator is re-presenting the situation for us or for our mind. This is the meaning of (ii) and hence we see that (ii) means the same as (iii) with the only difference that (iii) is an external representation and (ii) an internal one. One could, of course, think that the use of representations in the soul in the technical sense sought after here is a translation of some Greek word used in such a way. It is highly unclear, however, what Greek word that would be, and no ancient Latin author seems to have associated a Greek term with representations in the sense discussed here. One suggestion has been that Plato’s and Aristotle’s word ‘phantasia’ should be translated as a faculty of representation and ‘phantasmata’ as ‘representations’.12 Phantasia is for Aristotle what comes between aistesis and nous, that is, the end product of sensation and the start of intellectual activity. It seems like a natural interpretation of, for example, Aristotle to see the ‘phantasmata’ as sensory representations of external objects, but this remains an interpretation, which, given the story I am telling, is hardly uncontroversial. Furthermore, there seem to be no Latin Ancient commentator or other Latin writer that presents such an interpretation of Aristotle until the twelfth century. As an example it should be mentioned that Cicero translates ‘phantasia’ with ‘visum’.13 11 ‘Itaque enárgeia, cuius in praeceptis narrationis feci mentionem, quia plus est evidentia vel, ut alii dicunt, repraesentatio quam perspicuitas, et illud patet, hoc se quodammodo ostendit, inter ornamenta ponamus.’ (Quintilian, Institutio Oratoria, 8.3.61.) 12 See for example Wedin (1988) and Caston (1998) for such readings of Aristotle’s philosophical psychology. 13 ‘Plurima autem in illa tertia philosophiae parte mutauit. In qua primum de sensibus ipsis quaedam dixit noua, quos iunctos esse censuit e quadam quasi impulsione oblata extrinsecus, quam ille phantasia, nos uisum appelemus licet, et teneamus hoc quidem uerbum, erit enim utendum in reliquo sermone saepius; sed ad haec, quae uisa sunt et quasi accepta sensibus, adsensionem adiungit animorum, quam esse uolt in nobis positam et uoluntariam.’ (Cicero, Academics, I, 11, 40.)

Terminological and Conceptual Roots

15

The Latin verb ‘repraesentare’ is, however, used in relation to the notion of phantasia, and again it is Quintilian that makes the connection. By the time he writes the term phantasia has, however, become more detached from the technical philosophical vocabulary of Plato and Aristotle and come to mean something like fantasy or imagination in a non-technical sense.14 The orator who will be the most effective in moving the feelings of the audience is, according to Quintilian, the one that has acquired a proper stock of phantasiai. In Book IV of the Institutio oratoria, he writes: That which the Greeks call phantasiai and we may call clear visions are those things through which the image of things not present are so represented to the soul that we seem to see them with our very eyes and have them before us.15

Quintilian here uses the term ‘repraesentare’ in virtually the same sense as before. It thus seems that the classical usage of this term, which has bearing on the history of the notion of internal representation, is strictly limited to rhetoric. 3 Early Medieval Usage of Representation In the early Middle Ages, the classical usage of the terms ‘repraesentare’ and ‘repraesentatio’, of course, remained, although they seem to have been used rather infrequently in the sense of (ii). There seems, however, to be a whole tradition of theological writers using the notion of representation as image or example, that is, in the sense of (iii) above. One example can be found in Abelard’s Epitome Theologiae Christianae where he writes: In such a way are we obliged to have in the sacrament Christ before our eyes, in the way he was led to the passion, suffered and crucified for us. This representation of his love, which he himself has shown to us, makes us remember.16

Here Abelard stresses that we must in the Holy Communion have Christ’s suffering before our eyes as an example or as an image of his love in order for us to remember 14 See Watson (1988). It has also been claimed that the Stoics use of ‘phantasiai’ should be translated as ‘mental representations’, but also this is an interpretation and no Latin Stoic author used the word representation is this way. The Latin word for ‘representation’, as seen, seems to have had a different meaning in ancient times than what it has later in the history of philosophy, at least in relation to philosophy of mind and psychology. Maybe the Greek word ‘phantasia’ comes closer to our notion of ‘mental representation’, but this will always be a matter of interpretation. See the discussion in Long (1991). 15 ‘Quas phantasiai Graeci vocant, non sane visiones appellemus, per quas imagines rerum absentium ita repraesentantur animo, ut eas cernere oculis ac praesentes habere videamur.’ (Quintilian, Institutio Oratoria, 4.2.29–30.) 16 ‘Sic enim in sacramento Christum prae oculis habere debemus, tanquam ad passionem ductum, passum et crucifixum pro nobis, quae repraesentatio dilectionis illius nos memores facit, quam ipse nobis exhibuit.’ (Cap. XXIX; PL, Col. 1740D.)

16

Representation and Objects of Thought in Medieval Philosophy

that love. This representation seem to be an external representation in the form of Christ on the Holy cross, although we should retain it in our souls and it can, therefore, be seen as a representation in the sense of (ii).17 Abelard’s usage reminds us very much of Quintilian’s usage of the term, but with the specific Christian element added. The usage of the word representation in this sense and with this strong theological emphasis seems to stem from Tertullian. The word occurs quite frequently in his texts, which it does not in other texts from his time.18 Tertullian usage of this term has been studied before by Adhémar d’Alès in the book La théologie de Tertullien from 1905. He claims that there are three main uses of the verb ‘repraesentare’ in Tertullian, namely one physical, which has to do with something being really present; one mental, which has to do with representations in the imagination or the intellect; and finally one moral, which has to do with images of examples – preferably Christian.19 The first usage is in line with (iii) above and the third is basically the same as the one we have seen in Abelard, but the second usage mentioned by d’Alès is interesting. He, however, hardly presents any examples of this usage in Tertullian. His best example is from the book De spectaculis in which Tertullian writes: … we have already this [the coming of Christ] represented before us by the power of the Holy Ghost.20

This usage is the same as the one we have seen in Abelard. We have by the power of the Holy Ghost already before Christ’s return a representation or some mental image of this arrival. By his relatively frequent use of these terms Tertullian is laying the foundation for the different usage we can find in the early Middle Ages. He and everybody else in the early Middle Ages seems to use the notion of representation with Christian or theological connotations. Augustine, for example, very rarely uses these words,21 but none of his uses are interesting or relevant for the story told here. He, for example, uses the verb in the same sense as Abelard seen above. In the Trinity, he writes that the trinity is 17 Another similar usage can be found in Tertullian’s De anima. He writes: ‘Atquin in resurrectionis exemplis, com dei virtus sive per prophetas sive per Christum sive per apostolos in corpora animas repraesentat, solida et contrectabili et satiata veritate praeiudicatum est hanc esse formam veritatis, ut omnem mortuorum exhibitionem incorporalem praestigias iudices.’ (Tertullian, De anima, 57, 12.) 18 A search in Patriologia Latina reveals all in all 114 hits – although some are by the editors of PL. 19 See d’Alès (1905), 356–60 20 ‘Ut talia spectes, ut talibus exsultes, quis tibi praetor aut consul, aut quaestor, aut sacerdos de sua liberalitate praestabit? et tamen haec jam quodammodo habemus per fidem spiritu imaginante repraesentata. Caeterum qualia illa sunt quae nec oculus vidit, nec auris audivit, nec in cor hominis ascenderunt? Credo, circo et utraque cavea et omni stadio gratiora.’ (Tertullian, De spectaculis, cap. 30, PL, col. 0662A–0662B.) 21 A search in Patriologia Latina reveals 35 hits in his entire corpus and most are by the editors.

Terminological and Conceptual Roots

17

22

represented in reality. He also uses the noun once in City of God in the context of whether this is the time when God will re-present (restore or make present again) the sovereignty to Israel.23 None of these are of particular interest to us here. The Latin words ‘repraesentare’ and ‘representatio’ are not used in any of Boethius’s or the later medieval translations of Aristotle’s works into Latin.24 The only exception is again found in the Rethoric. It is William of Moerbeke’s thirteenth century translation in which he translates ‘mimesis’ with ‘repraesentatio’. See for example the following passage (I, 11 (1371b4–8)): And since learning and admiring are pleasant, all things connected with them must also be pleasant; for instance, a work of imitation, such as painting, sculpture, poetry, and all that is well represented, even if the object of representation is not pleasant; …25

In the early translation by James of Venice, the Latin term ‘imitatio’ was used to render ‘mimesis’.26 An early medieval author that many associate with the notion of having a representation in the mind is Anselm. The passage people then are thinking about is in the second chapter of Proslogion where he introduces the concept ‘somethingother-than-which-nothing-greater-can-be-thought’. This concept is in the mind of the person thinking the concept and it is of something, namely God, and one could say 22 ‘Attributa per Hilarium singulis personis. Trinitas in rebus factis repraesentatur. Quidam cum vellet brevissime singularum in Trinitate personarum insinuare propria, Aeternitas, inquit, in Patre, species in Imagine, usus in Munere.’ (Augustine, De trinitate, VI, 12. PL, col. 0931.) 23 ‘Illam sane novissimam persecutionem, quae ab Antichristo futura est, praesentia sua ipse exstinguet Jesus. Sic enim scriptum est, quod eum interficiet spiritu oris sui, et evacuabit illuminatione praesentiae suae. Hic quaeri solet, Quando istud erit? Importune omnino. Si enim hoc nobis nosse prodesset, a quo melius quam ab ipso Deo magistro interrogantibus discipulis diceretur? Non enim siluerunt inde apud eum; sed a praesente quaesierunt, dicentes: Domine, si hoc tempore praesentaberis, et quando regnum Israel [in a footnote as an alternate reading of the manuscript: si hoc in tempore repraesentabis regnum Israel]? At ille: Non est, inquit, vestrum scire tempora, quae Pater in sua posuit potestate.’ (Augustine, City of God, XVIII, cap. 53, PL, col. 0616.) 24 Representation in any form is not used in James of Venice’s translation (vetus translatio) of Aristotle’s De anima. See Anonymi, Magistri Artium (c. 1246–47), Sententia super II et III De anima. 25 Moerbecke’s Latin translation is the following: ‘Quoniam autem addiscere delectabile et mirari, et talia necesse delectabilia esse, scilicet imitativum, ut protractiva et statuificatio et poetica, et omne quodcumque fuerit representatum, et si non sit delectabile id cuius est representatio; …’ (Aristoteles Latinus, XXXI, 1–2, Rhetorica.) 26 As a note it might be interesting to mention that phantasia and mimesis are connected in late Classical thought. It is Philostratus that makes the connection in his Life of Apollonius from 217. Philostratus reports Apollonius to have said that phantasia is a much more skilful craftsman than mimesis. He says: ‘For mimesis will produce only what she has seen, but phantasia even what she has not seen as well; and she will produce it by referring to the standard of perfect reality.’

18

Representation and Objects of Thought in Medieval Philosophy

it represents God, but this is not what Anselm says. He does not use this particular word in this connection.27 He, however, uses the Latin words for representation, but not in any of his major works, only in his orations and letters,28 and none of them are of interest to us here. 4 The Introduction of Representation in the Soul in the Twelfth Century The twelfth century was an extremely dynamic time in medieval intellectual history and in a sense it is no wonder that it is in this century that we find the concept and terminology of representations in the soul being exploited seriously for the first time. This century saw the reappearing of Aristotle’s works in the West and numerous other works were also well being translated for the first time into Latin. It is also as part of this translation project that much of the foundation of the vocabulary of Latin Western philosophy was laid down. In fact, it is in this environment that the notion of representation in the soul is introduced. The terminology of representation in relation to the operations of the soul can be divided into three main groups and the introduction of the terminology seems also to fall into these groups, namely sense (visual) representation, internal (sense) representation and mental (conceptual) representation. This section of the paper is therefore divided into three separate subsections. In the first subsection, I will outline how the usage of representation was introduced in relation to sensation. In the second subsection, I will explain how representation becomes associated with the internal sense in the Latin translation of Avicenna. Finally, I will trace the history of conceptual or mental representation.

27 ‘Ergo, domine, qui das fidei intellectum, da mihi, ut quantum scis expedire intelligam, quia es sicut credimus, et hoc es quod credimus. Et quidem credimus te esse aliquid quo nihil maius cogitari possit. An ergo non est aliqua talis natura, quia “dixit insipiens in corde suo: non est deus”? Sed certe ipse idem insipiens, cum audit hoc ipsum quod dico: “aliquid quo maius nihil cogitari potest”, intelligit quod audit; et quod intelligit in intellectu eius est, etiam si non intelligat illud esse. Aliud enim est rem esse in intellectu, aliud intelligere rem esse. Nam cum pictor præcogitat quæ facturus est, habet quidem in intellectu, sed nondum intelligit esse quod nondum fecit. Cum vero iam pinxit, et habet in intellectu et intelligit esse quod iam fecit. Convincitur ergo etiam insipiens esse vel in intellectu aliquid quo nihil maius cogitari potest, quia hoc cum audit intelligit, et quidquid intelligitur in intellectu est. Et certe id quo maius cogitari nequit, non potest esse in solo intellectu. Si enim vel in solo intellectu est, potest cogitari esse et in re, quod maius est. Si ergo id quo maius cogitari non potest, est in solo intellectu: id ipsum quo maius cogitari non potest, est quo maius cogitari potest. Sed certe hoc esse non potest. Existit ergo procul dubio aliquid quo maius cogitari non valet, et in intellectu et in re.’ (Anselm, Opera omnia, I, 101–2.) 28 He uses the words 8 times in his entire corpus and only in orations and letters.

Terminological and Conceptual Roots

19

Sensory representation If not the first, then at least one of the very first uses of the notion of sensory representation can be found in the twelfth century translation of Nemesius of Emesa’s De natura hominis by Burgundio of Pisa. He translates the Greek word ‘parhistemi’ with ‘represento’. The Greek term means something like ‘to set before’. In the chapter on vision, he writes: Whenever truly in conformity with itself [vision] manifestly represents that which appears, since [whatever appears] is not seen from a long distance, but from a long distance a square tower, in fact, appears round.29

Here the word is used in exactly the sense of a sense representation of some external thing to the soul. The idea Nemesius wants to express is that vision, whenever it functions normally, correctly represents things as they appear, since at close range and under normal conditions things are as they appear. At a longer distance things are not always as they appear and in a visual appearance what is represented as round by the sense might in fact be square. The square tower that appears round from a distance is a famous and classic example, which goes back at least to Sextus Empericus in his famous work Against the Logicians from the second century,30 but it is reported as well by Tertullian in his De anima. Tertullian does not use the term representation in the context of this example.31 The same kind of example can be found in the early thirteenth century philosopher John Blund’s Tractatus de anima. He was heavily influenced by Avicenna, and, in his chapter on vision, he takes up the example mentioned and asks how indeed a round thing can appear round. When [a round] immovable [object] exists in the eye the external thing is represented as round, because the object, which it is in, is round, namely the eye, and this immovable external thing is thus represented. Therefore, when the soul judges in accordance with its representation of this immovable [object], [it judges in accordance with] whatever seems to appear round to it.32 29 ‘… quandoque vero secundum se ipsum manifeste repraesentat ea quae apparent, quando non a longe videt; turrim denique quadrangulatam ut rotundam videt a longe.’ (De natura hominis, Cap.VI, 79.14–17.) 30 See Against the Logicians, I, 208–9. 31 ‘Itaque mendacium visui objicitur, quod remos in aqua inflexos vel infractos adseverat adversus conscientiam integritatis, quod turrem quadrangulatam de longinquo rotundam persuadeat, quod aequalissimam porticum augustiorem in ultimo infamet, quod coelum tanta sublimitate suspensum mari jungat.’ (Tertullian, De anima.) 32 ‘Inmutatio autem existens in oculo representans rem extra est rotunda, quia suum subiectum in quo est [est] rotundum, scilicet oculis, et illa inmutatio representat rem extra. Ergo cum anima iudicet secundum representationem illius inmutationis, quodlibet videtur apparet rotundum.’ (John Blund, Tractatus de anima, cap. 9, §. 90.)

20

Representation and Objects of Thought in Medieval Philosophy

It does not seem a very good explanation as to why we see round things as round that our eyes are round, but that is beside the point here. He thinks the eye represents external objects to the soul and it is based on these representations that the soul judges things to be in certain ways.33 Presumably, if the representation is wrong, the soul’s judgments will be wrong as well. The way the notion of representation is introduced into the Western philosophical tradition, at least when it comes to representations in the sense organs, thus suggests a view of knowledge as accuracy of representation. If our representations of the world are incorrect, our judgments of the world will be incorrect as well. Internal representation The usage of the terms ‘repraesentatio’ and ‘repraesentare’ as internal representations in psychology or philosophy of mind seems to have its origin in the Latin translation of the works of Avicenna. In no other work before this translation can such a frequent use of these Latin words be found and, more important, they were not until this translation associated with the internal senses. The translation of Avicenna’s works into Latin was done at about the same time as Nemesius of Emesa’s De natura hominis was translated. The works I will be looking at are in Latin two separate works, namely the so-called Metaphysics (Liber de philosophia prima sive scientia divina) and De anima (Liber de anima seu sextus de naturalibus), but in Arabic only one very large work, namely the encyclopedic work Kitâb al-shifâ’ (The Book of Healing). Gerard of Cremona and Gundisalvi from Gundussalinas translated them around 1150.34 The Latin translation of the part of Avicenna’s al-Shifâ, which for some time came to be viewed as a commentary on Aristotle’s De anima, was extraordinary influential. Maybe its most profound influence was on terminology. Even though some, foremost nominalists, dismissed the faculty psychology they were still influenced by Avicenna’s terminology and one such terminological shift, which has had a profound influence on Western philosophy is the association of representations in the soul with our internal senses, that is, with fantasy, imagination and memory. This is, I think, the background against which all later talk of representations in the soul should be viewed. The notion of representation in the soul is, not surprisingly, used foremost in the De anima, but there seems to be no substantial difference in the usage of the terms 33 The same usage of representation can be found later on in Blund’s work, but then in relation to sound. He writes: ‘Contingat ita quod diversi soni sint quorum unus sit gravis et reliquus acutus, et eque cito perveniant ad aurem. Passio ergo generata in aure est confecta ex duobus sonis oppositis. Sed si duo sapores conmisceantur simul anima per gustum conmixtum apprehendet conmixtionem que est in passione: non enim iudicat nisi secundum representationem passionis invente per instrumentum sensus. Pari ratione anima habet iudicare mixtum esse sonum a gravi et acuto propter mixtam passionem in aure.’ (Ibid., cap. 13, §. 185.) 34 See Verbeke’s introduction to the editions Avicenna Latinus.

Terminological and Conceptual Roots

21

in the Metaphysics and in the De anima. I can only give here a few examples of how Avicenna uses the notion of representation, since a more comprehensive treatment must include an account of Avicenna’s whole psychology. I will only give a brief summery in order to show what space the translation into Latin of Avicenna gives to the inner representations. Avicenna’s psychology is a faculty psychology. According to the philosophers, there are five internal senses or faculties, and, according to the doctors, there are only three. In the front ventricle of the brain humans have two senses, namely the common sense, which receives all sense impressions, and the phantasia, which retains them. The third faculty is located in the middle ventricle of the brain and called the imaginative faculty, but it is also called cogitative when the intellect makes use of it. It combines the forms or images stored in phantasia. The fourth faculty is not in any part of the brain, and is thus not recognized by the doctors. It is the estimative faculty, which is shared by humans and animals and is a sort of understanding or, rather, instinctive understanding. It is with this faculty that, for example, the lamb knows that the wolf is dangerous. The fifth and last faculty is the memory in the back ventricle of the brain.35 The use of representation in Avicenna is associated with all the five faculties; the phantasia, or imagination (imaginatio) as it is called in the Latin translation, imaginative, estimative and memory faculties. The images received through the common sense and stored in phantasia are called representations. The images in the soul are thus on Avicenna’s terminology representations of perceived objects. He, for example, writes: … the acts of the imagination and its representations; …36 … the representation in the imagination are made stronger with the help of the intellect.37

The imaginative or cogitative faculty divides and puts together the representations that are collected in the phantasia into new representations, which might not have any real object corresponding to them. These representations are still images of some sort. Although they are not images in any naive sense, since the phantasia is suppose 35 For comprehensive and easily understandable overview of Avicenna’s psychology see Harvey (1975). See also Rahman (1957) and Hasse (2000). 36 ‘Quidam enim ex illis est fortis in memoriter retinendo sed debilis in recordando, eo quod est siccae complexionis quae retinet quod apprehendit, sed materia cum movetur anima non est oboediens actionibus imaginationis et eius repraesentationibus; …’ (Avicenna, Liber de anima, IV, cap. 3, 42.) The Arabic term here translated as representations is ‘isti‘râdât’, which is plural and means literally representations. 37 ‘Si vero impedita non fuerit, non egebit eo postea in suis propriis actionibus, nisi in aliquibus tantum in quibus opus est redire virtutes imaginativas et considerare ea iterum ad hoc ut percipiat principium aliud ab eo quod habuerat, et adiuvent repraesentare id quod appetitur in imaginatione, et repraesentatio eius in imaginatione firmetur auxilio intellectus.’ (Ibid., V, cap. 3, 104–5.) The Arabic term here translated with representation is ‘tamaththul’, which is a singular noun and means literally representation.

22

Representation and Objects of Thought in Medieval Philosophy

to compile sensation from all the five external senses. He for example writes about the representations in the imaginative faculty: … the imaginative represents then the forms which the soul is apt to choose and to mix together …38

The senses apprehend the sensible forms of the objects we perceive and the estimative faculty apprehends the ‘intentio’ or ‘ma’na’ of the perceived object. Avicenna’s standard example is the lamb perceiving the wolf. It is the apprehension of the ‘intentio’ of the wolf that makes the lamb flee. In this process all the internal faculties are involved and the notion of representation plays an absolutely crucial role in explaining this process.39 The ‘intentio’ of the wolf is also remembered and next time the lamb sees a wolf the response is much more automatic. The memory retains the representation of the wolf together with the estimation of that representation and thus whenever the representation is recalled the estimation follows with it. He writes: [A memory representation] which is created with the estimation converts to its [corresponding representation] in the imaginative faculty and represents whichever of the forms that are in the imagination … 40 38 ‘Naturalia autem sunt quae veniunt ex temperantia virtutum humorum cum spiritu gerente virtutem formalem et imaginativam: primum etenim non repraesentat nec occupatur nisi per hoc, et aliquando etiam repraesentat id quod est in corpore et quod est illi accidens, sicut cum virtus expulsiva spermatis movetur ad expellendum, virtus vero imaginativa repraesentat tunc formam quam anima solet diligere et commisceri cum ea; habenti vero famem, repraesentat fercula, et cui fuerit necesse expellere superfluitatem, repraesentat locum ipsius, et cui acciderit quod aliquod membrum eius calescat aut infrigidetur calore aut frigiditate, repraesentat ei quod illud menbrum mittitur in ignem aut in aquam frigidam.’ (Ibid., IV, cap. 2, 30.) The Arabic word which here has been translated with the Latin verb ‘repraesentare’ is the verb ‘hakâ’, which means something like ‘to be similar’ or ‘to be like’. This makes it even clearer that one should understand these representations as images of some form which purport to be like the objects they are of. 39 The same kind of usage of ‘to represent’ as can be found in Avicenna does also occure in Averroes major commentary on the De anima. In one place he writes: ‘Et sunt tres virtutes, quarum esse declaratum est in Sensu et Sensato, scilicet ymaginativa et cogitativa et rememorativa; iste enim tres virtutes sunt in homine ad presentandum (representandum in two manuscripts) formam rei ymaginate quando fuerit sensus absens, et ideo dictum fuit illic quod, cum iste tres virtutes adiuverint se adinvicem, forte representabunt individuum rei secundum quod est in suo esse, licet autem non sentiamus.’ (Averrois Cordubensis, Commentarium magnum in Aristotelis De anima libros, III, 449: 166–73.) The verb is also used on 475–6. ‘Praesentare’ and ‘praesentia’ is, however, used much more frequently. See 150; 364; 367; 374; 375; 377–8; 391; 419; 449; 468; 474; 476; 506; and 528. 40 ‘Quae virtus vocatur etiam memoralis, sed est retinens ob hoc quod id quod est in ea haeret firmiter, et est memoralis propter velocitatem suae aptitudinis ad recordandum per quod formatur cum rememorat post oblivionem, quod fit cum aestimatio convertitur ad suam virtutem imaginativam et repraesentat unamquamque formarum quae sunt in imaginatione,

Terminological and Conceptual Roots

23

The representations in the imagination are also the basis for intellectual activity. He writes: The sense represents by some form [an object] to the imagination and the imagination [represents] to the intellect …41

Thinking is always universal, according to Avicenna, and whatever is represented in the imagination is singular, hence, the intellect extracts the intention from whatever is presented or represented to it. If a form of the same species is presented to the intellect, it obviously does not extract another intention, or universal, from it. The same process is explained in the Metaphysics. He for example writes: In the intellect the form of animal is of such a kind that it in the intellect convenes to one and the same definition of many particular [things]. Wherefore one form in the intellect will be related to many, it is in this respect it is universal, because it is one intention in the intellect […] which is evident since, of which of those [things] which the form is a representative in the imagination, the intellect has plundered the intention of its accidents [and] acquired the form in the intellect.42

Here he describes the abstraction process, that is, how the representation in the imagination, which is of particulars, gets to be universal in the intellect. The universal forms are abstracted from the representations in the imagination and flow from the active intellect into the passive. Note, however, that the terminology of representations is never used in relation to the intellect. It is always the internal senses that represent in Avicenna and not the intellect or the external senses. In the table below, I have collected the whole list of Arabic terms translated with either the verb ‘repraesentare’ or the noun ‘repraesentatio’. They are also supplied with English translations.

ita ut quasi modo videat quod ipsae sunt formae eius.’ (Ibid., IV, cap. 1, p. 9.) The verb here translated as ‘repraesentare’ is in Arabic ‘ja‘ala’, which means literally ‘to represent’. 41 ‘Cum autem aliquam formam repraesentat sensus imaginationi et imaginatio intellectui, et intellectus excipit ex illa intentionem, si postea repraesentaverit ei aliam formam eiusdem speciei quae non est alia nisi numero, iam non excipiet intellectus ex ea aliam formam praeter quam acceperat ullo modo, nisi secundum accidens quod est illius propriam ex hoc quod est illud accidens, ita ut aliquando accipiat illam nudam, aliquando cum illo accidente.’ (Ibid., V, cap. 5, 129.) The Arabic term used here is ‘arada’, which means ‘to present’. The object represented in the imagination is thus presented to the intellect. 42 ‘In intellectu autem forma animalis taliter est quod in intellectu convenit ex una et eadem definitione multis particularibus. Quapropter una forma apud intellectum erit relata ad multitudinem, et secundum hunc respectum est universale, quia ipsum est una intentio in intellectu, cuius comparatio non variatur ad quodcumque acceperis animalium, videlicet quoniam, cuiusque eorum primum repraesentaveris formam in imaginationem, si postea exspoliaverit intellectus intentionem eius ab accidentibus, acquiretur in intellectu haec ipsa forma.’ (Avicenna, Liber de philosophia prima sive scientia divina, V, cap. 1, 237.)

24

Representation and Objects of Thought in Medieval Philosophy

Arabic

Latin

English

In the Metaphysics ahdara

repraesentare

to make present

hadara

repraesentari

to be present

sâdafa

repraesentari

to meet; to encounter

ja‘ala

repraesentare

to represent

hakâ

repraesentare

to resemble

muhâkâh

repraesentare

to resemble

‘arada

repraesentare

to present

‘ard

repraesentare

presentation

tamthîl

repraesentare

representation

isti‘râdât

repraesentatio

representations

mâ yalûh

repraesentatio

what appears (as something)

tamaththul

repraesentatio

representation

In the De anima

The influence of Avicenna on later medieval philosophical psychology is clearly seen if we again have a look at John Blund’s Tractatus de anima. He uses the same kind of terminology in relation to the imagination and memory.43 The influence becomes even clearer if we turn our attention to Thomas Aquinas. He uses the notion of representation in relation to phantasia in all his works, which deal with the powers of the soul. He hardly ever uses this terminology in relation to the external sense and never in relation to concept or what he calls the mental word (verbum). It always has to do with the internal senses or the abstracted universal species, which represents without material conditions, that is, the so-called sensible and intelligible species.44 43 See, for example, John Blund, Tractatus de anima, cap. 20, §. 271 and §. 273. 44 ‘Circa ea vero quae hic dicuntur, considerandum est, quare sensus sit singularium, scientia vero universalium; et quomodo universalia sint in anima. Sciendum est igitur circa primum, quod sensus est virtus in organo corporali; intellectus vero est virtus immaterialis, quae non est actus alicuius organi corporalis. Unumquodque autem recipitur in aliquo per modum sui. Cognitio autem omnis fit per hoc, quod cognitum est aliquo modo in cognoscente, scilicet secundum similitudinem. Nam cognoscens in actu, est ipsum cognitum in actu. Oportet igitur quod sensus corporaliter et materialiter recipiat similitudinem rei quae sentitur. Intellectus autem recipit similitudinem eius quod intelligitur, incorporaliter et immaterialiter. Individuatio autem naturae communis in rebus corporalibus et materialibus, est ex materia corporali, sub determinatis dimensionibus contenta: universale autem est per abstractionem ab huiusmodi materia, et materialibus conditionibus individuantibus. Manifestum est igitur, quod similitudo rei recepta in sensu repraesentat rem secundum quod est singularis; recepta autem in intellectu, repraesentat rem secundum rationem universalis naturae: et inde est, quod

Terminological and Conceptual Roots

25

Using the terminology of representation in connection with intelligible species is in itself an extension of Avicenna’s terminology. Concepts as representations (mental representation) In contemporary philosophy of mind the notion of mental representation is used foremost in relation to concepts and then they are often thought of as playing a linguistic role of some sort. They are also thought of as having meanings attached to them and in that sense function as signs of objects thought about. We have so far not seen any such use of the notion of representation in our history of how this notion was introduced. Remember also that the relevant sense of the word ‘intentio’ mentioned at the beginning of this paper is primarily used in relation to concepts. It thus remains to try to trace the history of this usage of the terms ‘repraesentare’ and ‘repraesentatio’. The first usage of the term representation that has a bearing on this sense of the word can be found in eleventh and twelfth century logic. Garlandus Compotista’s Dialectica and Abelard’s Dialectica contains discussions of a distinction between a words signification by imposition and by representation.45 A denominative like ‘white’ signify by imposition the substance which is white, while it signifies by representation the whiteness inhering in the substance. The white thing stands in for or is an instantiation of whiteness – white is re-presented in the object. Garlandus sensus cognoscit singularia, intellectus vero universalia, et horum sunt scientiae.’ (Thomas Aquinas, Sentencia De anima, lib. 2 l. 12 n. 5.) ‘Ad septimum dicendum quod licet species intelligibilis qua intellectus formaliter intelligit, sit in intellectu possibili istius et illius hominis, ex quo intellectus possibiles sunt plures; id tamen quod intelligitur per huiusmodi species est unum, si consideremus habito respectu ad rem intellectam; quia universale quod intelligitur ab utroque, est idem in omnibus. Et quod per species multiplicatas in diversis, id quod est unum in omnibus possit intelligi, contingit ex immaterialitate specierum, quae repraesentant rem absque materialibus conditionibus individuantibus, ex quibus una natura secundum speciem multiplicatur numero in diversis.’ (Q. d. de anima, a. 3 ad 7.) 45 Garlandus writes: ‘Sed quia adhuc istud totum convenit quibusdam nominibus differentibus solo casu et ab aliquo venientibus et secundum nomen, idest litteratura et sensu, sed non venientibus ad idem designandum – nisi per representationem, ut “viator” quodammodo viam representat –, ad differentiam illorum dicitur “habent appellationem”, idest habent iuxta se appellationem, idest veniunt ad idem designandum ita ut significatum primitivi videatur appellare derivativum illius primitivi ad se designandum.’ (Garlandus Compotista, Dialectica, 17: (De denominativis).) Abelard writes: ‘Sunt autem qui nec inter dictiones signa huiusmodi proprie velint admittere nec ipsa dicant propriam vocum significationem ex impositione tenere, sed quamdam indicationem facere de voce subiecta, qualiter ipsa accipiatur, circa omnes scilicet, non circa unum tantum, secundum quod Boethius in Secundo Divisionum huiusmodi signa determinationes appellat. Cum enim “homo” et circa unum per se et circa omnes possit accipi, cum “omnes” aut “quidam” ponitur, quodammodo determinat quot in nomine ipso accipiantur, hoc quoque signare dicitur, non per propriam vocis impositionem, sed tamquam per realem repraesentationem, qui et circulus.’ (Abelard, Dialectica, tract. II (De categoricis), 188.)

26

Representation and Objects of Thought in Medieval Philosophy

mentions the example of a traveler (viator), which can be said to represent a road (via). The term ‘traveler’ signifies by imposition the human being, which is a traveler, but it represents the road the traveler travels on. The same distinction can be found in Logica modernorum. In the words of that author the distinction is explicated in the following way: One should, however, note that a name does not nominally signify the substance and the quality at the same time … but it always names the substance, since it was imposed to do so. The quality, however, is signified non-nominally. More correctly it is a representation and determination of a substance in accordance with which the signification of the substance was imposed. This is why ever name has two significations; one by imposition of the substance and another by representation of the quality of the substance.46

It is not quite true as the author here suggests that every term (name) has two kinds of signification. Terms that have both these kinds of signification are appellative terms, and they should be distinguished from substance terms or natural kind terms, which only have signification by imposition. William Ockham will call such terms connotative and claim that their meaning is expressed by two kinds of signification, namely primary and secondary signification. This distinction is also related to Anselm’s distinction from De grammatico between per aliud and per se signification.47 The use of the term representation found here is of the more abstract sense as we would expect in talking about concepts or ideas as representations, but there are also differences. When concepts are described as representations the classical meaning of the word has changed, since concepts are not like or similar to the things they are about. In the classical usage, there were two notions involved, namely a standing in for something else (being a sign) and being similar to or a picture of (being an icon). The standing in for relation is preserved when concepts are representations, but the being similar or a picture of relation is lost. Garlandus’s and Abelard’s usage described above is certainly reminiscent of this although for them only appellative terms represent. Words and mental concepts are treated analogously, that is, a linguistic sign representing whatever it is of and by the very notion of it being a representation it has a content and a meaning. There is, however, no attempt in Abelard’s time or later when Avicenna is being translated to try to internalize this use of representation. In the late thirteenth century, John Duns Scotus developed a new theory of thought. According to him, thinking is having a mental act or concept in mind, which 46 ‘Notandum est tamen quod nomen non significat substantiam et qualitatem insimul nuncupative, ... sed substantiam tantum nominat, quia ei fuit impositum, qualitatem vero significat non nuncupative, immo representando et determinando circa substantiam propter quam tamen notandam substantie fuit impositum. Quare omne nomen habet duas significationes, una per impositionem in substantia, alteram per representationem in qualitate ipsius substantie.’ (de Rijk, Logica modernorum, II, i, 228 nota 1.) 47 See Marenbon (1996), 8.

Terminological and Conceptual Roots

27

is about whatever we are thinking of. A concept is an accident of the mind, which according to the terminology he develops, exists subjectively in the mind, while the content of the concept exists objectively in the mind. This allows him to draw a distinction between the thinking on the one hand and what the thinking is about on the other hand – a distinction that could not be drawn on a more crude Aristotelian theory of thought. To express the content of a concept he uses the terminology that the object is ‘sub ratione cognoscibilis seu repraesentanti’.48 Note, however, that this is still not the abstract use of representation we are looking for, since Scotus thought that the content of a concept was the form, in the Aristotelian sense, of the object thought about. Aquinas had thought in a more simplistic way that when one thinks of something the mind takes on the form of the object thought about. Scotus changed this and said that thinking is instead having a concept inhering in the mind, but the content of the concept is still the form. It is, therefore, primarily in the fourteenth and fifteenth centuries we find the terminology of representation being applied to concepts in a more abstract sense and then foremost in a nominalist tradition of thinkers, who develop a theory of thought involving a language of thought hypothesis. Ockham is the initiator of this tradition and it is carried on by thinkers like Buridan, Peter of Ailly and later also John Mair.49 According to these people, concepts function as signs of the thing thought about. Such a sign represents an object just because it is caused by the object, and it is mental simply because it is in the mind. Furthermore, a mental representation represents an object if it signifies that object. It hence functions as a ‘word’ for that object in a mental language. It should be stressed that neither Ockham nor Buridan use the terminology of representation very often when talking about a concept’s signification. Buridan uses ‘representation’ only once in this sense as far as I know,50 and it has been claimed that it was Albert of Saxony that introduced this terminology.51 I will here only give a taste of how they utilized the notion of representation to develop their theories. I have chosen Peter of Ailly, who lived in the late fourteenth and early fifteenth centuries, since he is unusually clear on terminology. In the beginning of his treaties on mental terms, the Conceptus, Peter states that there are three kinds of terms, namely, mental, spoken and written terms. He writes: A mental term is a concept, or an act of the intellective soul or the intellective power. A spoken term is an utterance (vox) signifying by convention (ad placitum). A written term is an inscription (scriptura) synonymous in signification with an utterance significative by convention.52 48 See John Duns Scotus, Ord., I, d. 3, pars 3, q. 1, n. 382. See also Tachau (1988), Chapter 2, Pasnau (1997), Pasnau (2003) and King in the present volume. 49 See Panaccio (1992). 50 See the quote from Buridan in footnote 5. 51 See Fitzgerald (2002), 17–18. 52 ‘Terminus mentalis est conceptus sive actus intelligendi animae vel potentiae intellectivae. Terminus vocalis est vox significans ad placitum. Terminus vero scriptus est

28

Representation and Objects of Thought in Medieval Philosophy

A mental term is a concept, which is the same as an intellectual act (or an act of the mind). While spoken and written terms signify by convention and are subordinate to mental terms. Mental terms, on the other hand, signify by nature. He makes a distinction, however, between mental terms properly so-called and improperly socalled, that is, between mental terms with, natural and conventional signification, respectively.53 I can, for example, have the word ‘animal’ in English in my mind without uttering it or writing it down – I am just thinking it. Although this is a mental term it is not a proper mental term, since it signifies by convention and is subordinate to a proper mental term that signifies by nature. The division of terms discussed here is, of course, a division of language as well. In this sense there are three distinct levels of language, namely, a written, a spoken and a mental language.54 Furthermore, there is a proper and an improper mental language. The terms of the proper mental language are concepts with natural signification and the terms of the improper mental language are mental words with conventional signification. The difference can also be explained by saying that the terms of the improper mental language are in, what we would call, some natural language, like English or Swedish, but this is not the case with the proper mental language – its terms are not in any natural language; it is the language other languages are based on. It is in this context Peter introduces the notion of representation. He does so in a passage on signification. He writes the following in the Conceptus: Next it must be noted that to ‘signify’ is the same as to be a sign of something. Nevertheless, a thing can be called a ‘sign’ of some thing in two senses. In one sense, because it leads to an act of knowing (notitia) the thing of which it is a sign. In another sense, because it is itself the act of knowing the thing. In the second sense, we say that a concept is a sign of a thing when such a concept is a natural likeness (similitudo) – not that it leads to an act of knowing that thing, but because it is the very act itself of knowing that thing, [an act that] naturally and properly represents that thing.55

In the first sense of to ‘signify’, terms signify in the improper mental, and in spoken or written languages. A concept signifies by being a natural likeness of whatever it stands for or, rather, represents. Peter writes earlier in the Conceptus that to scripura sinonima in significando voci significativae ad placitum.’ (Peter of Ailly, Conceptus et insolubilia, fol. Aib.) See also Spade (1980), 16. 53 See Peter of Ailly, Conceptus et insolubilia, fol. Aiiia, and, fol. Biib–Biiia, and also Spade (1980), 19–20, and, 36–7. 54 See Peter of Ailly, Conceptus et insolubilia, fol. Biia, and Spade (1980), 36. 55 ‘Notandum est deinde quod significare est idem quod signum rei facere hoc esse signum alicuius rei et veruntamen dupliciter aliqua res potest dici signum alicuius rei. Uno modo ut ducit in noticiam illius rei cuius est signum. Alio modo quia est ipsamet noticia rei. Secundo modo dicimus conceptum esse signum rei cuius talis conceptus est naturalis similitudo non quod ducat in noticiam illius rei, sed quia est ipsamet noticia rei naturaliter proprie representans rem.’ (Peter of Ailly, Conceptus et insolubilia, fol. Aiia.) See Spade (1980), 17.

Terminological and Conceptual Roots

29

‘signify’ is to ‘represent something, some things, or somehow, to a cognitive power by vitally changing it.’56 For a concept to signify something is for it to be the very act of knowing (notitia) that thing. Therefore, ‘a mental term, a concept or act of understanding (actus intelligendi), and an act of knowing (notitia) that apprehends a thing are the same.’57 Here we see that representation has taken on the contemporary or rather modern meaning as a mental particular with a causal role standing in for something else. Furthermore, we also see that the terms ‘representation’, ‘mental term’, ‘concept’, ‘act of understanding’, ‘act of knowing (notitia)’ all mean the same thing. Peter also uses ‘idea’, but not in the Conceptus.58 5 Concluding Remarks I have been trying to show that not until the Latin translation of the works of Avicenna do we get the right combination of ideas to be able to say that now we have a theory of cognition that relies on a notion of representation in the soul. The pressing question is of course: Why Avicenna? The most important historical fact is that it is not until then that the word ‘representation’ is used frequently to explain our cognition of the world around us. But it is not a question of word only. I would even say that the concept does not exist before the translation. The main reason for this claim is that the translators used the Latin word for representation to translate a whole range of different words and concepts in the Arabic. Whether this was conscious or not it had as a consequence that they literally formed the concept of internal representation. The concept introduced, however, particularly suited the specific characteristics of Avicenna’s theory of the soul. Avicenna’s theory of the soul is an interesting mix of Aristotelian and Platonic conceptions of the soul. He takes on the Aristotelian theory of cognition and places it in an independently existing Platonic soul with a strong conception of a self. This makes the terminology of representation applied to the imagination particularly suitable. We are aware of objects external to us through our sensation of their sensible forms, which we receive via our external senses. These are collected and then represented to us in the imagination. We can then at will attend to these representations and abstract the universal forms from them. It is in the context of this theory in combination with the terminology that the concept of internal representation is created.

56 ‘Significare autem est potentiae cognitivae eam vitaliter immutando aliquid vel aliqua vel aliqualiter representare.’ (Peter of Ailly, Conceptus et insolubilia, fol. Aib.) See Spade (1980), 16. 57 ‘Notandum est viterius quam terminus mentalis, conceptus sive actus intellidendi et noticia rei apprehensiva idem sunt.’ (Peter of Ailly, Conceptus et insolubilia, fol. Aiib.) See Spade (1980), 18. 58 See Pluta (1987), 9: 3, 55. In this passage, Peter, however, talks about the intelligible species and he does not equate ‘idea’ and ‘concept’.

30

Representation and Objects of Thought in Medieval Philosophy

As soon as this notion is established in the Latin translation of Avicenna it naturally spread through Avicenna’s authority to scholastic philosophy in general and we find it being used quite frequently in the thirteenth and fourteenth centuries. It is not strange that we sooner or later also find the notion being used about concepts in the intellectual soul or mind. From its widespread usage in scholastic philosophical psychology it naturally spread to later philosophy and into modern philosophy. Bibliography Abelard, Peter, Dialectica, ed. L.M. de Rijk, Assen: van Gorcum, 1956. d’Alès, A. (1905), La théologie de Tertullien, Paris: Gabriel Beauchesne & Co, Éditeurs. Anonymi, Magistri Artium (c. 1246–47), Sententia super II et III De anima, eds (commentary) Carlos Bazan and (De anima) Kevin White, Louvain – Paris: Éditions Peeters, 1998. Anslem, Opera omnia, ed. F.S. Schmitt, Edinburgh: Thomas Nelson and Sons, 1938. Aquinas, Thomas, S. Thomae Aquinatis Doctor Angelici Opera omnia, Rome: Commissio Leonina, 1882–. Ariew, R. (1999), Descartes and the Last Scholastics, Ithaca, New York: Cornell University Press. Aristoteles Latinus, XXXI, 1–2, Rhetorica, ed. B. Schneider, Leiden: Brill, 1978. Averrois Cordubensis, Commentarium magnum in Aristotelis De anima libros, ed. Stuart Crawford, Cambridge, Mass.: The Mediaeval Academy of America, 1953. Avicenna, Avicenna latinus: Liber de anima, ed. S. van Riet and with an introduction by G. Verbeke, Leiden: Brill, 1968–72. ——, Avicenna latinus: Liber de philosophia prima, ed. S. van Riet and with an introduction by G. Verbeke, Leiden: Brill, 1977–80. Blund, John, Tractatus de anima, D.A. Callus and R.W. Hunt (eds), Oxford: Oxford University Press, 1970. Brentano, F. (1924), Psychologie vom empirischen Standpunkt, Leipzig: Verlag von Felix Meiner. Buridan, John, Le Traité de l’Âme de Jean Buridan (De prima lectura), B. Patar (ed.), Leuven: Éditions Peeters, 1991. Caston, V. (1998), ‘Aristotle and the Problem of Intentionality’, in Philosophy and Phenomenological Research 58, 249–98. —— (2001), ‘Connecting Traditions: Augustine and the Greeks on Intentionality’, in D. Perler (ed.), Ancient and Medieval Theories of Intentionality, Leiden: Brill, 23–48. Cicero, Marcus Tullius, On the Nature of the Gods. Academics, trans. H. Rackham, Cambridge, MA.: Harvard University Press, 1969.

Terminological and Conceptual Roots

31

Descartes, René, Oeuvres complètes de Descartes, vols. I-XII, C. Adam and P. Tannery (eds), Paris, 1897–1913. Duns Scotus, John, Iohannis Duns Scoti Doctoris Subtilis et Mariani opera omnia, eds. P Carolus Balic et al. Typis Polyglottis Vaticanae, 1950–. Empericus, Sextur, Against the Logicians, R.G. Bury (trans) Cambridge, MA: Harvard University Press. Fitzgerald, M.J. (2002), Albert of Saxony’s Twenty-Five Disputed Questions on Logic, Leiden: Brill. Garlandus Compotista, Dialectica, L.M. d. Rijk (ed.), Assen: van Gorcum, 1959. Glare, P.D.W. (1982), Oxford Latin Dictionary, Oxford: Oxford University Press. Gyekye, K. (1971), ‘The Terms “Prima intentio” and “Secunda intentio” in Arabic Logic’, in Speculum, XLVI, 32–8. Harvey, R. (1975), The Invard Wits: Psychological Theory in the Middle Ages and the Renaissance, London: Warburg Institute. Hasse, D.N. (2000), Avicenna’s De anima in the Latin West: The Formation of a Peripatetic Philosophy of the Soul 1160–1300, London: Warburg Institute. Knudsen, C. (1982), ‘Intention and Imposition’, in N. Kretzmann, A. Kenny and J. Pinborg (eds), The Cambridge History of Later Medieval Philosophy, Cambridge: Cambridge University Press, 479–95. Long, A.A. (1991), ‘Representation and the Self in Stoicism’, in S. Everson (ed.), Companions to Ancient Thought 2: Psychology, Cambridge: Cambridge University Press, 102–20. Major, John, In secundum librum Sententiarum, Paris, 1519. Marenbon, J. (1996), ‘Anselm and the Early Medieval Aristotle’, in J. Marenbon (ed.), Aristotle in Britain During the Middle Ages, Turnhout: Brepols, 1–19. Nemesius of Emesa, De natura hominis, eds G.R. Muncho and G. Verbeke, Leiden: Brill 1975. Panaccio, C. (1992), Les mots, les concepts et les choses. La sémantique de Guillaume d’Occam et le nominalisme d’aujourd’hui, Montréal: Bellarmin; Paris: Vrin. Pasnau, R. (1997), Theories of Cognition in the Later Middle Ages, Cambridge: Cambridge University Press. ——— (2003), ‘Cognition’, in T. Williams (ed.), The Cambridge Companion to Duns Scotus, Cambridge: Cambridge University Press. Perler, D. (ed.) (2001), Ancient and Medieval Theories of Intentionality, Leiden: Brill. —— (2002), Theorien det Intentionalität im Mittelalter, Frankfurt: Vittorio Klostermann. Peter of Ailly, Conceptus et insolubilia, Paris, 1500. Pitkin, H.F. (1967), The Concept of Representation, Berkeley: California University Press. Pluta, O. (1987), Die philosophische Psyschologie des Peter von Ailly, Amsterdam: B.R. Grüner. Putnam, H. (1991), Representation and Reality, Cambridge, MA: MIT Press.

32

Representation and Objects of Thought in Medieval Philosophy

Quillet, J. (1971), ‘Universitas populi et representation au XIV siècle’, in A. Zimmermann (ed.), Der Begriff der representation im Mittelalters: Stellvertretung, Symbol, Zeiehen, Bild, Berlin: Walter de Gruyter, 186–200. Quintilian, Institutio oratoria, H.E. Butler (trans), Cambridge, MA: Harvard University Press, 1921. Rahman, F. (1952), Avicenna’s Psychology, London: Oxford University Press. de Rijk, L.M., Logica Modernorum: A Contribution to the History of Early Terminist Logic Vols 1–2 , Assen: Van Gorcum, 1962-67. Rorty, R. (1979), Philosophy and the Mirror of Nature, Princeton, NJ: Princeton University Press. Spade, P.V. (1980), Peter of Ailly: Concepts and Insolubles, Dordrecht: Reidel. Stalnaker, R.C. (1984), Inquiry, Cambridge, MA: MIT Press. Stroud, B. (1984), The Significance of Philosophical Scepticism, Oxford: Oxford University Press. Tachau, K. (1988), Vision and Certitude in the Age of Ockham, Leiden: Brill. Tertullian, De anima, ed. J.H. Waszink, Amsterdam: J.M. Meulenhoff, 1947. Tweedale, M. (1989), ‘Mental Representations in Later Medieval Scholasticism’, in J-C. Smith (ed.), Historical Foundations of Cognitive Science, Dordrecht: Kluwer, 35–51. Watson, G. (1988), ‘Discovering the Imagination: Platonists and Stoics on phantasi’, in J.M. Dillon and A.A. Long (eds), The Question of ‘Eclecticism’, Berkeley: University of California Press, 208–33. Wedin, M. (1988), Mind and Imagination in Aristotle, New Haven: Yale University Press. CD-Rom sources: Bibliotheca Teubneriana Latina, CD-ROM version of Teubner’s Latin text editions, Stuttgart: B.G. Teubner; Turnhout: Brepols Publishers, 1999 [BTL]. Patriologia Latina Database, electronic version of the first edition of Jacques-Paul Migne’s Patrologia Latina, published between 1844 and 1855, Cambridge: Chadwyck-Healey, 1993–95 [PL]. Thesaurus Linguae Graecae, electronic version from University of California, Irvine, of Thesaurus Linguae Graecae canon of Greek authors and works, ed. Luci Berkowitz, Karl A. Squitier with technical assistance from William A. Johnson, New York: Oxford University Press, 1986 [TLG].

Chapter III

Abstract Truth in Thomas Aquinas* Robert Pasnau

Thomas Aquinas holds that the proper objects of intellect are the natures of material objects, conceived of universally through intellectual abstraction. This paper considers two questions regarding that doctrine: first, what are these abstracted, universal objects and second, given that the world is concrete and particular, how can such abstract, universal thoughts yield true beliefs about the world? 1 Prologue: The Objects of Intellect A central methodological principle of Aristotelian psychology – perhaps the central principle – is that the capacities of the soul must be investigated in terms of what sort of object that capacity has. In the words of Aquinas, ‘the nature of any capacity lies in its relationship to its proper object’ (InDA, II.13.69–70). This is most obviously true for the five external senses. There are five external senses, no more and no less, because there are five kinds of sensible qualities that we need to investigate in the world around us.1 The principle holds equally of the human intellect, or at least it should. But here its application is more problematic, because it is unclear just what the object of intellect is. Thomas Aquinas holds that the proper objects of the human intellect – that is, those things that the human intellect is naturally suited to understand – are the natures or quiddities of material things in the world around us. As we will see, this claim shapes a great deal of what he has to say about the nature of intellect. Later authors, however, would take very different views. On one account, sometimes attributed to Henry of Ghent, the proper object of the human intellect is God. What Ghent had actually claimed is that God is both the first and the ultimate object of intellect: ‘The beginning and the end of our cognition lies in God himself: the beginning, with respect to the most general cognition of him; the end, with respect to the nude * I received extremely useful comments on this chapter from participants at a UCLA workshop on medieval theories of truth. I also owe special thanks to Gyula Klima for his extensive written comments and to Jenny Ashworth for her helpful advice. 1 For the case of Aristotle, see Sorabji (1971). For Aquinas, see ST, 1a 78.3. The case of touch is problematic, because it is not clear how the various sensible qualities associated with touch – temperature, texture, etc. – fall into a single kind (see ST, 1a 78.3 ad. 3–4).

34

Representation and Objects of Thought in Medieval Philosophy

and particular vision of him’ (Summa, 24.7c [144rH]). This claim is based on the more familiar idea that the first objects of cognition are the transcendental concepts of being, goodness, truth, etc. In virtue of having these concepts, we understand everything else. These concepts, according to Ghent, are fundamentally ideas of God, and so in that sense we begin with a very general conception of God, and work toward the clear and distinct idea of God obtained by the blessed in heaven. Aquinas would have accepted Ghent’s claims about the primacy of transcendental concepts,2 and would also have accepted that God is the ultimate object of intellect. Yet he would have denied that this makes God the proper object of the human intellect. Now it is not obvious that Ghent himself wanted to endorse that further claim, but this is how he would later be read by John Duns Scotus. Scotus stresses that the question of the intellect’s proper object is the question of what object the intellect is naturally disposed to apprehend. That which is first, temporally, or even first and last, is not necessarily the proper object of intellect. Thus Scotus remarks, ‘the first natural object of a capacity has a natural relationship to that capacity’ (Ordinatio, I.3.1.3, n. 126). Plainly, if this is so in Ghent’s view, it is so only with respect to God’s most general attributes, the transcendental attributes that apply to all being. But Scotus then reasons: If God is the proper object of intellect only under his general attributes, then it is really those attributes, rather than God, that are the proper objects of intellect. This seems right. In general, when the intellect apprehends the universal attributes of some particular thing, we do not say that the particular thing is the object of intellect. Instead, we think of those attributes themselves as the object of intellect.3 Part of what makes it attractive to identify God as the proper object of the human intellect is that this ties our proper cognitive object into our ultimate (hoped for) cognitive destiny. If the end of human life is to achieve a face-to-face vision of God, then there would seem to be something plausible about thinking of God as what our intellect is naturally suited to apprehend. There is, in the same way, something unsatisfactory about Aquinas’s account, inasmuch as the blessed in heaven would seem to be abandoning their proper intellectual object – the quiddities of material objects – in favor of something else, God. If that is our ultimate destiny, and if life on earth is just a brief prologue to the eternity of our life to come, then it is hard to understand Aquinas’s insistence that the material world is what our intellect is naturally suited to apprehend. Scotus brings this point out quite effectively in arguing against Aquinas’s view.

2 See, e.g., InDH, 2.9–18 [§20]. 3 ‘Contra istam opinionem arguo sic: primum obiectum naturale alicuius potentiae habet naturalem ordinem ad illam potentiam; Deus non habet naturalem ordinem ad intellectum nostrum sub ratione motivi, nisi forte sub ratione alicuius generalis attributi, sicut ponit illa opinio [Henrici]; ergo non est obiectum primum nisi sub ratione illius attributi, et ita illud attributum generale erit primum obiectum. … Sed particulare quod non intelligitur nisi in aliquo communi non est primum obiectum intellectus, sed magis illud commune. Ergo etc.’ (Ord., I.3.1.3 n. 126).

Abstract Truth in Thomas Aquinas

35

We identify the first object of a capacity as that which is adequate to it by reason of the capacity, not that which is adequate to the capacity in a certain state – just as the first object of sight is not held to be that which is adequate to sight solely when it is in a medium illuminated by a candle, but that which is naturally suited to be adequate to sight in its own right, with regard to the nature of sight. (Ord., I.3.1.3 n. 186)

Trapped in a dark cave, we might see only shades of gray, but that does not make gray the object of sight, even if we spend our entire lives in a cave. So too, Scotus argues, for the human intellect. Even if in this life we have cognitive access only to the material world, that does not mean that material objects are the proper object of intellect. Scotus therefore proposes an alternative account, that the proper object of the human intellect is being (ens). This is to say that there is no one aspect of the world that the intellect, in its own right, is especially suited to apprehend. Everything that exists is a potential object of intellect, and the intellect is equally suited to grasp all of those things, insofar as they are beings.4 This might look like a disappointingly bland conclusion, because Scotus is in effect simply denying the whole premise of the discussion, that our intellect has something to which it is especially attuned, in the way that each of the senses has its own proper object. But in denying that premise, Scotus is actually making quite a striking claim, that there is nothing intelligible to any intellect that is unintelligible to us. Whatever any mind can know, our minds can know, at least in principle.5 (Even God’s essence is intelligible, albeit never completely, to the blessed in heaven.) This has the negative methodological implication that there is no special object of the human intellect that can give us a grip on what the nature of our intellect is. But it has the exciting positive implication that our intellect is qualitatively the same in its nature as all other intellects. We may lack the information that angels have, since we are not illuminated by God in the way that they are, and our minds might anyway lack the capacity to grasp such illumination fully. But despite these quantitative differences in how much we know and how smart we are, our minds are fundamentally the same in kind as the minds of God and the angels. I have not found Scotus explicitly saying quite that, but this is the view that he would have to take, if he is to abide by the Aristotelian tenet that capacities are distinguished in virtue of their objects. Here it may begin to seem as if Scotus’s view is implausible. In insisting on our intellect’s connection to the material world, Aquinas of course has in mind our constant reliance on the senses. Surely it is reasonable to suppose that this constant downward orientation makes for a fundamental difference between our intellects and those intellects that are not attached to any body. Scotus is entirely willing to grant what Aquinas has to say about our intellectual dependence on the senses in this 4 See, esp., Ord., I.3.1.3 n. 137. I discuss Scotus’s view in more detail in Pasnau (2003), 293–6. See also Honnefelder (1979), 55–98, and Marrone (2001), vol. 2. 5 ‘Quod habet in natura sua intellectum possibilem, potest ex natura sua cognoscere quodcunque cognoscibile, hoc est recipere cognitionem eius, quantum est ex parte sui’ (QQ, 14.6).

36

Representation and Objects of Thought in Medieval Philosophy

life. He accepts that our intellect receives all its information through the senses, and accepts Aquinas’s insistence that we must continuously turn back toward phantasms in the course of our thinking. Hence he allows that ‘with respect to what moves the intellect, in this state, its first adequate object is the quiddity of a sensible thing’ (Ord., I.3.1.3 n. 187). But although this is so de facto pro statu isto, it reveals nothing about the intellect’s intrinsic nature, because ‘it is not so as a result of the intellect’s nature, that which makes it an intellect’ (ibid.). Both Scotus and Aquinas agree that the nature of intellect should be proportioned to the nature of its proper object.6 Yet, according to Scotus, a human intellect separated from the senses would not carry with it any distinguishing features to mark it off as directed by nature at the material world. Aquinas, in contrast, is committed to the idea that there is something intrinsic to intellect that suits it to apprehend material things. It does not just happen that intellects like ours are connected to bodies. ‘It is natural for us to cognize things that have existence only in individual matter, because our soul, through which we cognize, is the form of one kind of matter’ (ST, 1a 12.4c). So he concludes that although the natures of material things are not in their own right among the easiest things to grasp, being material, these nevertheless are the things that we are most capable of understanding.7 Aquinas’s commitment to this characterization of the human intellect runs so deep that even his account of the beatific vision gets explained in these terms. The blessed in heaven, he tells us, will see the divine essence through a purely intellectual vision. But if that kind of experience is the ultimate end of human life, then why shouldn’t we agree with Scotus that facts about how the intellect operates in this life are no more significant than facts about how sight operates in a dark room? Aquinas deals with this sort of objection by incorporating his conception of the intellect’s proper object into his account of the beatific vision. Why should we suppose that the ultimate happiness for human beings is a vision of the divine essence? Because only such a vision would show us the ultimate causes behind the natural world. Without grasping the divine essence, we can know that certain things are the case, but we can never truly know why they are the case. If someone were to lack that ultimate explanation, ‘there would still remain for him the natural desire to inquire into the cause. Hence he would not yet be completely happy (beatus)’ (ST, 1a2ae 3.8c).8 6 Scotus makes this point by remarking, as quoted earlier, ‘primum obiectum naturale alicuius potentiae habet naturalem ordinem ad illam potentiam’ (Ord., I.3.1.3, n. 126). For Aquinas, similarly, ‘obiectum cognoscibile proportionatur virtuti cognoscitivae’ (ST, 1a 85.1c). 7 ‘Nam magis sunt nobis intelligibilia quae sunt sensui proximiora, quae in se sunt minus intelligibilia’ (SCG, II.77.1584). ‘[C]um anima humana sit ultima in ordine substantiarum intellectivarum, minime participat de virtute intellectiva; et sicut ipsa quidem secundum naturam est actus corporis, eius autem intellectiva potentia non est actus organi corporalis, ita habet naturalem aptitudinem ad cognoscendum corporalium et sensibilium veritatem, quae sunt minus cognoscibilia secundum suam naturam propter eorum materialitatem, sed tamen cognosci possunt per abstractionem sensibilium a phantasmatibus’ (InMet., II.1.285). 8 Cf. ST, 1a 12.1c, InMat., V.2.434; InJoh., I.11.212; CT, I.104.

Abstract Truth in Thomas Aquinas

37

Complete happiness, Aquinas claims, requires the satisfaction of all desires. That is plausible enough. What is startling is the further implication that, for the human intellect, the beatific vision is beatific because it supplies the means for us to satisfy our true intellectual goal, a thorough understanding of the material world into which we were born. This is a very odd result. It would be as if, in leaving Plato’s cave, we took satisfaction in what we saw under the sun only insofar as that explained what we had been seeing for all those years underground. Surely, however, we would quickly lose interest in facts about the cave. Wouldn’t the same to be true for the beatific vision? If seeing God’s essence is indeed what would make us perfectly happy, surely the reward would not come from what we would learn about the natures of material things. Would learning about the different genera and species of butterflies really make us all that happy, let alone perfectly happy? There is a general question here about the heavily intellectual nature of Aquinas’s account: his assumption that our perfect happiness consists in our intellect’s perfect satisfaction.9 But even setting that aside, it is hard to see how a perfect grasp of the material world could be so satisfying. (And would it continue to be satisfying even after Judgment Day brings the end of the world as we know it?) Moreover, one might well suppose that the material world would be quite uninteresting in comparison to what we could learn about the nature of God. But the latter, Aquinas insists, is not the proper object of the human intellect. Aquinas might diminish the impact of this criticism by stressing the deep fascination human beings do in fact have with the natural world. For a connoisseur of butterflies, seeing in the divine essence the whole order Lepidoptera surely would approach a kind of perfect happiness. And if butterflies leave you cold, that may well be just because you don’t know enough about them. The beatific vision would be like a kind of virtual reality in which you could quickly become an expert on anything, and enjoy the same kind of pleasure in that subject that an expert enjoys. Still, one might wonder whether such pleasures would really carry much weight in comparison with what we might come to learn about God. Although I have not found Aquinas addressing this question, he has a natural reply. For he repeatedly stresses that God remains incomprehensible to us, even through the beatific vision.10 This is not to say that we can know nothing about God, since Aquinas of course thinks we can know some things about God even in this life, and will know more still in heaven. But given his claim that perfect happiness requires the satisfaction of all our desires, and that intellectual desires are satisfied only when we completely 9 See, e.g., QDV, 8.1c: ‘Constat enim quod cuiuslibet intellectualis creaturae beatitudo consistit in sua perfectissima operatione. Illud autem quod est supremum in qualibet creatura rationali est intellectus. Unde oportet quod beatitudo cuiuslibet creaturae rationalis in nobilissima visione intellectus consistat.’ 10 ‘Omnis autem substantia intellectualis creata est finita: ergo finite cognoscit. Cum ergo Deus sit infinitae virtutis et entitatis, et per consequens infinite cognoscibilis, a nullo intellectu creato cognosci potest quantum est cognoscibilis; et ideo omni intellectui creato remanet incomprehensibilis’ (InJoh., I.11.213).

38

Representation and Objects of Thought in Medieval Philosophy

grasp what a thing is, he can hardly hold that God’s nature is the principal object of inquiry in heaven. If that were what we were after, we would be doomed to failure, hence unsatisfied, and hence unhappy. It is better, then, that we seek to grasp the nature of the physical world, even once we have left that world. This is, no doubt, a less awe-inspiring object than God himself. But at least it is something that we can fully grasp. These theological reflections illustrate just how thoroughly Aquinas is committed to his distinctive view regarding the objects of intellect. At this point, as a good Aristotelian, Aquinas should use this result to show us something about the nature of intellect itself – approaching the soul’s capacities through their objects. Indeed, since other created intellects (those of the angels) do not have material natures as their objects, we might anticipate that Aquinas will now have something really interesting to say about how our minds are fundamentally different from the minds of the angels. Alas, we never get quite that far, because Aquinas thinks we are not now in a position to say very much about the intellect’s inner nature.11 What we can do, however, is say something interesting about how the intellect operates: we can say that it operates through the process of abstraction from sensible data. To go only this far is still very much in the spirit of the governing Aristotelian methodology, according to which ‘acts and operations are conceptually prior to their capacities … and prior to these are their objects, (De anima, II.4, 415a18–20). Instead of leaping all the way from the intellect’s objects to its very nature as a capacity, Aquinas makes a halfway leap from the intellect’s objects to its operations, saving for some future generation the more difficult question of the intellect’s nature. Let us too set aside that more difficult question, and focus on the process of abstraction. As the previous paragraph implies, Aquinas postulates abstraction not because it answers to any sort of introspective data about how our intellect works, but because it seems entailed by his view that the objects of intellect are the natures of material things. Since he is not a Platonist, he cannot accept that these natures exist on their own, outside of particulars. Yet since the intellect grasps the universal, not the particular, he also cannot allow that the intellect grasps these natures as they exist in particulars. Describing the difference between Aristotle and Plato in this respect, Aquinas writes that ‘Aristotle was led to hold that the things that are intelligible to us are not intelligibles existing per se; instead, they are made out of sensible things’ (SCG, II.77.1584) – that is, made by intellect. Of course we are not supposed to conclude that the intellect simply makes up these common natures. They are ‘made’ by intellect in the sense that the natures of material objects must be made intelligible by abstracting away the individuating conditions of the object. At the same time, the real object of intellectual inquiry is the material world – it is not as if we are primarily concerned with our own thoughts. Thus Aquinas arrives at the 11 Aquinas does reach some very general conclusions, in places like ST, 1a Q79, such as that the intellect is distinct from the soul’s essence, that it is passive, that there is also an active intellect, that both passive and active intellects are parts of the soul and different for each human being. But these results are of course highly schematic.

Abstract Truth in Thomas Aquinas

39

following conclusion: ‘it is natural to us, through intellect, to cognize natures that have existence only in individual matter – not as they are in individual matter, but as they are abstracted from that matter through the consideration of intellect’, (ST, 1a 12.4c). We can now turn to the two puzzles that will be the subject of this paper. First, what are these abstract natures that are so central to our intellectual life? Second, if the world is concrete and particular, then how can such abstract thought yield true beliefs about that world? 2 The Fruits of Abstraction When modern philosophers speak of abstract objects, they mean things that actually exist, but in a mode not subject to the rules governing ordinary, concrete existence. Such objects may, for instance, not have a location, or may have more than one location, and they may not exist in time, or be subject to change, or have any causal efficacy.12 Aquinas’s conception of abstractness involves none of these things. His objects of intellect are abstract only in their content; their mode of existence is perfectly concrete, albeit immaterial, inasmuch as they exist as forms within intellect. Indeed, Aquinas’s conception of abstractness might seem utterly irrelevant to the modern idea, if not for the fact that Aquinas uses that concept to do exactly the work that modern platonists want to do with their concept of abstractness. What both parties are after is an account of how we think and talk about universal concepts, including natural properties, logical concepts, and mathematical truths. For modern platonists, these things exist, abstractly. Accordingly, when we think and talk about universal properties, etc., we are referring to things that actually exist. Aquinas utterly rejects this idea that there is some mode of existence other than the concrete, particular mode. Accordingly, when we think and talk about universal properties, etc., our thoughts do not precisely correspond with anything that actually exists. Nevertheless, these thoughts are the product of what exists, formed through a process of abstraction that captures certain aspects of the world at the expense of other aspects. To be abstract, for Aquinas, is not a way of existing but rather a way of representing things in the world. If Aquinas’s account could actually do the work it sets out to do, then it would surely be preferable to the modern platonic conception of abstract objects. The obscurity of that modern conception makes Aquinas’s account far more attractive, other things being equal. Responding to the original Platonism, Aquinas writes that ‘because Plato did not consider this sort of abstraction, he was compelled to posit separated mathematical entities and species’ (InDA, III.12.300–303). The suggestion is that Plato simply didn’t consider the possibility of letting the abstractness hold only at the level of content. But of course it may be instead that Plato (and later

12 For some puzzled queries regarding this conception of abstract entities, see the introduction to Burgess and Rosen (1997).

40

Representation and Objects of Thought in Medieval Philosophy

platonists) have seen such a possibility and rejected it as unworkable. We therefore need to consider just how viable an account Aquinas is offering. Abstraction separates off the matter in favor of the form. But this simple formula obscures more than it reveals, because Aquinas thinks that there are different kinds of forms, and also different kinds of matter. With respect to form, Aquinas believes that the ultimate goal of intellectual abstraction is to arrive at the substantial form, but that more often we arrive at one or another accidental form. As we will see below, he thinks that human beings never, or almost never, grasp the substantial form itself. This is roughly to say that the essences of things are hidden from us. The best we can do, generally, is to grasp hold of certain accidents that are closely connected to that essence, in the sense that they are reliable marks of the essence. Aquinas also distinguishes between different kinds of matter, and here the kind of matter that we abstract from depends on the sort of inquiry we are engaged in. As is well known, he divides matter into two kinds, signate matter and common matter, the former being the distinct matter possessed by a given individual, and the latter being matter conceived of apart from its individuating conditions within a given particular. When I spoke in the previous paragraph as if the substantial form were the essence, that was an oversimplification: in fact the essence is the substantial form plus the common matter.13 Aquinas also divides matter in another way, into sensible matter and intelligible matter. The former underlies sensible qualities – that is, those qualities that can bring about change in the five external senses (color, heat, hardness, etc.). The latter underlies quantity – that is, shape, size, and number.14 Since these divisions are orthogonal, they yield four kinds of matter: Signate sensible matter

Common sensible matter

Signate intelligible matter

Common intelligible matter

The difference across the first row, as Aquinas describes it, is that common sensible matter is flesh and bones, whereas signate sensible matter is this flesh and these bones. The latter always gets abstracted by intellect, inasmuch as that kind of matter holds only of individuals. Common sensible matter, however, is not always abstracted. For if the goal is to understand the nature of some material substance, then the intellect must reach a generalized understanding of the matter required for such a nature.

13 ‘[A]d naturam speciei pertinet id quod significat definitio. Definitio autem in rebus naturalibus non significat formam tantum, sed formam et materiam. Unde materia est pars speciei in rebus naturalibus: non quidem materia signata, quae est principium individuationis, sed materia communis’ (ST, 1a 75.4c). 14 ‘Materia enim sensibilis dicitur materia corporalis secundum quod subiacet qualitatibus sensibilibus, scilicet calido et frigido, duro et molli, et huiusmodi. Materia vero intelligibilis dicitur substantia secundum quod subiacet quantitati. … ut numeri et dimensiones, et figurae, quae sunt terminationes quantitatum …’ (ST, 1a 85.1 ad. 2).

Abstract Truth in Thomas Aquinas

41

Natural entities are understood through abstraction from individual [i.e., signate] matter but not through abstraction from sensible matter entirely. For a human being is understood as made up of flesh and bones, but understood through abstraction from this flesh and these bones. (InDA, III.8.231–6)

There is some question about just what is being abstracted away in the move from this flesh to flesh in general. Often, he says that to grasp the singular is to grasp it as here and now, hic et nunc (e.g., ST, 1a 57.2c). This phrase must refer to more than just time and place, since the abstract understanding of a common nature will abstract from many other things as well. In using this phrase hic et nunc, Aquinas suggests a demonstrative grasp of the thing right in front of us, in all (or at least many) of its particular characteristics. In contrast, if we are aiming at a general understanding of human nature, then we will need to know the kind of body that human beings have, but not the particular color or smell of a given body. This first kind of abstraction, the abstraction of the natural philosopher, thus involves separating out the qualitative aspects of a material substance that are inessential to its nature. This is to abstract away signate sensible matter. Such abstraction also involves separating out the signate intelligible matter, since the natural philosopher will abstract away various accidental quantitative aspects, such as being six feet tall or fat.15 A second kind of abstraction, mathematical abstraction, goes one step further. It abstracts not only from both kinds of signate matter but also from common sensible matter. In explaining the difference, Aquinas uses Aristotle’s example of the difference between the concept of snub and the concept of curved. To have the concept of snub one has to have the concept of nose, or so Aristotle had insisted. As Aquinas puts it, ‘sensible matter – specifically, a nose – falls within the definition of snub’ (InDA, III.12.282–3). This is to say that snub is a natural concept, along the same lines as human being, vertebrate, or feathered. It cannot be understood apart from some reference to matter. The concept of curved, in contrast, can be abstracted not only from any determinate matter, but also from even a generalized conception of sensible matter. To have the concept of curved, all one needs is the general concept of extended stuff. [Q]uantities such as numbers and dimensions, and also shapes (which are the limits of quantities) can be considered without their sensible qualities, which is for them to be abstracted from sensible matter. But they cannot be considered without understanding a substance underlying the quantity, which would be for them to be abstracted from common intelligible matter. Still, they can be considered without this or that substance, which is for them to be abstracted from individual intelligible matter. (ST ,1a 85.1 ad. 2)

15 Oddly, Aquinas often refers exclusively to signate sensible matter in this connection: see ST, 1a 85.1 ad. 2; InDA, III.8.222–38; InDA, III.12.277–97. In contrast, De ente, 2.75–84 describes signate matter as matter ‘considered under determinate dimensions,’ and remarks that the definition of homo includes non-signate matter – evidently, matter not considered under determinate dimensions.

42

Representation and Objects of Thought in Medieval Philosophy

Mathematical concepts are therefore in a way material concepts, in as much as they cannot be grasped without a conception of extended matter. This account of mathematical concepts seems open to quite a few lines of objection, of which I’ll briefly consider two. First, it is not clear that it holds for number, even if it holds for shape and size. One can count angels, presumably, without making reference to matter. Aquinas grants that one can count the angels, and even the divine persons. He insists, however, that these claims are based on a concept of number that is metaphysical rather than mathematical. Our mathematical concept of number is based on the concept of a continuum, which in turn presupposes the concept of extension and therefore matter. Accordingly, the claim that number is a material concept rests on a further claim that cannot be evaluated here, that we have two distinct number concepts.16 Second, one might think that the concept of curved and the like can be understood as grounded on the concept of space, rather than the concept of matter. Now, from a physical standpoint, Aquinas denies that there can be space apart from matter – that is, he denies that there can be a vacuum.17 Still, one might insist that it is possible to have the idea of space without matter, and so possible to conceive of curved and the like independently of extended matter. Indeed, one might think that this is not just a possible but indeed the actual way in which we conceive of geometrical concepts. I think Aquinas must simply deny this. In the passage just quoted, he remarks that such concepts ‘cannot be considered without understanding a substance underlying the quantity.’ This is to say, I take it, that the concept of curved makes no sense apart from the concept of something’s being curved. But space, as Aquinas understands it, is not a thing all by itself. Hence empty space cannot be conceived of as curved. If it seems to you that you are conceiving of a curve in space, Aquinas would contend that you must be conceiving of that space as filled in with some thing that serves as the substance underlying the curve. Indeed, for Aquinas it seems that the concept of space is just one more concept that can be formed only through mathematical abstraction. A third and final kind of abstraction occurs in metaphysics, when we abstract away from all four kinds of matter and consider concepts such as being, one, truth, potentiality, and actuality. Such concepts can be understood apart from all matter; indeed, Aquinas supposes that such concepts actually are instantiated in immaterial substances.18 So conceived, metaphysics turns out to come after physics in the sense that it involves one further step in abstraction, beyond what is required for mathematics.

16 See ST, 1a 11.4 ad. 2, 30.3c, and 50.3 ad. 1. For discussion, see Klima (2000). 17 See InPhys, 4.12–14. 18 ‘Quaedam vero sunt quae possunt abstrahi etiam a materia intelligibili communi, sicut ens, unum, potentia et actus, et alia huiusmodi, quae etiam esse possunt absque omni materia, ut patet in substantiis immaterialibus’ (ST, 1a 85.1 ad. 2). For the case of truth, see I Sent., 19.5.1c. That same passage also discusses the case of time, which seems to fall into the class of mathematical concepts, inasmuch as time has underlying it the concept of motion.

Abstract Truth in Thomas Aquinas

43

This account of the varieties of abstraction raises many more questions than it answers, and I will address some of them shortly. But there is one attractive aspect of the theory that should be stressed immediately: the way it offers a coherent account of how human intellectual inquiry is structured. At the front end of Aquinas’s story, there is the empiricism that leads him to remark that ‘the whole of the intellect’s cognition is derived from the senses’ (InDT, 1.3c). The subsequent contribution of intellect turns out to be the stripping away of successive layers of sensory information: first the particular qualities and quantities of an object, next all qualities even in general terms, and finally even the very notion of extension, so that nothing remains that is distinctively material about the information. To a considerable extent, the unity of this account confirms the methodological claims made at the end of §1. Though Aquinas’s identification of material natures as the objects of intellect does not yield a substantive theory of the intellect’s nature, it does yield a comprehensive theory of how the intellect operates, through abstraction. What begins as an account of how the intellect manages to conceive universally of the material world develops into a comprehensive story about all human thought, even in areas like mathematics and metaphysics.19 By nature, our minds are designed to grasp the material world in its essential respects. Inevitably, the method we deploy there is the method we deploy everywhere, even in cases where we attempt to go well beyond our native field of inquiry. In this way, Aquinas’s account of the intellect’s proper object provides the key to his whole theory of intellectual cognition. Interestingly, the theory itself exemplifies the process it describes, in as much as the kinds of matter at issue are themselves abstractions. What there are in the physical world are composite substances. We can talk about them as composed of form and matter, but – with the one exception of the human soul – these are conceptual parts, not substantial parts (like a kidney or a hand) that might exist on their own. A thing’s ‘signate sensible matter’ is a part of it only in an abstract conceptual sense – that is, we form the concept of a thing’s particular sensible qualities by abstracting from its other characteristics. And once we’ve identified that aspect of a thing, by abstraction, we can invert our focus and decide to abstract away from the aspect that we’ve just singled out. In short, to talk of matter – whether it be signate or common matter, sensible or intelligible matter, or even prime matter – is already to engage in abstraction.20 There are obvious and well-known difficulties with this conception of human understanding. One of the best publicized objections is that of Berkeley, who sardonically remarked that ‘he who is not a perfect stranger to the writings and disputes of philosophers must needs acknowledge that no small part of them are spent 19 Aquinas extends the story to theology as well, remarking for instance that ‘incorporea, quorum non sunt phantasmata, cognoscuntur a nobis per comparationem ad corpora sensibilia’ (ST, 1a 84.7 ad. 3). 20 In Pasnau (2002), Part One, I argue that this understanding of the form–matter distinction is crucial for understanding Aquinas’s conception of how body and soul make up a single unified thing.

44

Representation and Objects of Thought in Medieval Philosophy

about abstract ideas’ (Principles, Intro. §6). Berkeley takes this to be disastrous, of course, arguing that one cannot have an abstract idea of, say, human being in general, only the idea of a white one, a black one, a tawny one, etc. (ibid., §10). Aquinas has a quick reply to this, which is that Berkeley is confusing imagination and thought.21 Aquinas can grant that imagination always concerns determinate particulars, and Aquinas himself insists that thoughts are always accompanied by imagination – this is what he calls the turn toward phantasms. Berkeley’s error, from Aquinas’s perspective, is to treat this constant stream of images as if they were the content of our thoughts. But for Aquinas, of course, intellectual thought is entirely distinct from imagination, and does not consist in sensory images at all. To decide whether you can have the abstract thought of a human being, don’t try to form a mental picture of such a thing – that’s the wrong level, the level of phantasms. Instead, ask yourself whether you understand what I am talking about when I talk about the human species, and whether you can do so without either of us making reference to any particular member of that species. If you understand my words in this way, then you have the concept. If we allow ourselves to accept the very idea of abstract ideas, then the next worry that naturally arises is how the intellect manages to make the transition from concrete sensory impressions to abstract ideas such as human being, curved, or truth. How, in other words, does intellect manage to separate out what is material and accidental, and focus on the thing’s essence? Here one might immediately raise questions about whether there are such things as essences at all. But I want to set aside that question until §4, and focus on the problem of how the intellect knows what to abstract and what not to abstract. It must be said that Aquinas does not have anything very illuminating to say about this problem. He says things like the following: And this is to abstract the universal from the particular, or an intelligible species from phantasms: to consider the nature of the species without considering the individual principles that are represented by the phantasms. (ST, 1a 85.1 ad. 1)

Here, and elsewhere, abstraction is simply a process of selective attention, whereby the agent intellect focuses on one thing (the form alone) and brackets off the rest (the particular material conditions). This capacity of agent intellect seems to be, for Aquinas, entirely primitive, in the sense that it cannot be analyzed or explained any further. It is just something we are able to do. Peter Geach has argued that Aquinas does not believe in abstraction, when that is understood as ‘the doctrine that a concept is acquired by a process of singling out in attention some one feature given in direct experience – abstracting it – and ignoring the other features simultaneously given – abstracting from them’ (1957, p. 18). It 21 In fact Berkeley does go back and forth between talk of imagining and talk of conceiving, e.g. in §10: ‘I have a faculty of imagining, or representing to myself. … [W]hatever hand or eye I imagine. … [T]he idea of man that I frame to myself. … I cannot by any effort of thought conceive the abstract idea above described.’

Abstract Truth in Thomas Aquinas

45

seems fairly clear, however, that this is Aquinas’s view, more or less. Certainly he does think of abstraction as a matter of selective attention, singling out some one feature while ignoring others. Now one might wonder whether the agent intellect singles out a feature ‘given in direct experience,’ as Geach puts it here. Obviously, there is a sense in which the concepts acquired are not sensory concepts at all. Still, Aquinas does insist that the information singled out by agent intellect is information that is somehow present at the sensory level. He says this fairly explicitly: If the senses apprehended only that which belongs to the particular, and in no way were also to apprehend the universal nature in the particular, then it would not be possible for a universal cognition to be caused in us by a sensory apprehension. (InPA, II.20.266–71)

It seems fair to say, then, that what the agent intellect abstracts is ‘given in direct experience,’ though of course the senses are incapable of recognizing it. Geach reaches his conclusion not through textual evidence but through the philosophical conviction that such an account would be utterly unworkable.22 That, combined with the principle of charity, gets him his interpretive conclusion. Now this may not be a model of how to go about studying the history of philosophy. But it seems to me nevertheless that there is something right in what Geach is claiming. What is right is that Aquinas did not think abstraction, all by itself, could contribute very much to the intellect’s ultimate goal of grasping the natures of material things. It is not as if the agent intellect encounters the phantasm of a dog and immediately singles out what it is to be a dog, the essence of doghood. This happens, if it happens at all, only at the end of a long process that involves the agent intellect, the possible intellect, and a continual recurrence to phantasms. Abstraction is just one small part of this process, the part that gets us from images of particulars to general concepts. Initially, those general concepts will be so general and uninteresting as to be of virtually no value – these will be concepts like being, goodness, and truth. (It is a mistake to suppose that the generality of these concepts makes them interesting.) In time we manage to reach an increasingly fine-grained understanding of the world, but we do so through reasoning, not through the brute force of abstraction. All of this makes it easier to accept that Aquinas conceives of abstraction simply in terms of selective attention. Though it is disappointing to learn that he has nothing more to offer, we can take solace in the fact that abstraction isn’t doing nearly as much work as one might suppose. This is not to say that an account of intellect’s operations can leave abstraction on the sidelines. Even if abstraction does not immediately yield a clear and distinct understanding of the natures of material things, through the magic of agent intellect, nevertheless the ongoing intellectual activity that aims at that goal is continuously an attempt at abstraction. For human beings, in large part, thinking is abstracting.23 The goal, always, is to single out those 22 For another negative assessment of the workability of abstraction, see King (1994). I take what follows in the main text to be a reply to King as well as to Geach. 23 I do not mean to suggest that this is all thinking consists in; I am setting aside the subsequent intellectual operations of composition–division and reasoning. These processes

46

Representation and Objects of Thought in Medieval Philosophy

aspects of the world around us that are essential, while bracketing off the rest. As we learn more about the world, we become increasingly good at focusing on what is essential, we learn how to see things, as it were, which is to say that we learn to recognize what things truly are, as opposed to what is accidental in them. Aquinas considers the objection that if the agent intellect were a power of the human soul, then we would all understand everything right away (ST, 1a 79.4 obj. 3). He replies that this is not so, partly because we need the right phantasms, and partly because we need ‘practice in activity of this sort – for one thing grasped by intellect leads to others, as terms lead to propositions and first principles lead to conclusions’ (ad. 3). This is to say that the agent intellect, in concert with the possible intellect and the senses, builds up a capacity to understand the world as it is. Abstraction does not happen by magic (or by divine illumination), but in the end the result of human thought is an abstract conception of the world around us.24 3 Abstract Truth Once we set aside the problem of how the intellect manages to form abstract concepts of the world, we can focus on the question of what these concepts tell us about the world. More specifically, there is a question of whether abstract thoughts can be true. For the platonist, there is of course no worry here, because abstract thoughts can straightforwardly refer to abstract objects. But Aquinas, as we have seen, rather quickly dismisses platonic objects of all sorts, replacing abstract entities with abstract thoughts. The obvious worry is that there is nothing for these thoughts to refer to. This worry is an old one, and Aquinas is well aware of it. In considering whether the intellect does indeed form its concepts through abstraction, he considers the following as the very first objection: Any intellect that understands a thing otherwise than it is is false. But the forms of material things do not exist abstracted from the particulars that phantasms are likenesses of. Therefore if we understand material things by abstracting species from phantasms, there will be falseness in our intellect. (ST, 1a 85.1 obj. 1)

Aquinas’s solution is familiar to the point of tedium. I will quote it nevertheless: Abstracting takes place in two ways. First, by way of composition and division, as when we understand something not to be in another or to be separated from it. Second, by way of a simple and absolute consideration, as when we understand one thing while not considering the other at all. So to abstract through intellect things that are not abstract in reality – by abstracting in the first way – is not without falseness. But there is nothing false

strike me as much less mysterious. 24 I discuss some of these issues at more length in Pasnau (2002), ch. 10.

Abstract Truth in Thomas Aquinas

47

in the second way of abstracting through intellect things that are not abstract in reality. (ibid., ad. 1)25

The passage goes on to give the example of an apple. One can understand the apple’s color while understanding nothing about the nature of the apple, and there is nothing false in doing so – just so long as one does not form the judgment that the apple is not red, or that red can exist outside of matter. This is at best a first step toward solving the real problem here, and it is not a terribly illuminating step. The first kind of abstraction described above involves the formation of a proposition, the understanding that something is (not) the case. Here, Aquinas readily concedes, we may well go wrong. The second kind of abstraction results in the simple understanding of some form in the abstract. The point, one might suppose, is that this kind of simple abstraction cannot be false, because the result is not a proposition at all, and so not the sort of thing that can be false. But this is not what Aquinas means, because ‘false’ on his usage can bear the extended sense of nonveridical, and hence can apply to any kind of cognition that does not depict the world as it is. So Aquinas’s point here must be that simple abstraction does not get the world wrong in this way. Other passages supply some guidance as to how this is so. He remarks that ‘the truth of an apprehension does not require that when one apprehends some thing one apprehends all the things in it’ (InDA, II.12.131– 4). Since abstraction just is the process of selectively apprehending certain aspects of a thing, it is not necessarily false. Of course, it is possible for one’s efforts at abstraction to fail. He remarks, of things that are conjoined in reality one can be understood without the other, and truly, as long as it is not the case that one of them is included in the other’s nature (ratio). For if Socrates is musical and white, I can understand whiteness while understanding nothing about music. On the other hand, I cannot understand human being while understanding nothing about animal, since animal is included in the nature of human being. (InDA, III.12.261–8)

So simple abstraction is false only if it abstracts in a way that cuts across the nature of what is being abstracted, leading to the formation of a fragmented and incoherent idea. At this point, one might once again want to raise the question of how the intellect manages to cut the world up in the right way – of why, in other words, such abstraction isn’t false all the time, given the difficulty involved in grasping the true natures of things. If the claims of the previous section were correct, then Aquinas 25 The specific formulation of the problem that Aquinas considers in 85.1, as well as his reply in ad 1, comes straight from Boethius, Second Commentary On Porphyry’s Isagoge, I.10–11 (trans. Spade, pp. 23–4). Aristotle had already considered much the same objection, in the context of mathematical ideas (Phys., II.2 193b33–7), and Aquinas’s commentary on that passage is thorough and interesting (InPhys., II.3.158–62). For other passages in Aquinas, see ST, 1a 13.12 ad. 3, InDA, III.8.239–63, InDA, II.12.116–39, InDA, III.12.261–97, InDT, 5.3.

48

Representation and Objects of Thought in Medieval Philosophy

should grant that the road to success here is long and difficult. Yet if that is right, then why do these passages seem so optimistic about our ability to grasp the truth through abstraction? This is a question I will return to at the end of the next section. For now I want to focus on why the explanation offered here, even if accepted, falls well short of being satisfactory. The heart of the problem is that Aquinas is willing to concede the following: [T]he object of our intellect is not something existing outside sensible things, as the Platonists claimed, but something existing in sensible things, although the intellect apprehends the quiddities of things differently from the way they are in sensible things. (InDA, III.8.242–7)

Within sensible things, these quiddities are of course particular, whereas within intellect they are universal. The concern, then, is that abstraction doesn’t merely involve a selective focus on certain aspects of reality, but that it in fact distorts reality, by depicting it in a way it is not. This last result is unacceptable to Aquinas. He remarks, for instance, that ‘the conceptions of intellects are a kind of likeness of the things understood. If, however, the conception of intellect were not made to be like the thing, then that conception of the thing would be false’ (De 108 articulis, prologue 9–12). Elsewhere, rather than speak of a likeness, he uses the language of correspondence: ‘The nature (ratio) is said to be in the thing inasmuch as there is something in the thing outside the soul that corresponds (respondet) to the conception of the soul’ (I Sent., 2.1.3c). Hence Aquinas must deny that the intellect depicts reality in a way it is not. Abstraction must not distort reality; it must only select out certain portions of reality. And indeed Aquinas takes great pains to stress that this is precisely what happens. After conceding, in the above passage, that ‘the intellect apprehends the quiddities of things differently from the way they are in sensible things,’ he immediately continues: For it does not apprehend them with the individuating conditions that are adjoined to them in sensible things. And the intellect can achieve this without any falseness, for nothing prevents one of two things conjoined to each other from being understood without the other’s being understood. (InDA, III.8.247–52)

So again we get the line that abstraction is simply a matter of selective attention, along with the idea that the quiddity conceived by intellect really is within the thing, but is there along with an accretion of individuating, material conditions. Elsewhere we get much the same picture: Although the nature of a genus and a species exists only in particular individuals, nevertheless the intellect understands the nature of the species and the genus without understanding the individuating principles, and this is to understand universals. Thus it is not a contradiction that universals do not subsist outside the soul, and that the intellect, in understanding universals, understands things that are outside the soul. (SCG, II.75.1551)

Abstract Truth in Thomas Aquinas

49

What Aquinas seems to insist on – what it seems he must insist on – is that the nature of a species exists within each member of that species, as a kind of essential kernel covered up by the shell of material conditions that distinguish each individual.

Evidently, this same account must hold not just at the species level, but at each higher genera, so that there will be kernel within kernel, as successive layers of differentiae are stripped away.

In this way, the accidental differences of a substance cover up layer after layer of formal structure, each layer qualitatively identical to a corresponding layer in other substances of that kind. If material substances were not structured like this, then it seems that the process of abstraction would distort reality, imposing a structure that is not there, fudging ineliminable differences between individuals. There is a considerable obstacle to this interpretation of Aquinas, for he is adamant that the logical framework of species and genus does not track a distinction between forms within a substance. His controversial unitarian stance regarding substantial form leads him to reject the pluralist account according to which a human being is rational in virtue of one form, sensory in virtue of another, living in virtue of a third, etc. Thus he remarks, ‘the distinction of species and genus does not require a real distinction between forms, but only an intelligible distinction’ (QDSC 3 ad 3). The insistence that there is no real distinction to be had here seems to undermine the layered structure described in the previous paragraph. The problem is most obvious at the genus level, since Aquinas is so insistent that all the genera fall out of the one substantial form that gives a thing its species.26 But the same problem arises 26 ‘[I]ta nec per aliam animam Socrates est homo, et per aliam animal, sed per unam et eandem’ (ST, 1a 76.3c).

50

Representation and Objects of Thought in Medieval Philosophy

at the species level, since Aquinas does not believe that two members of the same species share forms that are exactly alike.27 Even here, the classificatory scheme of genus and species does not track the ontological structure that Aquinas ascribes to the world. The following passage explicitly confronts this objection: It is not required that one treat diversity among natural things in terms of the diverse accounts (rationes) or logical conceptions (intentiones) that result from how one understands them. For reason can grasp one and the same thing in different ways. So, as was said, the intellective soul virtually contains whatever the sensory soul has, and more still. It follows, then, that reason can consider separately that which involves the power of the sensory soul – taken as something incomplete and material. And because reason finds this to be common to humans and other animals, it forms on this basis an account of the genus. Meanwhile, reason takes that in which the intellective soul exceeds the sensory as something formal and perfecting, and on that basis it forms the differentia of human being. (ST, 1a 76.3 ad. 4; see also QDSC, 3c)

The framework of logical conceptions, which just is the genus–species framework of the Porphyrian tree, need not correspond to any real differences within things. Even if there is no difference at the level of form between rational and sensory, still ‘reason can consider separately’ one or the other. Regrettably, the passage is not as forthcoming as it might be about the real difficulty that looms. One might wish that, here at least, Aquinas had resisted his chronic inclination to make matters appear more smooth and seamless than they really are. For if he is simply going to grant that our logical conceptions do not correspond to the formal structure of substances, then he needs to give us some account of why these logical conceptions describe the world as it really is. He tells us here that ‘because reason finds this [sensory capacity] to be common to humans and other animals, it forms on this basis an account of the genus.’ But he does not address the crucial question of what such commonality amounts to. There is a reason why Aquinas cannot say much more at this point. Earlier, at the end of §1, we saw how his theory of intellect goes from objects to acts, but then falls relatively silent at the level of capacities. That silence grows still deeper when Aquinas moves from the soul’s capacities to its very essence. When it comes to the question of what my soul has in common with the soul of a bear, Aquinas cannot give a very full account. He thinks he can rule out the view that the bear and I both have a sensory soul, and that I have an additional rational soul. Rather, the bear has its substantial form, and I have mine, and they have some common features. What features, exactly? In answering this question, Aquinas is forced to descend to the

27 ‘Manifestum est enim quod quanto corpus est melius dispositum, tanto meliorem sortitur animam, quod manifeste apparet in his quae sunt secundum speciem diversa. Cuius ratio est quia actus et forma recipitur in materia secundum materiae capacitatem. Unde cum etiam in hominibus quidam habeant corpus melius dispositum, sortiuntur animam maioris virtutis in intelligendo’ (ST, 1a 85.7c).

Abstract Truth in Thomas Aquinas

51

level of operations. The capacities for sensation and nutrition are common to me and to a bear; since Aquinas takes these powers to flow from the soul, there would seem to be some sort of commonality within soul. But possession of these capacities is just a sign of our sharing a genus. The reason we share a genus is that we have something common within our souls that makes it the case that we share these relatively superficial operations.28 Lacking good information about the essences of things, Aquinas must do the best he can. His position is analogous to that of a biologist who aspires to appeal to differences in DNA, but who is forced (for now) to settle for gross anatomical differences. Yet Aquinas’s position is worse, because whereas we now know what it means for two DNA strands to have more or less in common, Aquinas has no account of how two essences might have something in common. And though his language suggests the kind of structured view I described above, it is not at all clear that a soul – that is, the soul’s essence as distinct from its capacities – can have that sort of structure. He holds that the soul is ‘one and simple in essence’ (QDA, 10 ad. 17). This is difficult to square with the suggestion that the natures of material substances contain kernels within kernels, ready to be abstracted away by intellect. Rather than speculate further on how essences might or might not be structured, it seems better to conclude that the method of abstraction is not suited to go that far. We can use abstraction to form a conception of the genus animal, by focusing on the operations and hence capacities that all animals share. On that basis, we conclude that my essence has something in common with the essence of a bear. But at that level we have no real purchase on what commonality consists of, because we have absolutely no conception of what it would mean for two essences to have something in common. We can say that they give rise to common capacities and operations. But we do not understand essences well enough to perform any sort of abstraction directly on that level. Since abstraction is the mode in which our intellect functions, it truly is the case, as Aquinas regularly says, that the essences of things are unknown to us.29 In all of this, I have been taking for granted that common natures are not literally shared. In every case, the commonality in question is merely qualitative. On this point, Aquinas is quite definite: there is mere resemblance between particulars without any sort of numerical identity. It is not necessary that, if this is a human being and that is a human being, there is numerically the same humanity for each – just as in the case of two white things there is not numerically the same whiteness. Instead, what is necessary is that this one is like 28 ‘[R]ationale et sensibile, prout sunt differentiae, non sumuntur a potentiis sensus et rationis, sed ab ipsa anima sensitiva et rationali. Quia tamen formae substantiales, quae secundum se sunt nobis ignotae, innotescunt per accidentia, nihil prohibet interdum accidentia loco differentiarum substantialium poni’ (ST, 1a 77.1 ad. 7). 29 See, e.g., InDA, I.1.254–5, I Sent., 25.1.1 ad. 8, II Sent., 3.1.6c, QDV, 4.1 ad. 8, 10.1c, 10.1 ad. 6, InMet, VII.2.1277, VII.12.1552. I will return to this issue at the end of §4. For further discussion, see Pasnau (2002) §5.5.

52

Representation and Objects of Thought in Medieval Philosophy that one in that it has humanity just as the other does. It is the intellect, in taking up humanity not inasmuch as it belongs to this one, but as it is humanity, that forms an intention common to all. (II Sent., 17.1.1c)

Thus, as he often says, universals exist only within intellect; everything outside the mind is particular.30 There is, of course, unending controversy over whether one can solve the problem of universals without having recourse to universals in re. For present purposes, we might focus on two particular issues. First, because he settles for natures that are merely qualitatively alike, Aquinas seems to have no way of explaining what such likeness consists in. (For the realist, in contrast, two things can be alike in virtue of literally sharing some form or property.) Aquinas does hold that all similarity arises from agreement at the level of form,31 but in the present context that doesn’t help, because the question remains of what it means for two particulars to agree in form. This outcome does not seem to trouble Aquinas, however, because he shows no signs of supposing that there is anything more to be said about what agreement amounts to in such cases. The concept seems to be a primitive one. A second issue, more urgent in the present context, is that we again seem to face the problem of truth, because it seems once again that the intellect is not representing things as they are. In the mind, natures are conceived of as universal, but in reality they are always particular. Up until now, the strategy has been to show how this particularity could be stripped away, revealing a common nature. But, as we’ve just seen, the most this strategy will yield is a nature that is qualitatively common. Within intellect, however, the nature is universal. So we seem once again to have violated the principle that there be ‘something in the thing outside the soul that corresponds to the conception of the soul’ (I Sent., 2.1.3c). It seems to me that Aquinas has a reply analogous to the earlier strategy. Very often, he describes universality as something that the intellect adds to the concept, something that he calls an ‘intention of universality.’ In a useful passage from near the start of his Sentences Commentary, he describes three ways in which names signify things in the world. In one way, what is signified exists ‘outside the soul according to its complete existence,’ such as a

30 ‘Sic igitur patet quod naturae communi non potest attribui intentio universalitatis nisi secundum esse quod habet in intellectu: sic enim solum est unum de multis, prout intelligitur praeter principia quibus unum in multa dividitur. Unde relinquitur quod universalia secundum quod sunt universalia non sunt nisi in anima, ipsae autem naturae quibus accidit intentio universalitatis sunt in rebus’ (InDA, II.12.139–47). ‘[I]n Sorte non invenitur communitas aliqua, sed quicquid est in eo est individuatum’ (De ente, 3.80–82). 31 ‘Similitudo autem inter aliqua duo est secundum convenientiam in forma’ (QDV, 8.8c). It is tempting to say that likeness consists in sharing the same form, but we have just seen why that is not the case. The bear and I are similar with respect to being animals, but not in virtue of sharing animality or any other form. In some contexts, Aquinas follows Aristotle in reserving ‘likeness’ for cases of qualitative resemblance, and speaks of sameness between members of the same species or genus (see, e.g., InMet., X.4.1999). In the present context, nothing much seems to rest on this usage.

Abstract Truth in Thomas Aquinas

53

particular human being. In another way, what is signified has no existence outside the soul, like a chimera. Finally, there are some that have a foundation in a thing outside the soul, but the completion of their ratio, with respect to what is formal, is through the soul’s operation – as is clear in the case of universals. For humanity is something in a thing, but it does not there have the ratio of the universal, since there is no humanity outside the soul that is common to many. But in virtue of its being taken up in intellect, it has adjoined to it, through the operation of intellect, an intention, in virtue of which it is said to be a species. (I Sent., 19.5.1c)

This passage illustrates all the main strands of Aquinas’s account. First, ‘humanity is something in a thing,’ which is to say that the intellect’s abstract conception of a common nature does directly correspond to something in the world. Second, the humanity that exists in things is particular, because ‘there is no humanity outside the soul that is common to many.’ Third, the intellect’s understanding of humanity does not distort that concept but simply adds something to it: the concept has ‘adjoined to it … an intention … .’32 The point, I take it, is that the content of the thought exactly corresponds to the nature as it is in the world, within particulars. But the intellect attaches to this content something akin to a propositional attitude: it applies the concept universally to all members of the species. Just as in the statement ‘She fears her father’ we can distinguish between the attitude (fears) and the content (her father), so in ‘She understands humanity’ we can distinguish between the content (humanity) and the attitude (understands), which involves applying the concept generally to all particular instances. In this way, just as within particulars it is possible to distinguish a common kernel that corresponds to the abstract concept, so within intellect it is possible to distinguish the abstract concept from the intention of universality. This description of an abstract concept just is what Aquinas refers to as an ‘absolute consideration,’ with respect to which he famously remarks, ‘if it were asked whether the nature so considered can be called one or plural, neither reply ought to be granted, for each is outside the understanding of humanity, and each can accrue (accidere) to it’ (De ente, 3.37–40). On the account just presented, a nature considered absolutely is not some sort of mysterious platonic entity, separated from particulars. On the contrary, such natures never exist in separation from particulars – they exist absolutely both in intellect and in the world, and this is what, for Aquinas, guarantees the correspondence between abstract thought and reality. When instantiated in intellect or in material things, such natures are always made ‘one or plural’ (that is, universal or particular). But this is something added on, something that ‘can accrue to it.’ Underneath, the nature remains the same.

32 For other discussions of this ‘intention of universality,’ see ST, 1a 85.2 ad. 2 and InDA, II.12.96–151.

54

Representation and Objects of Thought in Medieval Philosophy

4 Sentential Truth To insure the truth of abstract thought, Aquinas has to insure a correspondence between concepts and reality. Yet this alone is not adequate as a general solution to the problem of abstract truth, because at best it deals with abstract concepts conceived in isolation. We further need an account of how such concepts work within sentences – that is, we need an account of sentential truth. Consider the sentences Anna is a human being. Tommy is a human being. On one reading, available to the platonist, these sentences claim that there is a property, humanity, that Anna and Tommy participate in. On another reading, favored by some nominalists, the sentences say that Anna and Tommy belong to the set of all human beings. These analyses are both extrinsic, as we might put it, in as much as they both appeal to something outside of the subject to which the subject is somehow related. Aquinas, in contrast, wants to give such sentences an intrinsic analysis, in as much as he takes them to be true in virtue of the subjects’ possessing a certain property. As natural as this approach is, it leaves Aquinas with an obvious difficulty. If it is true that Anna is a human being, and true that Tommy is a human being, then it seems that there is a property, call it humanity, that both Anna and Tommy possess. But, as we have seen, Aquinas denies that any properties are wholly possessed by more than one individual.33 Now there is no one, I trust, who will insist that these sorts of linguistic data decide metaphysical questions. Ordinary language may suggest the existence of universals in re, but there may be other analyses on which that suggestion is defused. Still, it seems very much incumbent on Aquinas to provide such an analysis. If it is true that Anna and Tommy are both human beings, and if this is true by virtue of their each having a property, humanity, then how can we avoid the implication that they each have the very same property? Aquinas has available to him a sophisticated means of handling this question, in as much as he distinguishes between two linguistic functions, signification and supposition. Very roughly, the supposition of a term is its reference, whereas its signification is its meaning. In the sentence Anna is a human being, the predicate and the subject both supposit for a person, my daughter. But ‘human being’ signifies the concept of humanity. It is in virtue of this signification that it supposits as it does, for things that have human nature. Thus ‘a name, properly speaking, is said to signify the form or quality based on which the name is imposed, and is said to supposit for that which it is imposed on’ (III Sent., 6.1.3c).34 On this account, the sentence Anna is a 33 Properties might be shared in various ways – e.g., if they are relational. But no property can be wholly possessed by one individual and wholly possessed by another. Thus he remarks, as quoted above, ‘there is no humanity outside the soul that is common to many’ (I Sent., 19.5.1c) 34 ‘Et nomen, proprie loquendo, dicitur significare formam sive qualitatem a qua imponitur nomen; dicitur vero supponere pro eo cui imponitur.’ See also ST, 1a 39.4 ad. 3

Abstract Truth in Thomas Aquinas

55

human being comes out true in the most straightforward way, in virtue of the subject and the predicate being coreferential.35 Sentential truth requires no abstract ontology. But the sentence nevertheless signifies something universal, which is why we can meaningfully say both that Anna is a human being and that Tommy is a human being. There is no equivocation in predicating ‘human being’ of several individuals, because the term has the same meaning (signification) in each case: Equivocation is brought on by a different form’s being signified by the term, not by a difference in supposition. For the term ‘human being’ is not used equivocally just because it supposits sometimes for Plato and sometimes for Socrates. So the term ‘human being,’ said of Christ and of other human beings, always signifies the same form: human nature. Thus it is univocally predicated of them. (SCG, IV.49.3847 [12])

Once again, Aquinas thinks he can have it both ways: a robust account of abstract truth, grounded on an ontology that is essentially nominalist.36 At the core of this semantic conception is Aristotle’s claim that spoken words refer to mental concepts or ‘passions of the soul,’ on the rather opaque medieval translation of De interpretatione, 1. Aquinas comments on this passage as follows: ‘Passions of the soul’ must here mean the conceptions of intellect that nouns and verbs and statements signify, according to the claim of Aristotle. For it cannot be that they immediately signify the things themselves, as is evident from their mode of signifying: for the term ‘human being’ signifies human nature in abstraction from singulars. Hence it cannot be that it immediately signifies a singular human being. Hence the Platonists held that it signifies the separate idea of Human Being. But since this, in virtue of its abstractness, does not subsist in reality, on Aristotle’s view, but exists in intellect alone, it was therefore necessary for Aristotle to say that spoken words immediately signify the conceptions of intellect and, through their mediation, signify things. (InPH, I.2.97–112)

The argument takes for granted that ‘human being’ signifies some one abstract thing. (If argument for this assumption were wanted, Aquinas could appeal to the point made in the previous paragraph: that otherwise the term would be equivocal across usages.) What could that one thing be? If we reject Platonic Forms, and we reject universals in re – a possibility that Aquinas characteristically doesn’t even deign to and III Sent., 6.1.2 ad. 4: ‘“Homo” significat humanam naturam, et supponit pro supposito subsistente in natura illa.’ 35 There is considerable controversy over whether this so-called identity theory of predication holds in all cases, for Aquinas, or even holds at all. Nothing in what follows rests on this debate. For some evidence in favor of an identity theory, at least in certain cases, see InMet., V.9.891, as well as the passage from SCG quoted in the main text at the end of the present paragraph. 36 The term ‘nominalist’ has a long and intriguing medieval history, and in the context of that story it would clearly be wrong to label Aquinas a nominalist. Yet in the modern context the label seems appropriate, inasmuch as ‘nominalism’ now standardly refers to any theory that rejects both platonic and in re universals. The common practice of referring to Aquinas as a conceptualist obfuscates the central issues.

56

Representation and Objects of Thought in Medieval Philosophy

mention – then the only remaining option is that ‘human being’ signifies a conception within intellect. Of course, more has to be said at this point, because we need to account for the apparent fact that both thought and language refer to things in the world. The last passage’s closing remark about how words indirectly signify things will look lame and unmotivated until it is coupled with the doctrine of abstraction considered in the previous section. Since conceptions within intellect really do correspond to the structure of reality, our language can refer to (supposit for) things in the world in virtue of signifying something within intellect. Universals, inasmuch as they are universal, exist only in the soul. But the natures to which the intention of universality applies exist in the world (in rebus). For this reason, the common names signifying those natures are predicated of individuals. (InDA, II.12.139– 44)

The universal meaning of words, combined with the correspondence between universal natures and natures in re, explains how we can use abstract language to talk about particular things. The following passage (quoted earlier in part) makes these connections utterly explicit: The nature (ratio) of anything is what its name signifies – e.g., the nature of stone is what its name signifies. But names are signs of intellectual conceptions. Hence the nature of anything signified by a name is the conception of intellect that the name signifies. This conception of intellect exists within intellect as in a subject, and exists in the thing understood as in the thing represented: for the conceptions of intellects are certain likenesses of the things understood. If, however, the conception of intellect were not made to be like the thing, then that conception of the thing would be false – as if one were to understand to be a stone what is not a stone. Therefore the nature of the stone exists within intellect as in a subject, but exists in the stone as in that which causes the truth in the conception of the intellect that understands the stone to be such. (De 108 articulis, prologue 1–17)

Natures exist within intellect, but correspond to things in the world. This correspondence – a ‘likeness’ between concept and object of the sort described in §3 – is what insures the truth (veridicality) of our concepts, and by extension is what insures the truth of language. As for what insures the truth of this correspondence, the last sentence of the passage is revealing. The nature of stone exists within intellect, and also exists in stones in the world. In the world, it is ‘that which causes the truth in the conception of the intellect’ – that is, the conception within intellect is the end result of a cognitive process that began with a sensory perception of the stone, and culminated in an abstract representation of the nature of stone. Our reason for being confident that our concepts are true is that we can tell a causal story about how those concepts were generated by the things themselves. At this point I want to set aside the semantic story introduced in the last three paragraphs, and focus on one implication of the causal story just introduced. The semantic story must be set aside, given my purposes, because this is not a story to which

Abstract Truth in Thomas Aquinas

57

Aquinas himself made any great contribution. Medieval theories of signification and supposition are among the greatest philosophical accomplishments of the era, but Aquinas here is simply following the lead of others. A detailed discussion of these matters would very quickly take us into the work of thirteenth-century logicians like Peter of Spain and William of Sherwood, and here I will advert to the fine work that has already been done in this area.37 What I wish to discuss instead, and in closing, is a certain aspect of the causal connection between sensation and concept formation. In as much as Aquinas thinks that all causal relationships involve agreement in form, and that likeness too always consists in some sort of formal agreement, he has a quick way of getting from a causal account of our cognitive processes to conclusions about the likeness between concept and object. But though the path may be direct, a serious obstacle stands in the way. For even if Aquinas can describe a causal chain running from the senses all the way to intellect, he has to allow that what the senses apprehend are the superficial sensible qualities of things, not their essences or natures. Given this disparity between the sensory input and the intellectual end-product, the causal story alone is hardly enough to insure the necessary kind of correspondence. Something more needs to be said about how we get from sensible qualities to essences. We have already seen one version of this problem in the previous section, where abstraction turned out to be inapplicable at the level of essences. But the problem spills over into semantics, too, inasmuch as the theory of reference just described requires some kind of correspondence between concept and object. Aquinas is well aware of the problem, and offers an extremely interesting solution. Rather than insist that we all do somehow grasp the real natures of things, he grants (as we saw earlier) that we do not. All that we grasp, typically, are accidental features of the thing, and in such cases we use these accidents to fix the reference of the term. Since essential differentiae are unfamiliar (ignotae) to us, we sometimes use accidents or effects in their place, as is said in Metaphysics VIII [ch. 2], and in virtue of this we name the thing. In such a case, that which is taken in place of the essential differentia is the basis on which the name is imposed. (QDV, 4.1 ad. 8)38

Aquinas goes on to use the example of a stone (lapis). Following the fanciful lead of Isidore of Seville, he supposes that ‘lapis’ was imposed on the basis of one of the 37 See, in particular, the extensive discussion of semantics in Kretzmann et al. (1982), and Spade (1996). For Aquinas’s own semantic commitments, I am heavily indebted to Klima (1996) and Ashworth (1991). The latter is an excellent source for further references to thirteenth-century literature. For a useful discussion of Aquinas on supposition, see Schoot (1993). 38 Aquinas insists on this point partly because it puts him in a stronger position to explain the meaningfulness of talk about God. If it is generally the case that we manage to talk about things without knowing their essences, then this is not a special problem in the case of the divine names. See, e.g. ST, 1a 13.2 ad. 2, 13.8c and ad. 2, and 18.2c.

58

Representation and Objects of Thought in Medieval Philosophy

effects of stones, their hurting the foot (laedere pedem). In such a case, even though the accidents or effects fix the reference of the term, they are not the thing signified by the term. It is still the thing’s nature that the term signifies, even though it may be the case that no one who uses the term knows what that nature is. These considerations introduce an important modification into Aquinas’s semantic framework. According to that framework, as we have seen, (a) a term signifies the nature of the thing; (b) the term supposits for what has that nature; and (c) the supposition is fixed by the thing’s nature (see III Sent., 6.1.3c, as quoted earlier). It now turns out that (c) holds only in certain cases. It holds, he thinks, for sensible qualities, where on his view there is no gap between what the thing is and how we pick it out.39 But when we lack such direct acquaintance, as is the case with the natures of material substances, then (a) that which is signified comes apart from (c) that which picks out the supposition. ‘Sometimes, that on the basis of which the name is imposed to signify is different from what the name is imposed to signify’ (ST, 1a 13.2 ad. 2). The signification or meaning of the term remains the same in such cases, but because we do not have access to that meaning, it must be the case that something else – some accidental feature of the thing – serves to pick out the reference. In this way, Aquinas is able to free his semantic account from worries about how we grasp the essences of things. Even if we never do grasp such essences, still this is what our natural kind terms signify. Meaning, on this account, is divorced from historical facts about how the term was imposed or how language users in fact pick up on the term’s reference. But what then does determine the meaning (signification) of the term? To answer this question, we need to have in front of us a better example of the phenomenon than the one that Aquinas favors. I like to think that the case of the stone is meant to be funny – if not, it is hard to see what value it has as an example. Obviously, hurting the foot will not at all do as the sort of placeholder attribute that we use to fix the reference of a natural kind term. Not only are there many other things that can hurt the feet, but of course there are also some stones that feel rather nice to walk on. So whatever it is that we use to fix the reference of ‘stone,’ it cannot be something like hurting the foot. Even if this were plausible as an etymological claim, it would be worthless from the semantic point of view. Aquinas offers a more serious example, however, when he discusses the term ‘life.’ We base this term, he remarks at ST, 1a 18.2c, on ‘some sort of external appearance’ – that is, on a thing’s moving itself. But this kind of behavior is not what it is to live – to live is to exist with such a nature as to be capable of moving itself or (more generally) bringing itself to perform some operation. This is what the predicate ‘lives’ signifies, even though what fixes the reference of the term in practice is some sort of superficial behavior. It may be a 39 See ST, 1a 13.8c: ‘Si qua vero sunt quae secundum se sunt nota nobis, ut calor, frigus, albedo, et huiusmodi, non ab aliis denominantur. Unde in talibus idem est quod nomen significant, et id a quo imponitur nomen ad significandum.’ This is of course a problematic choice of examples, given the familiar modern debate over the very nature of such so-called secondary qualities.

Abstract Truth in Thomas Aquinas

59

considerable puzzle to understand just what sort of intrinsic nature a thing must have in order to be alive. But we can use the term ‘life’ all the same, by keying the reference to signs of life that track with reasonable precision – that is, with as much precision as there is in our actual usage of the term – the presence of the underlying nature.40 Presumably, with behaviorism long discredited, no one will be tempted to say that the term ‘life’ signifies the superficial behavior. That is not what it is to be alive. Terms like ‘life’ and ‘stone’ signify the underlying nature that explains the superficial behavior and properties of the thing in question. Even without knowing what that nature is, we can use these terms to refer to things of the right kind, as long as we have managed to single out properties that reliably indicate the presence of the nature in question. It is potentially misleading for Aquinas to say, as he often does, that we use ‘accidents’ to fix the reference of terms, because this suggests we are using properties that are only very loosely connected to the thing’s nature. If the term is to capture the extension of the natural kind, then the property used must be a special sort of accident, a proprium, something ‘not part of the essence of the thing, but caused by the essential principles of the species’ (ST, 1a 77.1 ad. 5). As before, the crucial point for Aquinas is the causal relationship between the nature of the thing and superficial signs that we use to fix the reference of our language. We can talk about the natures of things, even without knowing exactly what they are, in virtue of our being able to describe them through the sensible effects that they produce in the world around us. This whole scheme works only given a robust kind of realism regarding natural kinds. Aquinas does not think that he needs to introduce abstract, universal properties into the world. They can stay where they belong, in the mind. But for abstract thought to be true, the world must be structured in a certain way. There must be real similarity between members of a kind, and that similarity must be isomorphic with our conceptual scheme. Since the ultimate grounds of similarity within a species are often hidden from us, we must take our cues from what we can observe, and what we can observe must be closely correlated with the underlying essences we wish to understand. On this depends the very meaningfulness of thought and language. Bibliography Aquinas, Thomas, S. Thomae Aquinatis Doctoris Angelici Opera Omnia, Rome: Commissio Leonina, 1882– [CT, QDV, QDSC, QDA, InDA, De ente, InDH, InDT, InPA, InPH, De 108 articulis].

40 Another example of this same procedure would be the one discussed in §3: the way we refer to the genera and species of living things on the basis of their capacities. We do not know what all animals have in common at the level of essence, but we can refer to that common genus in virtue of the sensory capacities that are a sign of animality. See ST, 1a 77.1 ad. 7, quoted in note 28.

60

Representation and Objects of Thought in Medieval Philosophy

——, Scriptum super libros Sententiarum, ed. P. Mandonnet and M.F. Moos, Paris: P. Léthielleux, 1929–47 [Sent.]. ——, Summa theologiae, ed. P. Caramello, Rome: Marietti, 1950–53 [ST]. ——, Super Evangelium S. Matthaei lectura, ed. R. Cai, Rome: Marietti, 1951 [InMat.]. ——, Super Evangelium S. Ioannis lectura, ed. R. Cai, Rome: Marietti, 1952 [InJoh.]. ——, In octo libros Physicorum Aristotelis expositio, ed. M. Maggiòlo, Rome: Marietti, 1954 [InPhys.]. ——, Liber de veritate Catholicae fidei contra errores infidelium, seu Summa contra gentiles, ed. C. Pera, P. Marc, and P. Carmello, Rome: Marietti, 1961–67 [SCG]. ——, In duodecim libros Metaphysicorum Aristotelis expositio, ed. M.R. Cathala and R.M. Spiazzi, Rome: Marietti, 1971 [InMet.]. —— (1996), Thoughts, Words and Things: An Introduction to Late Mediaeval Logic and Semantic Theory, version 1.0 (http://www.pvspade.com/Logic/docs/ thoughts.pdf). Ashworth, E.J. (1991), ‘Signification and Modes of Signifying in Thirteenth-Century Logic: A Preface to Aquinas on Analogy’, Medieval Philosophy and Theology, 1, 39–67. Berkeley, George, The Principles of Human Understanding, in D.M. Armstrong (ed.), Berkeley’s Philosophical Writings, New York: Macmillan, 1965. Burgess, John P. and Rosen, Gideon (1997), A Subject with No Object: Strategies for Nominalistic Interpretation of Mathematics, Oxford: Clarendon Press. Duns Scotus, John, Ordinatio in C. Balic et al. (eds), Opera omnia, Vatican: Scotistic Commission, 1950–. ——, Cuestiones Cuodlibetales, ed. F. Alluntis, Madrid: Biblioteca de Auctores Cristianos, 1968 [QQ]. Geach, Peter (1957), Mental Acts: Their Content and their Objects, London: Routledge. Henry of Ghent, Summa quaestionum ordinariarum, Paris, 1520, (repr. 1953, St Bonaventure). Honnefelder, Ludger (1979), Ens inquantum ens. Der Begriff des Seienden als solchen als Gegenstand der Metaphysik nach der Lehre des Johannes Duns Scotus, (Beiträge zur Geschichte der Philosophie und Theologie des Mittelalters, NF 16), Münster: Aschendorff. King, Peter (1994), ‘Scholasticism and the Philosophy of Mind: The Failure of Aristotelian Psychology’, in T. Horowitz and A. Janis (eds), Scientific Failure, Lanham, MD: Rowman and Littlefield, 109-38. Klima, Gyula (1996), ‘The Semantic Principles Underlying St Thomas Aquinas’s Metaphysics of Being’, Medieval Philosophy and Theology, 5, 87–141. —— (2000), ‘Aquinas on One and Many’, Documenti e Studi sulla Tradizione Filosofica Medievale, 11, 195–215.

Abstract Truth in Thomas Aquinas

61

Kretzmann, Norman, Kenny, Anthony and Pinborg, Jan (eds) (1982), The Cambridge History of Later Medieval Philosophy, Cambridge, UK: Cambridge University Press. Marrone, Steven (2001), The Light of Thy Countenance: Science and Knowledge of God in the Thirteenth Century, Leiden: Brill. Pasnau, Robert (2002), Thomas Aquinas on Human Nature: A Philosophical Study of Summa theologiae, 1a, 75–89, New York: Cambridge University Press. —— (2003), ‘Cognition’, in T. Williams (ed.), Cambridge Companion to Scotus, Cambridge: Cambridge University Press, 285–311. Schoot, Henk J.M. (1993), ‘Aquinas and supposition: the possibilities and limitations of logic in divinis’, Vivarium, 31, 193–225. Sorabji, Richard (1971), ‘Aristotle on Demarcating the Five Senses’, The Philosophical Review, 80, 55–79. Spade, Paul Vincent (1994), Five Texts on the Mediaeval Problem of Universals, Indianapolis: Hackett.

This page intentionally left blank

Chapter IV

Representation in Scholastic Epistemology Martin Tweedale

Perhaps the most basic and most perplexing problem for epistemology is what is currently called the problem of intentionality, i.e. how it is possible for any being to have acts like thinking, imagining, perceiving, and states like belief and desire which are about something other than themselves and frequently about something external to the being in question altogether. Among the thirteenth and fourteenth-century scholastics the approach to this problem was almost always through the notion of representation. This chapter will first attempt to describe the ideas the scholastics inherited from the ancients and from earlier thinkers in the Islamic world and then proceed to concentrate on the notion of intentional existence and how views on it changed radically in the late thirteenth and early fourteenth century. Necessarily an appraisal of this length can give only a brief sketch of developments, but I hope to suggest what the basic motivations were for the varying approaches that were propounded. 1 Origins At the very beginning of De interpretatione1 Aristotle tells us spoken sounds are signs of certain pathemata in the soul, which are themselves homoiomata of things in the world. A homoioma is a likeness or image while a pathema is some condition the soul passively receives. Aristotle refers us to his works on the soul for further explanation, but in fact the notion of a likeness or image is not prominent in his psychological pieces. Rather Aristotle relies on the idea that what is cognized, either through sense perception or mind, is a form which exists both in external objects and in the cognitive faculty. The theory is limited to the primary objects of cognition, which, in the case of sensation, means the forms of the very properties of things which excite the sense organs to activity, and, in the case of the mind, to the abstract natures of things encountered in the world we perceive. It is clear enough that Aristotle wants to attribute two sorts of existence to these forms: in the one existence they exist as the definitive natures of material things and give rise to actual instances 1

De int. 1, 7–9

64

Representation and Objects of Thought in Medieval Philosophy

of types of things; in the other existence they do not give rise to other instances but allow for some form of cognition of the type of thing in question. Aristotle says that in the case of sense perception ‘the sense has the power of receiving into itself the sensible forms of things without the matter.’2 He compares this to the way wax receives the shape of a signet ring without taking on any of the gold or iron of which the ring is composed. Still it seems clear that the sense or sense organ does not take on the form in a way which gives us another thing that has the quality in question in the same way as the external object does. In the case of sight there is the suggestion that when a color is perceived the transparent liquid in the eye is said to take on that color, but only in a different sense from that in which the external object is colored. The external object is colored kath hauto, by virtue of itself, while the eye liquid is colored by virtue of something else, in the way the sea is blue.3 All of this suggests the idea that something in the organ of sense is a likeness of what is being perceived. In the case of non sensory cognition, i.e. intellectual cognition, it is the abstract and non sensible natures of things that are the objects, and these are the forms of physical things, but they are intelligible objects only when they exist in the immaterial mind. In fact the existence of these forms in the mind is the full actuality of the mind itself.4 There are two important differences, however, between the intellectual and the sensory forms of cognition. First of all, in the case of sense perception Aristotle never says that what is sensed is something internal to the sense itself. Although in his view the full actuality of the sense object is identical with the activity of the sense faculty, an activity which does occur in the sense organ, Aristotle comes close to some such position, by and large the assumption seems to be that we perceive something that exists in external reality. But with the intellect and Aristotle holds the view that the object is a universal and universals exist only in minds, and this is why the mind can call up its object at will once it has apprehended it.5 In this the mind is like the imagination, only the imagination works with phantasmata, i.e. images retained from sensation. It is also claimed that the mind when it thinks has to make use of these images.6 It would be highly unlike Aristotle to claim that it is the images themselves that we are thinking about or even imagining; rather we are thinking about and imagining what the images are images of. Hence there is implicit here a belief that representation plays a role in thought and imagination, if not in sense perception itself. Centuries later the peripatetic philosopher and commentator, Alexander of Aphrodisias (fl. c. 200 CE), treated sense perception as an assimilation of the sense faculty to the sense object via some alteration of the sense organ. He made it clear, 2 De anima, II 12, 424a17–20. 3 See Tweedale (1992), 227. The key texts are De anima, II 7, 418a28–b5, and De Sensu, 439b1–5. 4 See De anima, III 4, 429a28. 5 See De anima, II 5, 417b19–25. 6 See De anima, III 7, 431b1.

Representation in Scholastic Epistemology

65

however, that the sensible form that the organ receives does not make the organ genuinely have the quality being perceived. In the case of sight, where it is colors that are perceived, the colors exist in the organ only in the way they do in a mirror or in water.7 This is not so far from what I suggested was Aristotle’s view, and is a clear attempt to find a way for a form to exist other than the way in which it makes something genuinely have a certain quality. Still later we find Themistius (c. 320–90), whose commentary on De anima was known to the Christian scholastics in the latter part of thirteenth century, proposing that the sensible form exists in the sense faculty but not in its matter. At one point he says: But the senses do not come to be the materials of the sense objects, for the sense does not whiten or blacken or get heavy or sharp.8

Themistius’s theory marks a point where the existence of the form whereby it either is cognition or a pre-requisite for cognition is described as an immaterial existence. Its subject is the soul itself rather than any organ of the soul’s faculties. This line of thought is perpetuated by John Philoponus (late sixth century) in the following passage: Since the body is affected by heat, the tactile sense is also affected, but it is not the same affection. Rather the sense has been affected cognitively by just the form of the hot thing, while the sense organ or flesh is like matter which in virtue of both form and matter becomes the subject for the heat itself and is affected by the whole thing that heats it as a whole. It is no wonder if sense is affected by the sensible objects in a different way than is the sense organ and bodies generally, for the being of the colors, flavors, sounds, heats, colds themselves is different from the being of the sensible object. For this reason colors, flavors, sounds and the rest exist even when sense does not, but sensible objects do not exist if sense does not lay hold of them.9

The sensible objects then have a different sort of existence from the sensible qualities, an existence dependent on being apprehended by the sense faculties. The implication is that the matter of the sense organs is not what supports this existence, although in other places Philoponus grants that in sensation alterations are produced in the organ. When we turn to the Islamic thinkers we find that Avicenna had quite a complex view of sense perception, which involves elements of the preceding views but worked up in quite an original way. He held that sensation involves the sensed form

7 The relevant texts here are found in Alexander of Aphrodisias, Alexandri de anima liber cum mantissa, 38, 21–39, 2; 39, 12–14; 62, 11–16. See Tweedale (1992), 224–5. 8 See Themistius, Themistii in libros Aristotelis de anima paraphrases, 78, 8. See Tweedale (1992), 223–4. 9 See Philoponus, Ioannis Philoponi in Aristotelis de anima libros commentaria, 438,10–20. See Tweedale (1992), 222–3.

66

Representation and Objects of Thought in Medieval Philosophy

having some sort of existence in the sense organ and that in sight ‘likenesses’ are radiated into the eye. Those [apprehensive powers] that apprehend by the agency of what is outside are the five or eight senses. Of these one is sight, which is a power belonging to the optic nerve for apprehending the form of that which is formed in the crystalline humor from likenesses of bodies having color, which likenesses come through bodies radiated in actuality onto the surfaces of smooth bodies.10

There is definitely a suggestion here that the object of sight is something in the eye rather than something external. Avicenna was aware that this makes the sensing of something external indirect, as the following passage shows. For the primary sensed object is most definitely what is represented in the sense’s instrument, and that is what [the sense?] apprehends. Further, it seems that when we say that what is outside is sensed our meaning is more than what we mean by saying that what is to be sensed is in the soul. For the meaning of ‘what is outside is sensed’ is that its form is assimilated in my sense; but the meaning of ‘what is to be sensed is in the soul’ is that its very form which has been imaged is in my sense, and on account of this it is difficult to count it as one of the sensible qualities found in bodies. But we must definitely know, since some body affects the sense but that body is not affected, that in the body there is a peculiar quality which is the origin of a change in the sense, even though a sense does not change it.11

Here we have an adoption of the assimilation theory found in Alexander combined with Philoponus’s point that the form in the soul is not the same as the external sensible quality. It is a representation or likeness of it. Imagination too, according to Avicenna, requires a representative likeness, as is particularly evident in the following: A corporeal instrument is also necessary for the apprehension of singular forms by a complete abstraction from matter and by ridding the abstraction of any sort of material concomitants, as is the case in imagination. For imagination can imagine only if the imaginable form is represented in it in a body by a representation which is common to the power and to the body. For Socrates’s form, which is represented by the imagination in respect of its figure and its outline, and in respect of the position of his various limbs to each other (which appear in the imagination just as though they were seen) can only be imagined as such if the parts and dimensions of his limbs are represented in a body in such a way that the dimensions of that form are in the dimensions of the body and its parts in that body’s parts.12

10 Avicenna, Avicenna Latinus. Liber de Anima seu Sextus de Naturalibus, part I, ch.3, p.59, 23ff. 11 Ibid., part II, ch. 2, p.120, 42ff. 12 Ibid., part IV, ch. 3, p.45, 41–6, 52.

Representation in Scholastic Epistemology

67

We can see from these passages that the existence of the representative likeness in both sense and imagination is very much tied to the corporeal organ, in a way that differs from Themistius’s and Philoponus’s accounts. This accords better, I would say, with the original Aristotelian position. But then Avicenna added an entirely new dimension to his view of sense perception. He claimed that besides the sensible forms the faculty of sense perception also apprehends ‘intentions’ of sensible things. Consider the following passage: Now of the powers of apprehension that are moved by what is inside, some apprehend the intentions of sensibles. Moreover, of the apprehensive powers there are some which apprehend and operate at the same time, some which apprehend and do not operate, and some apprehend principally and some secondarily. Now the difference between apprehending a form and apprehending an intention is this: The form is what the interior sense and the exterior sense apprehend together, but the exterior sense first apprehends it and then sends it on to the interior sense. For example, when a sheep apprehends the form of a wolf, i.e. its shape, character, and color, the sheep’s exterior sense first apprehends it, and then the interior sense. An intention, on the other hand, is what the soul apprehends of the sensible even though the exterior sense did not earlier apprehend it. For example, the sheep apprehends an intention which it has of the wolf, which is the reason why it is compelled to fear the wolf and flee, even though the sense does not apprehend this in any way. But what the exterior sense first and then the interior sense apprehend of the wolf is here called a form, while what the hidden powers apprehend without the senses is here called an intention.13

The Arabic word for ‘intention’ also translates the Greek ‘logos’, and thus must be something like what a word means or an account of what something is. Nevertheless, Avicenna is quite willing to attribute such knowledge to brute animals; without it animals would not recognize the need to flee what is dangerous or, presumably, to pursue what is good to eat. On top of all this, but, I think, on a continuum with it, is Avicenna’s view that intellectual knowledge is also an intention, but here a universal intention which is just some essence existing in the mind. Avicenna also treats this as the intelligible form common to many things of the same type.14 Since Avicenna held to the immateriality of the mind, we can be sure that the intelligible form or intention does not require any material basis. Neither is there any indication that the intentions formed by the interior senses are based in bodies. In Avicenna the intentions, but not the likenesses caused in the exterior senses, appear to be based in the soul alone, not any corporeal organ. Averroes, the great twelfth century commentator on Aristotle, erased the distinction Avicenna had drawn between the likenesses of sensible qualities that exist in the sense organs and the intentions that are only perceived by the interior senses. His view is that only intentions are perceived.

13 Ibid., part I, ch. 5, p.85, 88–6,6. 14 Avicenna, Avicenna Latinus: Liber de Philosophia Prima sive Scientia Divina.

68

Representation and Objects of Thought in Medieval Philosophy … we are to think that the reception of sensible forms by each sense is a reception abstracted from matter. For if it received them with matter they would have the same being in the soul as outside the soul. Thus in the soul there are intentions and apprehensions, while outside the soul there are neither intentions nor apprehensions, but rather material things that are not apprehended at all.15

Averroes has interpreted Aristotle’s dictum that in the senses the form is received without the matter as meaning that in the senses they have some immaterial or ‘intentional’ existence, rather than as simply meaning that the material of external objects is not imported into the sense organ. A little further on Averroes elaborates on this proposal: Next he [Aristotle] says: ‘In like fashion each of the senses is affected [by what has color or flavor or sound].’ I.e., in this way each of the senses is affected by the items it is naturally suited to be affected by, whether color or sound; but it is not affected by them in virtue of the fact that it is a color or sound, since, if that were the case, it would turn out that when it had received it it would be a color or sound, not an intention. And this is what he means to say when he says: ‘but this not [in virtue of its being called each of these]’, i.e., not in virtue of being called each but in virtue of being an intention, for the intention of a color is different from the color. And then he says, ‘in virtue of being in this condition and in intention’, by way of guarding against the intentions which the intellect receives, for the latter are universals while the former are only these.16

In other words, the senses do not receive the colors or sounds or tastes etc. themselves, but only their ‘intentions’, and it is these intentions which are directly apprehended. In this respect the senses are, evidently, like the intellect, which also receives intentions, but ones that are universal rather than particular. Obviously, an account of this sort is problematic for the sort of direct realism Aristotle himself seems, at least for the most part, to have adopted. Averroes recognizes that he is opening up a gap between what exists in the external world and what the soul apprehends. Witness this passage: Someone could say that sensibles do not move the senses in the way that they exist outside the soul, for they move the senses in as much as they are intentions, and intentions in actuality do not exist in matter but only intentions in potentiality. But someone cannot say that that diversity is due to a diversity of the subjects, so that intentions come into being on account of the spiritual matter which is the sense rather than on account of a mover outside. For it is better to hold that diversity of forms is the cause of diversity of matter, not that diversity of matter is the cause of diversity of forms. Since this is so, it is necessary to posit an outside mover in the case of the senses that is different from the sensibles, just as was necessary in the case of the intellect. It seems, then, that, if we grant that diversity of forms is the cause of diversity of matter, it will be necessary for there to be an outside mover. But Aristotle says nothing about this in the case of the senses, since 15 Averroes, Averrois Cordubensis Commentarium in Aristotelis de anima Libros, 317, 13ff. 16 Ibid.

Representation in Scholastic Epistemology

69

there it is not obvious, whereas it is in the case of the intellect. But you ought to think about this, since it requires scrutiny.17

Averroes rejects here the idea that the existence of intentions in the senses is due to some special, ‘spiritual’ matter that the external sensibles affect. Instead, he sees a need for some other ‘outside’ agency to bring about the sensible intentions, just as Aristotle had posited the need for an agent intellect to bring about universal intentions in the mind. The problem of how apprehending such intentions is going to give us knowledge of the external world and its sensible qualities is clearly now a very pressing one. To sum up, in the tradition the late scholastics inherited sense perception is treated in two quite different ways. On the one hand, we have the idea that the sense organs are affected by the sensible qualities in such a way that a “likeness” or representation of that quality is formed in the organ. This, in accord with Aristotle, is thought of as the form of the sensible quality having an existence in that organ whereby it brings about cognition but does not make something in the organ genuinely have that sensible quality. The other approach is to say that this form exists immaterially in the soul itself. In Averroes this is read as thinking that an ‘intention’ exists in the sensory soul, and in this way the sensory soul is held to be much more analogous to the intellect where cognition is also accomplished through intentions existing immaterially. In either case a problem arises as to how, if what is existing in the sense organ or in the sensory soul itself is what we perceive, perception can give us knowledge of a world outside the senses where what is inside can have no existence. 2 The Theory of Species In the thirteenth century scholastics the term used for a representation, likeness or image was ‘species’, a term which had roughly this meaning in late ancient times.18 To give some idea of the way this notion is used at that time I shall talk about the views of Albert the Great and his student, Thomas Aquinas. Species figure prominently in Albert’s two theories of sensation, an earlier one found in the section of his Summa de creaturis called De homine and a later one in his questions on Aristotle’s De anima.19 In De homine he treated the form of the sensible quality existing in the sense organ as a species that has ‘spiritual’ rather than ‘natural being’. This terminology is meant to distinguish the way the form exists as a species from the way it exists in the external world in material objects. In other words, it deals with the distinction we saw made as early as Aristotle and Alexander to protect the theory from having to claim that something in the sense organ has the 17 Ibid., 221. 18 Augustine uses the term this way in several places, e.g. Contra Secundinum Manich. 2 (PL 42: 579), Contra adversarium legis 1.10.13 (PL 42: 610), De Trin. 11.2 (PL 42: 987). My thanks to Rega Wood for drawing these passages to my attention. 19 This summary of Albert’s views relies heavily on Dewan (1980).

70

Representation and Objects of Thought in Medieval Philosophy

very sensible quality it perceives and in the same way as external material objects have it. In the case of sight, but not the other senses, Albert also thought the species existed in the medium, i.e. the air or transparent fluid lying between the external object and the eye. The object, evidently, transmits a flow of species, and some of these lodge in some sort of spiritual matter in the eye. It is clear from this that the ‘spiritual being’ that the species has is a physical sort of existence but in a material that is able to support the species while not actually taking on the quality it is a species of. This means that species of opposed forms, i.e. forms which could not coexist in the same subject in natural being, can co-exist in the same spiritual matter. Albert’s later theory in his questions on De anima is more complex and in some ways reminiscent of Avicenna’s. Actual sense perception now requires an act of judging on the part of the interior, ‘common’ sense, situated in the brain. For this to occur the species that we find in the sense organs of the exterior (‘proper’) senses must be transmitted to the brain and made to exist in a still more spiritual way. This ‘clear spirit’ in the brain has an active role in this process, doing for the sensible species in the external organs what light does for colors.20 Here, I believe, Albert has taken seriously Averroes’s proposal that some special agent is needed as much in the case of sensation as in the case of the intellect. Albert represents a widespread effort among thirteenth century scholastics to have a theory of cognition based on species, and in the case of sense cognition to make this a theory where everything is physical, although it acknowledges that some forms of physical existence are not the usual concretions of form and matter on which the Aristotelian ontology is based. This sort of existence, often referred to as intentional as well as spiritual, allows for a sort of halfway house between the standard matter/form composites and the totally non-physical way in which species exist in the intellect. On this view all cognition involves a to some degree de-materialized existence for the form cognized, although in sensation and sense perception some special matter is still required. In his later theory Albert allowed that these species exist in the media for all the external senses, thus making even clearer that we are dealing here with a physical sort of existence. The species, i.e. the representation or likeness of the sensible form, can exist even in inanimate bodies, although only in animate ones does it go on to produce sensation, for only in them do we find the especially clear spirit of the internal sense organs. Aquinas carried this line of thought, I believe, to its logical conclusion. Cognition consists in forms existing immaterially, i.e. intentionally, and these he called species.21 Nevertheless, Aquinas followed Albert in thinking that sensible species exist in the inanimate media as well as in the sense organs, and, since Aquinas certainly did not want to attribute cognition to things without soul, a contradiction arises. It is not clear what Aquinas would have to say about this if it were brought to his 20 Albertus Magnus, In Aristotelis librum de anima commentarium, II, IV, 12, 165.20–30. 21 This interpretation of Aquinas’s remarks has been adopted both by me in Tweedale (1992), 216–18, and Robert Pasnau in his excellent study, see Pasnau (1997), ch.1.

Representation in Scholastic Epistemology

71

attention, but I am inclined to believe that, like Albert, he would have acknowledged degrees of ‘spirituality’ in physical things and held that the inanimate media are insufficiently spiritual to produce cognition although they allow for intentional existence of species.22 The other notable feature of Aquinas’s account is his refusal to let his belief in the existence of species, both sensible and intellectual, get in the way of epistemological direct realism. We saw that in Philoponous, Avicenna and Averroes there is a willingness to allow that what is directly apprehended by the senses is something internal to the senses, not the sensible quality in the external world. Aquinas, I think, sees that this is leading toward very un-Aristotelian doubts about our knowledge of the external world and almost always insists that the species is the means of cognizing what it represents; it itself, at least not in the primary act of cognition, is not what is cognized. This was not a universally held view; Durand of St Pourcain, writing in the early fourteenth century argues against species on the grounds that, if we cognized sensible qualities through such a representative, we would cognize the species first. But he takes it as obvious that it is the quality we first apprehend.23 Robert Pasnau has explored in depth what Aquinas’s view might have amounted to,24 and he holds that it allows for an apprehension of the species which permits cognition of the external cause of the species in the sense that it is that external object which we are prepared to make judgments about. This seems to me to be very likely the correct interpretation. To put it differently, I think Aquinas was thinking of the way in which we can look at a picture and focus not on the picture but on the things pictured, being prepared to make judgments not about the former but about the latter, although the latter is certainly something we apprehend. Just as it takes a special reorienting of our attention to take the picture as something we want to make judgments about, so it takes a special reorienting to assess a sensible or intelligible species. But if Aquinas thinks that this view escapes the sceptic’s reach, I think he is mistaken. Once we view perception this way, it is perfectly reasonable to ask whether, even in normal cases, the species represents the external object as it really is, in just the way we can question this when we are relying on a picture of something. Although in the primary act of cognition the species is used as a means to cognizing something else, Aquinas does allow that the intellect is capable of acts of reflection by which it cognizes the very species by which it cognizes its objects.25 Also Aquinas allows for the intellect to take a sensible image as its object rather than the thing the image is an image of.26 It does not appear that in such passages Aquinas 22 Pasnau (1997) holds the view that inanimate media stand at the bottom of a continuum of things that are capable of retaining information representationally. This, I think, is compatible with my proposal, but he is more inclined than I to exculpate Aquinas from a major oversight. 23 Ibid., 17–18. 24 Ibid., ch. 6. 25 See Summa Theologica, I, qu. 85, art. 2. 26 See Pasnau (1997), 206.

72

Representation and Objects of Thought in Medieval Philosophy

is just referring to the way the intellect might recognize species as some entity needed to construct a coherent theory of cognition. Rather, it is more likely that he thinks we can have some knowledge of the intrinsic character of these species. But what do we know about them other than that they represent certain things to the sense or mind? Like Albert, Aquinas denies that the species is a ‘natural’ likeness of what it represents. It does not literally have the same qualities. But it is a representational or intentional likeness. Although there were scholastics who insisted that the similarity between the species and the represented object had to be ‘natural’,27 I think Aquinas’s view is defensible. Pictures can accurately represent things without being similar to them to any significant degree. A painting or photograph of a house is not anything like a house, but it can still show you what the house is like. Admittedly, just how this is possible in the case of species, particularly intelligible species, is hard to say. But perhaps Aquinas does not have to say. He could, if pressed, just claim that the soul is innately equipped to read or interpret the species analogously to the way we seem able to interpret pictures without any special training. However, this relatively modern idea of ‘interpreting’ likenesses is not one Aquinas and the late scholastics in general seem to have had in their philosophical toolkit. Consequently, it was a problem for them to see how the species could represent anything, given it had no natural likeness to what it was supposed to represent. I think this is part of what leads to dissatisfaction with the species theory generally, and the adoption by such figures as Peter John Olivi and William of Ockham of theories which rely simply on the act of cognition itself without any prior existence of a species.28 Ockham, however, still referred to the act itself as a likeness of what it cognizes and Olivi thought of the act as itself a species, so representation was still involved. But here I think it is likely that the act is representative of what it cognizes because it is cognizing that thing, rather than being cognitive of that thing because it represents it. Or at least there is no more reason to say one is because of the other than the other way around. A robust species theory such as Aquinas’s makes the species in both the senses and the intellect present an object to the cognitive power in question, and this must occur before an act of cognition can take place. What is the attraction of such a theory? It seems to me that probably thinkers like Aquinas thought there was something very unreasonable in supposing that an act of the soul could be directed to an object without there being in the soul or its organs something apart from the act that determined what object the act was directed to. Olivi and Ockham challenge this assumption by holding that the act, having been partially caused by the external object, carries in itself all that is needed to be directed toward an object. It is not necessary, then, to see the prior cause of the directedness of the act in something internal to the cognizer.

27 Notably Roger Bacon and William Crathorn. See Pasnau (1997), 66 and 90. 28 See ibid., 19–17 and 41–2 of Tweedale (1990).

Representation in Scholastic Epistemology

73

3 Esse Objectivum Toward the end of the thirteenth century a distinction between esse subjectivum and esse objectivum comes into common use among the scholastics. Perhaps the first thinker to make heavy use of it is John Duns Scotus. The idea here is that something might have two ways of existing: (1) a real existence in no way dependent on being the object of any mental act or state; (2) existence as an object of some mental act or state. The former is esse subjectivum; the latter, esse objectivum. Something can have either of these without the other, or both at once. If someone imagines something that in fact does not exist, what the person imagines has only esse objectivum. It can still be described in the vocabulary we use to describe real things, but the truth of such descriptions does not imply it has any real existence. We can also think of something that does really exist, and then that thing has both forms of existence. Esse objectivum comes in various forms depending on what mental act or state is taking the thing as its object. Scotus and others speak of esse cognitum, esse volitum, esse intellectum, and esse representatum, as all forms of esse objectivum. Even where the ens objectivum is something that has esse subjectivum as well the ens objectivum may have properties that the ens subjectivum does not. Thus Scotus makes universality belong to the common nature as an ens objectivum, but not as an ens subjectivum, because an ens objectivum can have a certain indeterminateness that an ens subjectivum cannot.29 To understand this think of a picture of some person who really exists. The picture may not show us whether the person has long hair or not, and thus the ens objectivum here is not determinate in respect of that feature, although certainly the real person is. We can talk about what the picture represents without supposing that what it represents has any real existence, but then we have to admit that there is indeterminacy in what it represents. Scotus treats concepts and intentions as entia objectiva. He speaks about things of first or second intentions, by which he means some property, and concepts of first or second intentions, by which he means a class of entia objectiva. This is quite a radical departure from the treatments of intentions and esse intentionale that we have examined previously. There intentional or spiritual being is a kind of real being, but one that only species can have. In the case of sensible species it is even a sort of physical real being. In contrast, Scotus treats esse intentionale as a sort of existence which does not of itself entail any real existence. Nevertheless, he retains species but these are forms having a real existence in the organ or soul. This approach clarifies matters considerably. The species is analogous to a picture and the ens objectivum is analogous to what the picture represents taken without any assumption of whether this has real existence or not. Obviously, when we dream or hallucinate or imagine, there is something that we apprehend even if what we apprehend has no real existence. 29 Unfortunately Scotus does not give treat esse objectivum extensively in any one place. One has to draw conclusions from many different passages. One which implies that universals have this sort of existence occurs in Ordinatio, I, d. 8, pt. 1, qu. 3 in Opera Omnia, vol. 4, sects. 146–7. I have explained how I read Scotus on this matters in Tweedale (1999), vol. II, 608–10.

74

Representation and Objects of Thought in Medieval Philosophy

The entia objectiva provide objects here for those acts. But they are there in veridical thought and perception as well, but not as something other than the really existing thing being thought of, for one and the same thing can have both sorts of being. Given cognition is through a species and of what that species represents, we have to admit that once there is a species a cognitive act of what it represents is possible and this cognitive act can occur even though what is represented does not exist. Later in the fourteenth century Peter Aureol used this device to explain sensory illusions.30 Scotus uses it to explain God’s knowledge of possible things that will never be created, although in this case no species in the usual sense is required.31 God’s essence itself is what does the representing, and since the entia objectiva that result have no real existence we are not involved in claiming something real other than God exists without being created. Unreal objects of cognitive acts had been proposed in the twelfth century by Peter Abelard,32 but he did not see them involved in veridical perception. The reason Abelard treats these ‘images’ as unreal is that they are described as having the features which real physical things have, but it would be absurd to think that just by imagining something that is, say, round and tall, some real, round and tall object comes into existence. He makes it clear, too, that he is not attributing some real but mental existence to images. They have neither real physical nor real mental existence. Since Abelard did not think such images were involved in veridical perception, he did not open himself up to the skeptically oriented criticism that he had interposed between the perceiver and what he perceives some third entity that mediates the perception. Scotus’s view, and even more Aureol’s, does, however, seem vulnerable to that kind of doubt, and this is largely why Ockham, after first having some sympathy with entia objectiva, abandons the whole idea. To defend against the threat of skepticism one has, at least, to emphasize that in veridical perception the really existing object is indeed the thing which also has esse objectivum. In perceiving the ens objectivum one is not thereby perceiving something other than the external object. One is just perceiving that object as it is represented by the species that presents it to the cognitive faculty. To then go on and ask how we can be sure that it is presented as it really is, is a form of doubt that Scotus replies to by saying that it is self-evident that a faculty does not err with respect to its appropriate objects unless it is disordered, and we can know when a faculty is disordered.33 That reply 30 See Pasnau (1997), 71–6. It is not clear from what I have read how close Aureol is to Scotus in his treatment of esse objectivum. 31 This doctrine is prominent in Lectura, I, d. 36, qq. 1–2 (in Opera Omnia, vol. XVII, 461–76; see paras 26, 30) and in Ordinatio, I, d.35, q.1 (in Opera Omnia, vol. VI, 245–70; see paras 32, 41–2.) It can also be found in Ordinatio, I, d. 3, pt. 1, qu. 4 (in Opera Omnia, vol. III, 123–72; see para. 268. The topic is well discussed in Perler (1994a), (1994b) and (1995). 32 See his commentary on the Peri ermenias, in Logica Ingredientibus, pp. 314(25)– 15(17). 33 Ordinatio, I, d. 3, pt. 1, qu. 4 (in Opera Omnia, vol. III, 123–72).

Representation in Scholastic Epistemology

75

is perhaps not totally reassuring, but this sort of skepticism arises whether we have a representational theory of cognition or not. It is clear, however, that the ens objectivum which is the object of a veridical cognition had better not have some quasi-real existence that makes it distinct from the species doing the representing, for then indeed we have some intermediate thing which is the genuine object rather than the external, fully real thing which we suppose ourselves to know. In the remainder of this paper I want to describe how William of Alnwick, a student and follower of Scotus’s confronted and developed an alternative to such a view in the first question of his Quaestiones Disputatae de esse intelligibili.34 Alnwick describes the view he wants to refute as follows: Some recent thinkers say to this question that the esse representatum of some object signifies an entity distinct from what represents the object, and that the esse cognitum of some object designates an entity distinct from the cognition.35

The former esse is what something gets just be being represented, something required of all cognition on the view we have been considering. The latter esse is attributed only when something is the object of an act of cognition. Alnwick goes on to ask the question we all want to see answered: But what is this being and what is this distinction between esse representatum and esse cognitum on the one hand and the item doing the representing and the cognition on the other? To this those holding the above view say that there are three sorts of being taken generally: namely, real being [esse reale], intentional being [esse intentionale] and being of thought [esse rationis].36

Real being is the actual existence of the thing while intentional being is equated with esse representatum. Being of thought or esse rationis belongs only to types of things which come into existence on account of acts of the intellect in which it compares and relates the contents of the mind. Most of the scholastics of this era think universals have only esse rationis, since they exist only by the intellect predicating one content of another. It is important to realize that all the thinkers who use the notion of esse representatum or esse intentionale distinguish it from esse rationis and in fact claim that things must have the former sort of being before there is any question of arriving at items with the latter sort. The mind has to have objects before it can compare and relate them. The view Alnwick is describing equates esse representatum with esse intentionale. Alnwick goes on to explain further the reasoning his opponents use: Therefore, they say that intentional being is not real being, because it can belong to a thing that does not exist in its own distinctive nature, nor is it being of thought, since a being

34 William of Alnwick, Quaestiones Disputatae de esse intelligibili, 1–29. 35 Ibid., 3–4. 36 Ibid., 6.

76

Representation and Objects of Thought in Medieval Philosophy of thought [enti rationis] is opposed to existing in reality. Thus this intentional being is intermediate, as they say, between real being and being of thought.37

Alnwick states his own view as follows: But to me it does not seem that these views are correct. Thus, I show that esse representatum is really the same as the form that does the representing and esse cognitum is really the same as the cognition, for every positive entity that does not depend on the soul is a real entity, because such an entity would have existence even if the soul did not. For this reason The Philosopher and The Commentator, in Metaphysics VI, divide being by a first division into being in the soul and being outside the soul, and they say that a being outside the soul is a real being because it is that which they divide into the ten categories, each of which is a real being or a real entity. For it is obvious that what is no thing is nothing; consequently, it is obvious that being which is not dependent on the soul is real, but represented being is a positive being and a certain positive entity, as even they [Alnwick’s opponents] allow, and it is not dependent on the operation of the intellect or soul, as they also allow, since the species would represent something even if the intellect were not thinking. Therefore, if esse representatum is real being or a real entity, it is not something other than the entity of what does the representing, because, if it were some other real entity outside the soul, it would have subjective being [esse subjectivum] really distinct from what does the representing. Thus esse representatum is really the same as the form that does the representing.38

What Alnwick is arguing for here is the genuine reality of representation; he is not arguing for the reality of what is represented, although of course often enough what is represented is real. Being represented is something that really happens to things; it is just as real as what does the representing, for, in fact, there is no real distinction between them. If there were such a real distinction, then either there could be something represented, an ens representatum, without representer, and then we do genuinely have something with its own subjective being, a view even his opponents do not want to hold, or we could have a representer without any ens representatum, in other words representation but nothing represented, an obvious absurdity. Alnwick will admit, however, that esse representatum is a ‘diminished’ sort of being for the thing represented, for it does not imply the real existence of that thing (nor, of course, does it imply that the thing does not have real existence.) What his opponents have done is confuse this point with the sort of reality we want to ascribe to the esse representatum itself. Alnwick asserts this in the following: … when it is argued that the esse representatum of a stone is not an entity of thought because it precedes the act of the intellect, I grant that. And when it is argued further that it is not a real entity because it is a diminished being, I answer that, although esse representatum is a diminished being of the stone that is represented, it is nevertheless a real being which is really the same with the being of the form which does the representing,

37 Ibid., 6–7. 38 Ibid., 8–9.

Representation in Scholastic Epistemology

77

just as, although being in thought is a qualified being of the item that is thought of, it is nevertheless a true thing and an act of thought in the mind of the one who is thinking.39

Although the esse representatum, which, as we saw, is a real being, is attributed to the object, this does not mean that the object thereby has to be real, for it is attributed by ‘extrinsic characterization’, not as an intrinsic form. … for when I say that a stone is represented and cognized by a species or by the divine essence, the characterization is made in virtue of either an intrinsic form or an extrinsic form. It is not made in virtue of an intrinsic form inhering in the stone, because then it would not be the case that it would belong to the stone even if the stone were non-existent. Also it would follow that the known being of the stone would have a being formally inhering in the stone, and thus our intellect in thinking of the stone would cause some form to inhere in the stone, which is false. Therefore, when I say a stone is represented or cognized, the characterization is made only in virtue of a form that characterizes extrinsically, i.e. a form which is only representative or a form of thought. Thus, just as when it is said that a stone is located an extrinsic characterization is made in virtue of the place that surrounds it, even though when it is said that the stone is ‘where-ified’ [ubicatus] a characterization is made in virtue of the ‘where’ existing in the stone, so, when it is said that the stone is represented or cognized by a species, the characterization is made only in virtue of the species that does the representing and the cognition that terminates in the stone.40

In this way Alnwick can claim that although esse representatum is a real form of being, the thing to which it is attributed need not really exist in order to have it. As for the item represented (ens representatum) if it is a really existing thing, then it is really distinct from and really related to the form doing the representing, but when it is not a really existing thing it is neither, as the following passages show: … when the very item that is represented is a real thing then there is a real relation of the item that is represented to what does the representing … and thus what does the representing and the item represented are really distinguished and also the being of what does the representing and the formal, intrinsic being of the very item represented are distinguished really.41 But when the creature represented does not have actual existence, represented being belongs to it only by extrinsic denomination and not in the sense which implies a real relation, because a real relation requires an actual subject.42

This is not the place to explain the Scotist doctrines on real relations and real distinctions. Suffice it to say that both require more than one really existing thing. Also real identity, as Alnwick holds exists between both the item represented when it does not really exist and its esse representatum on the one hand, and the form 39 40 41 42

Ibid., 20–21. Ibid., 15. Ibid., 22. Ibid., 24.

78

Representation and Objects of Thought in Medieval Philosophy

doing the representing on the other, does not, for Scotists at least, preclude other sorts of distinction, less than the real distinction, which will permit certain things to be correctly asserted of one of the items but not of the other, even though these items are really identical. All that the real identity of x and y amounts to is the logical impossibility of either existing without the other. Alnwick has worked out here the logic and ontology of representation in a very subtle way that may avoid the pitfall of implying that even in veridical cognition we must cognize something internal before we cognize the external thing we suppose ourselves to be apprehending. The key move, and the greatest contribution of the Scotist theory of esse objectivum, is to clearly distinguish the representer or likeness from its content. It is not the form doing the representing that we apprehend but the content of the representation, i.e. the ens objectivum. That content, in contrast to the representer, can have a real existence external to the cognizer. I suggest, although further investigation is needed to either confirm or refute this hypothesis with any certainty, that the prior theories of representation suffer from the problems that they have largely because of a failure to make this distinction between representer and content in any clear way. Bibliography Abelard, Peter, Logica Ingredientibus, ed. Bernard Geyer as part of Peter Abaelards Philosophische Scriften (in Bd.xxi, heft 1–4 of Beitraege zur Geschichte der Philosophie des Mittelalters), Monasterii Westfalorum Aschendorff, 1919. Alexander of Aphrodisias, Alexandri de anima liber cum mantissa, ed. Ivo Bruns, Berlin: Reimer, 1887. Aquinas, Thomas, Sancti Thomae de Aquino opera omnia, iussu Leonis XIII P.M. edita. Cura et studio Fratrum Praedicatorum 1876–. Aristotle, The Complete Works of Aristotle, ed. J. Barnes, Princeton: Princeton University Press ,1984. Averroes, Averrois Cordubensis Commentarium in Aristotelis de anima Libros, ed. F.S. Crawford, vol VI,1 of Corpus Commentariorum Averrois in Aristotelem, Cambridge MA: The Medieval Academy of America, 1953. Avicenna, Avicenna Latinus. Liberde Anima seu Sextus de Naturalibus, ed. S. van Riet, IV-V, Leiden: Brill, 1968. ——, Avicenna Latinus. Liberde Anima seu Sextus de Naturalibus, ed. S. van Riet, I, II and III, Leiden: Brill, 1972. ——, Avicenna Latinus: Liber de Philosophia Prima sive Scientia Divina, 2 vols, ed. S. van Riet, Louvain: E. Peeters & Leiden: E.J. Brill, 1977 and 1980. Dewan, L. (1980), ‘St Albert, the Sensibles, and Spiritual Being’, in James A Weisheipl (ed.), Albertus Magnus and the Sciences, Commemorative Essays 1980, Toronto: Pontifical Institute of Medieval Studies, 291–320. Duns Scotus, John, Iohannis Duns Scoti Doctoris Subtilis et Mariani opera omnia. ed. P. Carolus Balic et al. Typis Polyglottis Vaticanae 1950–.

Representation in Scholastic Epistemology

79

Magnus, Albertus, In Aristotelis librum de anima commentarium, in Opera Omnia, ed. C.Stroik, tom VII, part 1, Monasterii Westfalorum: Aschendorff, 1968. Pasnau, R. (1997), Theories of Cognition in the Later Middle Ages, Cambridge: Cambridge University Press. Perler, D. (1994a) ‘What am I thinking about? John Duns Scotus and Peter Aureol on Intentional Objects’, Vivarium, 32, 72–89. —— (1994b), ‘Peter Aureol vs. Hervaeus Natalis on Intentionality’, Archives d’Histoire Doctrinale et Litteraire du Moyen Age, 61, 227–62. —— (1995), ‘Intentionale und reale Existenz: Eine spaetmittelalterliche Kontroverse’, Philosophisches Jarhbuch, 102/2, 261–78. Philoponus, Ioannis Philoponi in Aristotelis de anima libros commentaria, ed. M. Hayduck, Berlin: Reimer, 1897. Themistius, Themistii in libros Aristotelis de anima paraphrases, ed. R. Heinz, Berlin: Reimer, 1890. Tweedale, M. (1990), ‘Mental Representations in Later Medieval Scholasticism’, in J-C. Smith (ed.), Historical Foundations of Cognitive Science, Dordrecht: Kluwer, 35–51. —— (1992) ‘Origins of the Medieval Theory That Sensation Is an Immaterial Reception of a Form’, Philosophical Topics, 20.2, 215–31. —— (1999), Scotus vs. Ockham: A Medieval Dispute Over Universals, Lewiston: Edwin Mellen Press. William of Alnwick, Quaestiones Disputatae de esse intelligibili, ed. P. Athanasius Ledoux, O.F.M., (Tom X of the Bibliotheca Franciscana Scholastica Medii Aevi), Firenze-Quaracchi: Collegium S. Bonaventurae, 1937.

This page intentionally left blank

Chapter V

Rethinking Representation in the Middle Ages: A Vade-Mecum to Medieval Theories of Mental Representation* Peter King

The object sufficiently represents itself in a cognition. –William of Ockham, Reportatio, 2, qq. 12–13 (OTh, 5, 274. 4–5)

The Christian Aristotelianism of the High Middle Ages had the conceptual resources to explain the representationality of mental representation – that is, the feature or features in virtue of which a mental representation represents what it represents – in four separate ways:1 (R1) The mental representation and the represented item have the same form. (R2) The mental representation resembles, or is a likeness, of the represented item. (R3) The mental representation is caused by the represented item. (R4) The mental representation signifies the represented item. These several accounts were often uncritically taken to go together. When Socrates confronts a sheep (say), the sheep causes a particular mental event to occur in Socrates (R3), namely the sheep’s form coming to inhere in his soul (R1); this selfsame quality in Socrates’s soul, namely the inherent form, is thus a natural likeness of the sheep (R2), thereby signifying the sheep and playing the role of the mental or inner word (verbum) for it (R4). Even on this first pass there are obvious problems. Do

* Versions of this chapter were read in Oslo on 25 November 2000 and in Cincinnati on 10 May 2002. Its subtitle is an unapologetic nod to Fodor (1985). 1 It should go without saying that I’m concerned with the modern notion of mental representation as it shows up in the High Middle Ages, not with the medieval (limited and restricted) use of repraesentatio. That is a part, but only a small part, of their account of mental representation; it’s more profitable to track the concept than the terminology.

82

Representation and Objects of Thought in Medieval Philosophy

(R1)–(R4) have to go together? Which of (R1)–(R4) actually does the representing? And the like. A rough approximation of what I want to argue for here is that in the course of the High Middle Ages an important shift takes place in the theory of representation, namely a shift from accounts of representation that favor (R1) and (R2) to accounts that favor (R3) and (R4). This is all the more surprising in that (R1) and (R2) are clearly Aristotle’s preferred account of representationality, if anything is. The trajectory of the debate begins with Thomas Aquinas and is epitomized, as so many medieval philosophical discussions are, in William of Ockham. In the spirit of Fodor (1985) and Haugeland (1990), I’ll take a top-down approach to the historical sources, concentrating on the logic of the positions and their development. My account will therefore track mainstream medieval philosophy of psychology. Richard Rufus’s attack on naive representationalism, for example, won’t be considered here since it appears not to have affected the course of the debate, interesting though his arguments were. My focus is rather on scholastic ‘common wisdom’ about mental representation, to the extent there was any, in the High Middle Ages. 1 Conformality 1.1 The Simple Version According to (R1), a mental representation represents an object just in case it has the same form as the object – hence the name ‘conformality’ for this account of representation. More exactly, the inherence of the form in the appropriate kind of matter makes that matter into the very thing or the kind of thing it is, whereas the presence of the form in the soul doesn’t turn the soul into the thing itself, except metaphorically; instead it produces a sensing or a thinking of that thing, namely when the form is present in the sensitive or in the intellective soul respectively. The conformality account is usually embedded in a much larger and longer theory, for the most part meant to be a causal theory, of the reception or acquisition of such forms in the soul, involving the transmission of forms through the intervening medium (the species in medio doctrine), their affection of the sense-organ and reduction of the associated sense-faculty from potency to act, the production of a phantasm or sensible species through the common sense, and so on, with the agent intellect and the possible intellect getting their licks in too. But we can ignore the mechanical details here, since they aren’t important for explaining how a mental event is ‘about’ an external item (in the paradigm case).2 Philosophers who hold an illumination theory of cognition, for instance, can hold that God directly causes the presence of a form in the intellect, which is thereby thinking of the item whose form it is, without any direct causal link between the thought and the external item itself. (Bonaventure 2 See the recent surveys in Tachau (1988), Spruit (1994), and Pasnau (1997) for the broader theory of the reception of forms in the soul.

Rethinking Representation in the Middle Ages

83

and Matthew of Acquasparta are examples of this view, and arguably Henry of Ghent as well.3) Hence (R1) is independent of (R3). A mental representation represents something by virtue of having its very form in mind, no matter how it got there. The precise details of the conformality account depend on the answers to a series of metaphysical questions. Are there individual forms or only non-individual forms that are (non-formally) individualized in individuals? What is the principle of individuation? Does an individual have a plurality of substantial forms or only one? But even without settling these and related questions, we can explore the paradigm case in which Socrates, confronted with a sheep, thinks about it by virtue of having the sheep’s form, whether individual or specific, in mind.4 Socrates might think about, and hence represent, the sheep confronting him in a ‘thin’ fashion, namely by thinking about all sheep, or about sheephood wherever it may be found – but though thin, it is nevertheless a case of thinking about this sheep in the end. (Whether we can identify an individual sheep with its non-individual form is a separate and disputed metaphysical question.) The point is that a general representation still represents what it does by virtue of the presence of the same form. Yet what is it about the presence of form in the soul that makes it represent things that have the form? Put another way, why doesn’t the sheep represent Socrates’s mental state, by virtue of the presence of the form in the sheep, as much as vice-versa? Conformality, as a kind of identity or sameness, is symmetrical, but representation isn’t (or usually isn’t taken to be). Three possible replies suggest themselves, each problematic. The first reply: The defender of conformality could simply bite the bullet and maintain that conformal items represent one another, in spite of the counterintuitive consequences: external objects represent mental states, and each thing represents itself. This line of reply was quickly dropped for its obvious drawbacks. The second reply: There might be something interestingly special about the subject in which the form is present, namely the soul, so that the presence of a form in the (intellective) soul counts as representing its object whereas its presence elsewhere does not.5 Yet this merely names the mystery rather than explains it; what 3 For Bonaventure see especially his Quaestio disputata de cognitionis humanae suprema ratione; for Matthew of Acquasparta see his Quaestiones disputatae de cognitione q. 1 ad. 22, q. 2 ad. 1 and ad. 12, Quaestiones de anima, 13, q. 5. Both are discussed in King (1994). For Henry of Ghent see his Summae quaestionum ordinariarum, art. 1 q. 2 (amplified and modified in art. 58 q. 2), Quodlibeta, 8, q. 12 and 9, q. 15; see also the analysis in Marrone (1985). 4 This strategy assumes that having the specific form of sheephood in mind is a way, though not necessarily the only way, of representing an individual sheep. If not, then a separate account of singular thought, and in particular how singular thought can represent an individual, is required. Aquinas is the best-known case of a philosopher who rejects the assumption, and his account of singular thought, if indeed he has one, is at best obscure: the conversio ad phantasmata. 5 This is the version of the conformality account Cummins (1989) ridicules, saying that ‘mind-stuff’ is what makes representationality. But it’s only one possible way to go

84

Representation and Objects of Thought in Medieval Philosophy

is the special feature of intellect such that the presence of form in intellect, though not elsewhere, becomes a representation? But a deeper objection, I think, is that this reply winds up giving no account of representationality at all. Instead, the presence of the form determines which thing the representation represents, while the burden of the claim that there is representation going on falls squarely on the (unexplained) nature of intellect. And in the High Middle Ages that was too high a philosophical price to pay. The third reply – Aquinas’s way out and the most promising of the three – holds that it isn’t the subject but rather the mode of the form’s presence that makes all the difference. The form of the sheep in Socrates’s soul doesn’t inform the soul in such a way as to make it into a sheep, which is what it usually does (given appropriate matter), and indeed it’s not clear that the notion of a corporeal form informing an incorporeal substance in anything like the ordinary sense of ‘inform’ makes a great deal of sense. Rather, the form must be present in the soul, but present in a special way. Aquinas says that the form is present not ‘really’ but spiritually or intentionally.6 The suggestion is ingenious. On Aquinas’s reading, Aristotelian physics is deeply committed to the notion that a form may occur in something without literally informing it, e.g. a color in the medium which doesn’t tint the intervening air (De veritate, q. 27, art. 4, ad. 4); why not make use of this idea in psychology? Furthermore, unlike the second reply, it does make something relevant to the form, namely its presence, be the key to explaining representation, though like the second reply it makes the form explain only why the representation represents this sheep rather than something else, not why representation occurs at all.7 Promising as it is, there are two drawbacks to this third reply. First, since it doesn’t depend on any special features of the subject in which the form intentionally exists (otherwise it would be a version of the second reply), there is no good reason to rule out representation wherever forms may intentionally exist – say, a color existing intentionally in the air, as the blackness of the sheep’s wool is said to exist intentionally in the air, which is the intervening medium between the sheep and Socrates’s eye; the intervening air, though not itself black, would thereby be said to represent the black color, or perhaps even the sheep itself. For many this would qualify as a reductio ad absurdum, but Aquinas seems to have bitten this particular bullet: he declares that ‘air and water are perceptive of color’ (aër et aqua …sunt perceptiva coloris: In De anima, 3.1, §570). To defend this claim we might point out on Aquinas’s behalf that representation isn’t the mark of the mental; the statue of Hercules in the park represents Hercules, and a mirror-image represents that of in developing the conformality account, and not common at that. Cummins also conflates conformality (R1) with likeness (R2) in his discussion: Cummins (1989) 3–4 and Chapter 3. It may be that conformality is no longer a contender in the race to explain representationality, but it should lose in its own right, not as a caricature of itself. 6 See for example In De anima, 2.14, §418 and 2.24, §553, Summa theologiae, 1a, q. 14, art. 1 corp., De veritate, q. 2, art. 2 corp., et passim. Aquinas probably took this idea over from Albert the Great: see e.g. Albert’s De anima, 3.3.12 and Spruit (1994), vol.1, 144. 7 See De veritate, q. 10, art. 4 corp. for an exceptionally clear statement of this point.

Rethinking Representation in the Middle Ages

85

which it is the image. Hence Aquinas’s use of ‘perceptive’ is not entirely without precedent. However, the second drawback to this third reply is more serious. What is it for a form to be present only ‘intentionally’? Aquinas never says, or, to the extent that he does, his account was opaque to his disciples and detractors alike, then and now.8 Aquinas’s failure to say what intentional presence consists in makes representationality into a mystery again, this time centered on the non-informing presence of the form in the representer; it may well explain why Aquinas had few followers in philosophy of psychology during the High Middle Ages. 1.2 The Composite Version Duns Scotus starts from a new direction, using a top-down approach to psychology based on simple introspection.9 His reasoning is as follows. Thinking is episodic; sometimes we think and sometimes we do not. In general, episodic processes are metaphysically identified as accidents, since they may be present or absent. Well, accidents require substances to inhere in, of course, and in this case there is a handy (quasi-) substance available: the soul. Thus a mental event, such as Socrates’s thinking of a sheep, is metaphysically analyzed as an accident (the thinking-of-thesheep) inhering in a substance (Socrates’s soul).10 How does this substance–accident analysis square with the claim that to think of something is to have its form in mind? The natural answer is that the representational form just is the accident inhering in the mind. But the natural answer won’t do. Aside from the metaphysical difficulties it faces in holding that (say) a corporeal substantial form such as sheephood can accidentally inform an incorporeal (quasi-) substance such as Socrates’s soul,11 we’ve seen above some of the problems to which this answer leads. Sheephood cannot straightforwardly inform Socrates’s soul since it doesn’t make his soul into a sheep, 8 Cohen (1982) argues that the form’s intentional mode of presence is actually a physical event; Haldane (1983) offers textual grounds for reading it physically and also for reading it non-physically. Pasnau (1997) maintains that ‘Aquinas … gives the theoretical outlines of an account but leaves the specific details to be filled in,’ although this ‘lack of specificity can hardly be seen as a weakness in the account’ (41–2). But this isn’t a mere detail that further research could fill in; it’s the brass ring itself. 9 See his Ordinatio, 1, d. 3, p. 3, q. 2, n 422 and the parallel claim in Quodl., 15.6: ‘A thought is something new in us, as we all know by experience.’ 10 To get straight on the details here requires understanding how the substantial form of a composite can itself be the subject of accidents. There is a long story to be told, but for our purposes it’s enough to follow the medievals in talking of the soul as though it were simply a substance, however the details are worked out-roughly, whether we follow the Franciscan tradition in thinking that the soul is substantial in some fashion or follow Aquinas in radically insisting that the soul is just a form. 11 Whether one and the same form can be the substantial form of one object and an accidental form of another is one of the issues raised in discussions of the Incarnation: see Cross (2002), Chs 2–5.

86

Representation and Objects of Thought in Medieval Philosophy

and the reason for this, it seems, has to be sought in either the subject or the mode of inherence of the form. The former tack makes representationality a mysterious feature of minds, the latter postulates a novel and as yet unexplained mode of being, ‘intentional presence’; neither is a case of a simple substance–accident relation. Thus whatever we say about representation, it seems as though the form of the object in the mind should not be identified with the accidental quality inhering in the mind. But the argument that thinking is somehow an accidental quality, given above, still has force. Therefore, we have grounds to postulate not one but two forms in the mind: one that corresponds to the thinking (as an accident), the other that determines what the thinking is about (by conformality). Now obviously these two forms are not independent of one another. Very roughly, we might take the former to be ‘directed’ at the latter, or to ‘include’ it. It doesn’t matter which of these largely metaphorical ways of speaking we adopt, for in the end they come to the same thing, namely the introduction of a distinction between something in the mind and its content. We can either talk of how the form of the external object present in the mind is that which ‘terminates’ the mental act or concept,12 or alternatively talk of how the object of thought is included in or is a part of the concept or the act of thinking; each enshrines the distinction. Take a moment to realize just how extraordinary a move this was. There isn’t any room in the ordinary Aristotelian framework for the distinction, since, on the conformality account (Aristotle’s ‘official’ view), the ‘content’ of a thought is given by the nature of the form involved, and its presence or inherence in the soul just is the occurrent thought – the sort of thing that led Aristotle, and Aquinas in his wake, to speak of the knower becoming the known, a suggestive but obscure claim.13 But as we’ve seen, this simple account, pushed to its extreme in Aquinas, eventually breaks down. Hence the distinction between mental act (or concept) and its content, resulting in a view of thought as essentially composite. The post-Aquinean proposal is therefore that a mental representation represents in virtue of having the same formal content as the external object, so that when Socrates thinks of a sheep he does so in virtue of having a concept whose content is the form of the sheep. Concepts or mental acts can thus be sorted by their (formal) contents.14 Now thinking, as an occurrent mental act, is an accident inhering in the soul as its (quasi-) substance – that is to say, the thought, concept, or act of thinking is present in the soul as in a subject: it is ‘subjectively’ in the soul. The determinate 12 Even Aquinas talks this way occasionally: Summa contra Gentiles, 1.53. 13 These and other dark sayings of the Philosopher are most naturally read as expressing the view that the form of the external object inheres directly in the soul and (somehow) imparts to it the qualities that it engenders in the external object: the soul is ‘assimilated’ to the thing; the mind, as the ‘form of all forms,’ successively becomes each of the things it thinks about. 14 Duns Scotus, Quodlibeta, 15.30: ‘The (intelligible) species also seems to be classified according to the object, not as an intrinsic formal principle but instead as an extrinsic principle’ (Videtur etiam sortiri speciem ab obiecto, licet non sicut a principio formali intrinseco, tamen sicut a per se principio extrinsico). Scotus uses ‘sortiri’ here as the deponent verb ‘to sort or classify’ rather than in its classical sense ‘to select by lot.’

Rethinking Representation in the Middle Ages

87

form, however, is not present subjectively. Hence the coinage of a new vocabulary to express the presence of this second form: it is present ‘objectively’ in the soul, which is to say that it has ‘objective being’ (esse obiectivum) in the soul. This turn of phrase emerged in the work of Hervaeus Natalis, Giacomo di Ascoli, and Duns Scotus (who uses it synonymously with esse repraesentativum and esse deminutum). The virtues of this new approach are evident. It clarifies the underlying ontology, and, at the price of doubling the ‘forms’ involved (one existing subjectively and the other objectively), it seems to provide a clear model for mental representation. Indeed, to the extent that we identify a thing with its substantial form, or at least don’t take its matter to be essential to it, we can now speak of the thing as ‘existing’ in the thought. This isn’t quite real being, of course; it’s only a lesser or ‘objective’ kind of existence (the existence had by objects of thought). Such conceptual contents mediating the thinker and the world can come in handy for a variety of other philosophical problems, especially with regard to thinking about impossible or unreal things, such as universals (for Ockham’s early view) or perceptual and conceptual illusion (as proposed by Peter Aureol). Unfortunately, the drawback of this approach is also evident. Put as a question: What is the ontological standing of such mental contents? The most tempting answer – ‘Nothing’ – won’t do. Nor will it do to say that such conceptual contents are mere extrinsic denominations of the items thought about. Apart from the obvious problem that they will no longer come in handy for nonexistents and impossibilia, since there aren’t any such items to extrinsically denominate, to treat conceptual content as mere extrinsic denomination, i.e. to sort or classify mental acts solely by referring to the external objects they are about, gives up entirely on conformality: there is a form existing subjectively in the sheep, and a distinct form existing subjectively in Socrates’s soul, and that’s the end of it; the form ‘existing objectively’ is really just a way of talking about (‘denominating’) the form in the sheep. Yet if there is no shared form, there is no explanation why a given mental act represents a given external object. The only recourse, then, is to grant them some independent ontological standing. This raises a host of metaphysical problems. What kind of being do they have? Doesn’t this result in a medieval Meinongian ontology? Aren’t they accidents of accidents?15 And so on. William of Ockham recapitulates the public debates over objective being in his own philosophical development. It’s an oft-told tale how Ockham initially accepted an act/content distinction, supporting it with a view of universals as ficta, but faced with arguments from Walter Chatton (among others) eventually found it metaphysically insupportable.16 I won’t go through the details, in part because it is 15 The content of a mental act seems to be an accidental feature of that act, which is in its turn an accident subjectively inhering in the soul – which itself is a form, not properly a substance! The standard reading of Aristotle didn’t countenance accidents of accidents. 16 Ockham describes in some detail, and at different times seems to endorse, at least three distinct theories of the nature of concepts: (i) conceptual content exists objectively in the mental act, that is, something fashioned by the mind that is the object of the mental act,

88

Representation and Objects of Thought in Medieval Philosophy

oft-told, in part because these are problems that are, in the last analysis, problems in the ontology of mind rather than problems having to do with the account of representation. Their net effect, of course, was to deter philosophers by Ockham’s time from explaining representation in terms of conformality, despite its Aristotelian pedigree.17 Fortunately, other candidates were available. 2 Likeness According to (R2) a mental representation represents an object just in case it is similar to the object, or is a likeness of it. Statues literally resemble their subjects, and mirror-images what they reflect; so too mental representations. The genus of representation isn’t confined to the mental, as we noted above. Now it’s clear that (R2) can be combined with (R1), namely by holding that the possession of a form in the mind constitutes the presence of a mental likeness. This is a substantive claim, and it isn’t obviously true. If anything it seems false. Why should the possession of a form automatically lead to there being anything mental we might want to call a ‘likeness’ on independent grounds? (The proviso ‘on independent grounds’ does real work, since it’s trivial that the form in the mind resembles the form in the external object in being the same form, after all.) Hence (R1) and (R2) are distinct. Of course, all the philosophical work for (R2) is done in explaining when one item is ‘like’ another. There were two approaches to this, the second having at least two distinct branches. 2.1 The Literalist Proposal The first approach proposes that one item is ‘like’ another just in case they literally have the same quality. This fits ordinary usage, Latin as well as English. My sweater is like your shirt since each is red. Furthermore, submerged technical terminology is at namely a fictum; (ii) there is a mental quality distinct from the mental act; (iii) the concept is simply the mental act itself. The most plausible interpretation of the presence of these different theories is a developmental hypothesis which leads to (iii) as Ockham’s ‘mature’ theory. According to this developmental reconstruction, Ockham began by endorsing (i), but serious difficulties regarding the ontological status of such fictions forced him to abandon this position in favor of holding that concepts must have real existence or subjective being in the soul, as (ii)–(iii) maintain, and considerations of parsimony eventually pushed Ockham to (iii). See Gal (1967), Read (1977), Adams (1977), Adams (1987), Chapter 3, Tachau (1988), 148–53, Pasnau (1997), 76–85. 17 It’s an interesting question how ‘objective being’ and associated notions re-enter the scene in Late Scholasticism, just in time to be Descartes’s downfall (see Caterus’s First Objections: Adam-Tannery, 7, 92.12–94.4.). Spruit (1994), 280, suggests that it has to do with the use of Aquinas as the basic teaching text during the Counter-Reformation, but, if my analysis is correct, it’s precisely Aquinas’s failure to get clear about intentional being that led later thinkers to the mental act/content distinction, in which case Spruit’s suggestion won’t work.

Rethinking Representation in the Middle Ages

89

work here. For philosophers in the Aristotelian tradition, there are three fundamental forms of identity: among substances, called ‘sameness’; among quantities, called ‘equality’; and among qualities, called ‘likeness’. Hence two items are alike when they are qualitatively identical. Thus items are alike when they literally have the same quality. There are obvious problems in applying this ‘literalist’ proposal to the case of mental representation. How can something immaterial, such as a mental representation, be literally ‘like’ something material? Wouldn’t it resemble anything immaterial much more than it resembles anything material? Likewise, wouldn’t the mind have to actually be red (say) in order to count as representing red through likeness, i.e. qualitiative identity, on this score? William Crathorn endorses the literalist proposal.18 Take the last question first; Crathorn raises it in his second and fourth objections to the proposal (Sent. 1, q.1, n.7, 119.13–15 and 119.24–6): Secondly, if the aforementioned likeness of color were genuine color, then a soul understanding color would be genuinely colored (and a soul understanding heat would be genuinely hot), which is false … Fourthly, the color that is seen by the soul and exists outside the soul would then color the soul itself.

He bites the bullet in his replies (120.30–34 and 121.16–17 respectively): As for the second objection, we declare that the argument holds. A soul seeing and understanding a color is genuinely colored, even with no color existing outside the soul but only its likeness, which is genuine color … As for the fourth objection … I grant that [external] color really causes color in the soul.

Crathorn finds himself driven to this extreme – Robert Holcot jeered that Crathorn’s soul must therefore be a chameleon (see Pasnau 1997, 91) – because he thinks that only a given quality can resemble itself. White cannot be a likeness of red, for instance (117.29–30). Crathorn concludes that only conspecific items can be alike (117.23–5). And since literally the same color is in the stone and in the soul, the first pair of worries about the immaterial and the material don’t arise. The immaterial soul is ‘genuinely colored,’ whatever this may mean for an immaterial object, and so straightforwardly resembles the material colored stone. Crathorn tries to blunt the edge of this paradoxical conclusion by arguing that a form such as redness can be either indivisible and unextended in an immaterial subject, or divisible and extended in a material subject, while nevertheless remaining the same form (120.17–19). Of course, without more explanation these claims won’t help. Yet even if we were to persevere in the face of these difficulties – a price most medieval philosophers thought too high to pay – there is a further problem for the literalist, namely that

18 Tweedale holds that medieval philosophers generally adopt ‘a fairly literal interpretation of the view that species are likenesses of external objects’ (Tweedale, 1990, 36). But who besides Crathorn was a literalist?

90

Representation and Objects of Thought in Medieval Philosophy

the proposal is the same as the simple version of conformality. It is one and the same quality present in the external object and the mind that grounds representation in virtue of engendering the same quality in each subject, which is the essence of conformality.19 Thus likeness, under the literalist proposal, is no better than simple conformality as an explanation of representationality. 2.2 Picturing The second approach to explaining likeness proceeds not literally, as Crathorn tried to do, but through a traditional yet fertile conceit: a mental representation represents what it does – is similar to what it represents – by picturing it. The notion of ‘picturing’ at work here can be taken more or less strictly, of course, but we can broadly divide this second approach into two main branches. On the first of these, the notion of picture or likeness is taken more strictly: a mental representation is an image of what it represents, just as drawings, paintings, statues, photographs, and the like are images of their subjects. A sketch of a sheep, for example, is an image or likeness of the sheep, but it doesn’t literally have the same properties as the sheep. The sketch is an inanimate two-dimensional plane figure; the sheep is none of those things. Furthermore, the claim that a mental representation is an image dovetails nicely with the fact that we have, and often think by using, mental images.20 In addition, the (mental) representation, understood pictorially, embodies something like the ‘form’ of the item represented. The lines and shading on the paper that make 19 Strictly speaking, it isn’t the conspecificity of the shared form but the fact that each engenders the same quality in its possessor that grounds representation. But since it’s the nature of the form that determines the quality it engenders – redness wouldn’t be redness unless its inherence made, or usually made, things red – the distinction is too fine to make a difference. 20 There is a delicate point here. Much of the discussion of mental representation in the Middle Ages takes place in the context of explaining concept-acquisition, arguing over which mental mechanisms have to be postulated to this end, and in particular whether there needs to be an intelligible species. Two of the many jobs performed by the intelligible species are (a) to be representative of the object, and (b) to be impressed on the possible intellect, typically by the agent intellect, thereby reducing the possible intellect from potency to act, which constitutes the mind’s actual thinking of the object. Two features of the intelligible species, evident even from the bare description of (a)–(b), seem not to match up with mental images. First, the intelligible species is that by means of which the object is thought of, not the object of thought itself, whereas mental images are often part of the content of thought. Second, intelligible species are clearly pre-conscious, whereas mental images can be introspectively examined, But neither of these is a barrier to identifying the intelligible species with mental representations as mental images. We may think with mental images, but that isn’t to be confused with thinking of mental images; Socrates thinks of the sheep by calling up a mental image, but it’s another matter for him to think about his mental image of the sheep (rather than the sheep itself). Again, the intelligible species may be generated prior to any conscious thought, but on the standard account it is then stored in memory to be used later; hence one and the same thing can be pre-conscious and also accessed as the object of thought.

Rethinking Representation in the Middle Ages

91

up the sketch of the sheep, for example, arguably preserve various formal features of the sheep: its shape, color, and overall visual appearance. The mental image may likewise preserve formal features of what it represents. These considerations make it plausible that mental representation is a matter of pictorial resemblance. This account avoids the problems that plagued the literalist proposal, since it allows us to hang on to a large amount of dissimilarity between the representation and what it represents. Pictures can fail to resemble their subjects in all sorts of ways. Even photographs, the most ‘realistic’ of representational media, are utterly unlike their subjects. A photograph of Socrates, unlike Socrates himself, is a flat colored sheet of (developed) photographic paper. Hence the immateriality of the mental image need be no barrier to its representing material objects, just as the flatness of the photograph is no barrier to its representing the three-dimensional Socrates. Aquinas makes this point via a distinction between natural and representational likeness. A picture isn’t a natural likeness of Socrates, he asserts, since they don’t ‘agree in their nature,’ but it nevertheless is a representative likeness of him (De veritate, q. 2, art. 3, ad. 9). Pictorial resemblance (representational likeness) is far removed from the way in which twins resemble one another (the natural likeness of the literalist proposal); it need not even be symmetric. If mental representation is to be explained in terms of pictorial resemblance, as contrasted with natural likeness, we need a better understanding of how it works. And here we run into difficulty with the more strict reading of picturing. Think of the sketch of the sheep: an inanimate two-dimensional plane figure. The sheep has none of these properties. What makes this object a representation or image of the sheep? Well, perhaps the sketch has the same color as the sheep, and we might even convince ourselves that the lines traced on the paper are in fact the actual shape of the sheep without the cognitive processing involved in stereoscopic vision (the sketch is sheep-shape). The lifelessness of the sketch, on the other hand, is a characteristic, along with its size and other features, that is not to be taken into account in explaining how the sketch pictures the sheep. Even putting aside the difficulty in distinguishing features that matter from those that do not, this explanation cashes out pictorial resemblance in terms of literal likeness, that is, through the natural resemblance of some features: the same color, the same shape, and the like. So too with the mental image of the sheep, here setting aside the immaterial medium in which the sheep’s color is exemplified. The fact that these instances of literal sameness are surrounded by other dissimilar or discounted features is irrelevant. It seems as though all the problems with the literalist proposal haunt the more strict reading of pictorial resemblance.21 A way out of this dilemma is to interpret pictorial resemblance less strictly, the ‘second branch’ mentioned above. Imagine a monochrome sketch of a sheep. The 21 Many medieval philosophers use imagistic terminology to describe mental representations without apparent worry over the literalism such terminology seems to entail. But they should be worried even if they aren’t, and without telling more of a story we have no real explanation of representation.

92

Representation and Objects of Thought in Medieval Philosophy

sheep isn’t literally colored a shade of gray, but that gray shade rather than another corresponds to the sheep’s rich and creamy merino color. Likewise, the sheep isn’t a two-dimensional closed plane figure, but is sketched as one according to the ‘laws’ of perspective. In short, none of the sheep’s properties are literally present in the representation; the representation, instead, pictures what it does in virtue of having intrinsic features that correspond to the properties of the external object. The image need not ‘look like’ its subject at all, as long as the appropriate correspondence holds. An architectural blueprint for a building, or a circuit diagram for an amplifier, represent their subjects without any ‘natural’ resemblance. Nowadays philosophers speak of correspondence, projection-rules, mapping, or transformation-rules, but the underlying idea is recognizably the same. A (mental) representation represents what it does in virtue of having features that systematically correspond to the properties of the represented object (and perhaps other, irrelevant, features as well) according to some scheme. Likeness is a matter of ‘picturing’ in this extended sense.22 The virtues and the vices of such a correspondence-account of representation are familiar.23 But its most appealing feature to medieval philosophers deserves special mention. When a transformation-rule is applied to some item, the result is, ideally, something with features that systematically correspond to properties of the original item. What is it that such a transformation-rule preserves? Well, the natural answer is: form. (Nowadays people say ‘structure’ but that’s an acceptable translation of forma.) It is because the sheep is the way it is that the pattern of lines and shading on the sketchpad is the way it is, and the two-dimensional relations among lines is a ‘projection’ of the three-dimensional volumetric relations in the world. The one pattern is not the other, but it is the transformation of the other, and, if the transformation-rule is a good one (in some sense to be spelled out), we can say that it has the same form. This happy meeting of (R1) and (R2) made the attractions of a general theory of pictorial resemblance a clear winner for medieval philosophers, and they helped themselves to it freely. For example, Roger Bacon wrote that mental representations signify things ‘according to conformality (conformitatem) and the configuration of one thing to another in its parts and proper characteristics, the way images and pictures and likenesses and so on do’ (De signis, §5, 83). Seductive as this picture of picturing is, it only qualifies as an explanation of the representationality of mental representation if it is supplemented by a full (or at least fuller) account of the natural transformation-rules embodied in sense and intellect, as well as of the transformed ‘analogous’ features in the mind. Unfortunately, not only did mediaeva philosophers not provide such an account, there isn’t any sign they ever even tried to, and so leave us in the end with no more than a suggestive conceit. 22 Sellars (1960) tries to re-develop this theory in light of modern concerns with isomorphism and picturing. His attempt is noteworthy for his exploitation of the (neoAquinean) idea that the features of the representation are the ‘analogous’ properties systematically correlated by some scheme with the actual properties of the represented item. 23 Here’s a sample difficulty: is a sketch of a sheep shaded gray a monochrome representation of a merino sheep or a colorful representation of a gray sheep?

Rethinking Representation in the Middle Ages

93

Yet medieval philosophers are hardly the only ones to leave ‘picturing’ at a more intuitive level than they should. We might be more tolerant here than in the case of ‘intentional being’: the latter seem like an obfuscatory word, whereas there is a rich body of theory surrounding pictorial representation. We might even be inclined to praise medieval philosophers for their restraint, in not designing theories when they had no way of making good on them.24 But this tolerant attitude misses the point. Without anything more to say about how picturing does its work, the lack of an articulated theory ultimately leaves mental representation mysterious – it too misses the brass ring – and renders it vulnerable to a surprise attack from another quarter. William of Ockham discusses the nature of representation carefully in Ordinatio, 1, d. 3, q. 9, where he takes up the question whether creatures are somehow indications of their Creator. He begins by distinguishing two kinds of representation: images (imagines) and impressions (vestigia). The paradigm of the former kind is a statue of Hercules, of the latter an animal’s hoofprint, but Ockham is clear that these categories of representation are much wider; an ‘image’ can be any univocal effect at all, even if not intended as such.25 It’s clear that he is working with the generalized notion of pictorial representation described above; he elsewhere points out that the image can be entirely dissimilar to that of which it is the image and yet represent it.26 Such cases of pictorial resemblance are instances of representation, Ockham argues; they differ from what they are of (what they depict), and they lead to the notion of what they are of through acquaintance with it.27 And Ockham grants that images of all kinds do in fact represent their subjects. But at the very least they can’t be the whole story, he maintains, since they are intrinsically general, and therefore can’t explain what we nowadays call ‘singular thought.’

24 Pasnau (1997) takes this tack on Aquinas’s behalf (his emphasis): ‘One might say that Aquinas doesn’t have a theory of representation at all, in the sense that he doesn’t give a determinate account (of the mechanisms behind representational likeness) … It is one of the merits of Aquinas’s approach, I would suggest, that he does not rest his account of mental representation on any particular kind of likeness’ (112). One indeed might say. 25 Ockham gives three senses of ‘impression’ in Ordinatio, 1, d. 3, q. 9 (OTh, 2, 548.8–549.2) and of ‘image’ in q. 10 (553.2–25), the strictest of which is the statue of Hercules and the broadest of which is ‘anything univocally produced by another.’ 26 In Ordinatio, 1, d. 2, q. 8 (OTh, 2, 277.3–278.12), Ockham approvingly cites Augustine, who emphasizes the lack of similarity between the picture and what it is said to depict-indeed, Augustine emphasizes the arbitrariness of the picture. For example, Augustine describes imagining the city of Alexandria, which he had never seen, and notes that it would be miraculous if it were anything like Alexandria; equally, when reading the Bible, one can fashion mental images of the Apostles and of Christ which are probably quite unlike their actual appearance. Aquinas would call it a lack of natural, not representational, likeness. 27 See Ockham’s analysis of ‘representation’ in Quodl., 4.3 (OTh, 9, 310.9–19). There are nuances having to do with whether a mental representation leads to what it represents immediately or through a mediating notion, and if the latter whether memory must be involved, but they aren’t essential to Ockham’s general attack on pictorial resemblance.

94

Representation and Objects of Thought in Medieval Philosophy

Ockham claims that impressions and images, by their very nature, represent no one individual any more than another individual that is extremely similar to it (simillimum: 546.6–8). A moment’s reflection on pictures illustrates why his contention is correct. A photograph by itself will not determine whether it is a picture of Socrates or a picture of Socrates’s twin brother. Whether Socrates has a twin brother is a fact about the world, not about the photograph, and so is not settled by the intrinsic features of the photograph. Ockham returns to this claim in Reportatio, 2, qq. 12–13, pointing out that the intellect couldn’t distinguish which of two extremely similar whitenesses might be the individual quality a pictorial mental representation was trying to represent (OTh, 5, 281.24–282.12). Nothing turns on the particular example; Ockham repeats it in a more detailed version using two equal amounts of heat (287.19–289.7), and once again with two men (304.6–20).28 Yet the problem isn’t due to indiscernibility, in the sense that we inspect the image and can’t then determine what it is an image of. We needn’t be consciously aware of our mental representations. Ockham’s point is that images, conscious or not, are by their nature applicable to many-that the correspondence-rules aren’t guaranteed to have unique inverses (i.e. the rules don’t in general yield one-to-one mappings).29 But since we can and do think about individuals, mental representation must not be solely a matter of pictorial resemblance. The upshot is that ‘likeness isn’t the precise reason why we understand one thing rather than another’ (similitudo non est causa praecisa quare intelligit unum et non aliud: 287.17–19). Images do represent things, but they aren’t sufficient to represent individuals as individuals. Given the lack of detail about how the transformation-rules in fact work, the net result of Ockham’s attack is to make the siren-call of pictorial resemblance even less attractive. Despite their Aristotelian credentials, neither (R1) nor (R2) can, in the end, provide a satisfactory account of mental representation. Hence medieval philosophers turned elsewhere to clarify representation, namely the second-string choices: covariance and linguistic role, whose combination at the start of the fourteenth century marked a new departure.

28 Ockham makes the same claim, in the same context, with regard to intuitive cognition in Quodl., 1.13 (Oth, 9, 76.89–98): see §3 below. 29 Does Ockham’s argument work in general? There’s certainly no reason in principle why transformation rules can’t be one-to-one. But given that we’re interested in mental representation, we might argue from the known limits of perceptual distinguishability to limitations on the information the intellect can make use of, at least under the assumption that mental representations are preprocessed by the senses. We could always grasp the other horn of Ockham’s dilemma, too, and deny that we have singular thought. But there is a better reading of Ockham’s argument available. He could be taken as pointing out that knowledge of the (inverse of the) transformation-rule is distinct from knowledge of the mental representation itself, and thus, to the extent that representation is encapsulated in the image, it does not of its nature determine what it represents. (It might represent many things.) Put a different way, Ockham is objecting that the representationality of the representation isn’t a matter of its intrinsic features but depends on further knowledge of transformation-rules. On that score he seems right.

Rethinking Representation in the Middle Ages

95

3 Covariance and Linguistic Role According to (R3), a mental representation represents an object, at a first approximation, just in case it is caused by that object. The mental representation is present in the presence of the item and absent in its absence: the object and the representation ‘co-vary.’ Unlike (R1) and (R2), there was a rich body of theory on causality available, ready to be used to supplement the causal account account of representation. Furthermore, covariance fits nicely with the (largely) causal account of perception and thought in which it is to be embedded, as noted at the start of §1.1 above. The horse’s hoofprint represents the horse, as the impression in the sealing-wax represents the seal. The hoofprint or impression is a sure sign that the appropriate causal agent has been at work in the vicinity. Now obviously this net is cast too widely; we might balk at saying that a sunburn represents the Sun, that smoke represents fire, that the child represents the parent. We can loosely speak of the result of any causal activity as an ‘impression,’ but more strictly an impression is only something left behind as the consequence of some proper causal activity (and better yet left through the causal activity of only part of the agent such as the hoof), as Ockham tells us in Ordinatio, 1, d. 3, q. 9, so that we don’t count ordinary univocal causality as representational (OTh, 2, 548.20–549.2). A mental event that occurs as the result of an object’s causal activity counts as an impression in this restricted sense, so that the thought Socrates has upon seeing the sheep, as an impression, represents the sheep – at least, so long as Socrates’s thought covaries with the sheep. The intuition behind (R3), then, is that the thoughts we have when we look at sheep are thoughts of the sheep in virtue of the fact that they are the thoughts sheep naturally and regularly cause us to have.30 There is no need to suppose further that such thoughts involve the sheep’s form in any substantive way, as required by (R1). Likewise, there is no need to suppose further that such thoughts ‘resemble’ the sheep, as required by (R2), except for the trivial case where the projection-rule is just causal impression. Hence (R3) is independent of (R1) and (R2). According to (R4), a mental representation represents an object whenever it signifies that object, i.e. to the extent that it functions as the (mental) ‘word’ for the object. Christian doctrine provided a source for interpreting thought as a form of ‘inner language’; by Ockham’s time it was well-understood that angels were telepaths who communicated in the language of thought.31 Now the idea that thoughts are somehow language-like has a long history and was exploited for various purposes, not always mutually compatible, and a lot of work needs to be done to clarify the vague outlines of the proposal. What is relevant to our purposes is the suggestion that a concept represents what it does – it is the concept it is – only if it is connected 30 I’ve adapted the characterization of covariance given in Cummins (1989), 36, there attributed to Locke; it fits the medieval case as well or better than the early modern case. 31 Aristotle’s suggestion in De interpretatione, 1, 16a, 3–4 that thoughts are another level of language was taken to provide independent philosophical confirmation of this Christian view, despite the fact that on Aristotle’s account thoughts were more like the semantics of a language than like language itself.

96

Representation and Objects of Thought in Medieval Philosophy

to other contentful concepts in the appropriate ways, as words in a language are. Roughly, Socrates’s concept |sheep| represents a sheep in virtue of its having the right (linguistic) role: it is subordinate to |animal| and |living creature|, able to be the subject in propositions, a constituent of the belief that sheep are edible, and the like. Anything that plays this complex role is ipso facto the concept |sheep|, that is, represents sheep. There is no need to suppose further that whatever fills this role will somehow exemplify the sheep’s form, as (R1) would have it; nor that it ‘resemble’ the sheep, as (R2) would have it. Hence (R4) is independent of (R1) and (R2). It is clearly independent of (R3) as well, since it is no part of the linguistic role played by a mental item how it came to be; nor does the simple covariance sketched in the preceding paragraph determine that the relations among concepts will work out appropriately. Although (R3) and (R4) are distinct, their combination is powerful. It is no less than a medieval version of functionalism, the idea that determinate content is fully specified by inputs (covariance) and outputs (linguistic role).32 Unlike modern functionalism, medieval functionalism is holist only in a shallow sense, since human mental structure was understood to be innate and fixed rather than individually variable. But the meaning of terms is cashed out by their place in the (determinate) structure of thought, which itself constitutes a language. The efflorescence of theories of Mental Language and the flurry of research on causality at the beginning of the fourteenth century – in each of which Ockham played an important part – underwrite this new approach in psychology to mental representation. Mental Language, especially Ockham’s theory of Mental Language, has been the subject of intense investigation for several decades; there are several high-level accounts of it available.33 In addition to providing a framework for logic (to which Ockham devotes the Summa logicae), Mental Language also claims to be the truth about cognitive psychology, making good on the claim that there is a ‘language of thought’ in staggering detail. For our purposes, it’s enough to note that mental representation is going to be generally explained, at least on its functionalist ‘output’ side, in terms of Mental Language. Covariance will be needed to explain the ‘input’ side and, in particular, to resolve the problem Ockham took to be fatal for (R2): how singular thought is possible, or, linguistically, how we can explicate proper names in Mental Language. As we’ve seen in §2, Ockham maintains that likeness isn’t ‘the precise reason’ why we think of one thing rather than another. He argues instead that covariance is the correct explanation. He begins his discussion in Ordinatio, 1, d. 3, q. 9 by making the point that impressions (vestigia) by their nature are general, just as images are: 32 Likewise today: Fodor (1987) combines his ‘Language-of-Thought hypothesis’ (namely that mental representations are language-like symbols) with the ‘crude causal theory’ that symbol tokenings denote their causes and symbol types express the property whose instantiations reliably cause their tokenings. Fodor’s theory is exceptional in recognizing (R3) and (R4) as distinct components, though the theory is common enough. 33 See for example Adams (1987), Chapter 10, and Normore (1990).

Rethinking Representation in the Middle Ages

97

a hoofprint might have been made by any horse, and no amount of inspection will determine which horse made a given hoofprint (OTh, 2, 546.6–8). But impressions differ from images in that ‘it’s part of the very notion of an impression that it be caused by that of which it is the impression’ (547.6–7). An image or resemblance need not be fashioned from the original, whereas an impression must be. More exactly, Ockham holds that it is the nature of an impression to be producible by a given individual rather than another, i.e. that it is apt to be so produced, even if God supplants the causal chain. He states his view succinctly in Quodl., 1.13 (76.89–96): Intuitive cognition is a proper cognition of a singular not because of its greater likeness to one than another but because it is naturally caused by the one and not by the other; nor can it be caused by the other. If you object that it can be caused by God alone, I reply that this is true, such a sight is always apt to be caused by one created object and not by another; and if it were caused naturally, and if it is caused naturally, it is caused by the one and not by the other, and it is not able to be caused by the other.

The point is reiterated in Reportatio, 2, qq. 12–13 (OTh, 5, 289.8–18): Suppose you were to object that a given concept (intentio) can be immediately and totally caused by God, and so through that given concept the intellect would no more understand one singular than another extremely similar one, since it would be as much similar to one as to the other; nor does causality make it be of one and not of the other, since it is caused by neither but rather immediately by God. I reply that any given concept of a creature that is caused by God can be partially caused by the creature, even if it weren’t actually so caused. Hence a given singular is cognized through that cognition by which it would be determinately caused were it caused by a creature; this is a feature of one thing and not another; therefore, etc.

(Ockham says’partially’ in his reply because he holds that God is a necessary cocause of any effect.) Thus Ockham rejects (R2) in no uncertain terms, insisting on a counterfactual causal account of singular thought, that is, he endorses a causal theory of proper names in Mental Language.34 Thus Ockham epitomizes the philosophical struggles of his generation in bringing psychology to a new functionalist paradigm in place of the old conformality and likeness theories. Beginning in the fourteenth century, philosophers had a new way of thinking about mental representation, one that may have had less pure Aristotelian roots but looked more promising as a theory. Yet there is a sense in which they gave up on the notion of representation entirely.

34 Pasnau (1997) says that ‘Ockham is at best tentatively moving away from a likeness account of mental representation’ (105 my emphasis). There isn’t anything tentative about it. (He mentions this on 105 n.45 but doesn’t give the point its due.) Tabarroni (1989) gets it right when he claims that Ockham ‘abandoned the iconic model’ of mental representation (214).

98

Representation and Objects of Thought in Medieval Philosophy

Conclusion: Against Representation Ockham is notorious for his attack on the intelligible species, arguing inter alia that the intelligible species isn’t needed for the purposes of mental representation, one of its traditional roles (Reportatio, 2, qq. 12–13; OTh, 5, 272.17–20). Although he preserves the traditional terminology, declaring that ‘the act of understanding (intellectio) is the’likeness’ of the object’ (287.15), it’s clear that this is an empty formula: as noted in the citation at the beginning of this paper, ‘the object sufficiently represents itself in a cognition’ (274.4–5), a point Ockham later repeats: ‘[the object] can be present qua object to the intellect, without any species’ (300.1–2). It’s not that Ockham thinks there are no mental events. Rather, there is no need for mental representation as traditionally conceived. There are mental acts of thinking, but there is no need to postulate independent contents, or indeed any discernible intrinsic structure to the mental act; it is what it is in virtue of its functional inputs and outputs, not because of its inner nature. On his mature theory of mind, Ockham countenances only the spartan ontology of mental acts of thinking, which are then paired with their external objects directly, not requiring any mediation. In short, Ockham, at least in his mature view, argues against what is traditionally called a ‘representationalist’ theory of mind and for what is usually called ‘direct realism.’ The final result of rethinking representation in the Middle Ages, then, is to junk it. Final logically, that is, not historically; Ockham had few followers in psychology, and the discredited accounts of representationality hung around long enough to be the targets of abuse from Hobbes, Descartes, and Locke.35 Direct realism is sometimes portrayed as the simple initial position, the shortcomings of which lead to more complex forms of representationalism. The history and development of medieval philosophy of psychology shows otherwise. Bibliography Primary Sources Albert the Great, Alberti Magni opera omnia, ed. Berhardt Geyer et al. Aschendorff: Munster Westfalen 1940–64. Aquinas, Thomas, Sancti Thomae de Aquino opera omnia, iussu Leonis XIII P.M. ed. Cura et studio Fratrum Praedicatorum 1876–. Bonaventure, Sancti Bonaventurae opera omnia, ed. Collegium sancti Bonaventurae, Ad Claras Aquas (Quaracchi) 1882–1902.

35 It’s an interesting historical question why Mental Language seems to have played itself out, and for that matter why the discredited accounts of representation weren’t immediately discarded – it’s even a good question why Ockham had few followers in psychology – but these are historical questions whose answers don’t alter the logic of the positions sketched here.

Rethinking Representation in the Middle Ages

99

Duns Scotus, John, Iohannis Duns Scoti Doctoris Subtilis et Mariani opera omnia, ed. P. Carolus Balic et al. Typis Polyglottis Vaticanae 1950–. ——, Obras del Doctor Sutil Juan Duns Escoto (edicion bilingue): Cuestones Cuodlibetales, Introduccíon, resúmenes y versíon de Felix Alluntis, Biblioteca de autores cristianos: Madrid 1968. Henry of Ghent, Summae quaestionum ordinariarum, Paris 1520; reprinted by the Franciscan Institute, St Bonaventure, NY, 1943. Matthew of Acquasparta, Quaestiones disputatae de fide et cognitione, ediderunt Collegium sancti Bonaventurae, Ad Claras Aquas (Quaracchi) 1957; Quaestiones disputatae de anima XIII, A-J. Gondras, Études de philosophie médiévale, vol. 50 (1961). William Crathorn, Sententiae 1, ed. Fritz Hoffman, Crathorn: Quästionen zum ersten Sentenzbuch, Einführung und Text. Beiträge zur Geschichte der Philosophie und Theologie des Mittelalters, Neue Folge, Band 29, Ashendorff: Munster 1988. William of Ockham, Guillelmi de Ockham opera philosophica (OPh), tom. 1–8. Cura Instituti Franciscani Universitatis S. Bonaventurae. Ediderunt Stephanus Brown et al. S. Bonaventure, NY: impressa Ad Claras Aquas (Italia) 1974–88. ——, Guillelmi de Ockham opera theologica (OTh), tom. 1–10, Cura Instituti Franciscani Universitatis S. Bonaventurae, ed. Stephanus Brown et al. S. Bonaventure, NY: impressa Ad Claras Aquas (Italia) 1967–86. Secondary Sources Adams, M.M. (1977), ‘Ockham’s Nominalism and Unreal Entities’, Philosophical Review, 86, 144–76. —— (1987), William Ockham, vols 1—2, Notre Dame: University of Notre Dame Press. Cohen, S. (1982), ‘St Thomas Aquinas on the Immaterial Reception of Sensible Forms’, Philosophical Review, 91, 193–209. Cross, R. (2002), The Metaphysics of the Incarnation, Oxford: Oxford University Press. Cummins, R. (1989), Meaning and Mental Representation, Cambridge: MIT Press. Fodor, J. (1985), ‘Fodor’s Guide to Mental Representation’, Mind, 94, 55–97; reprinted in A Theory of Content and Other Essays, Cambridge: MIT Press 1990, 3–29. —— (1987), Psychosemantics, Cambridge: MIT Press. Gal, G. (1967), ‘Gualteri de Chatton et Guillelmi de Ockham controversia de natura conceptus universalis’, Franciscan Studies, 27, 191–212. Haldane, J. (1983), ‘Aquinas on Sense-Perception’, Philosophical Review, 92, 233–9. Haugeland, J. (1990), ‘The Intentionality All-Stars’, in James A. Tomberlin (ed.), Philosophical Perspectives 4: Action Theory and Philosophy of Mind, Ridgeview Publishing Company, 383–427.

100

Representation and Objects of Thought in Medieval Philosophy

Hoffmann, F. (1998), Ockham-Rezeption und Ockham-Kritik im Jahrzehnt nach Wilhelm von Ockham in Oxford 1322–1332, (Beiträge zur Geschichte der Philosophie und Theologie des Mittelalters Neue Folge, Band 50), Ashendorff: Munster. King, P. (1994), ‘Scholasticism and the Philosophy of Mind: The Failure of Aristotelian Psychology’, in Tamara Horowitz and Al Janis (eds), Scientific Failure, New York: Rowman & Littlefield, 109–38. Marrone, S.P. (1985), Truth and Scientific Knowledge in the Thought of Henry of Ghent, The Medieval Academy of America. Normore, C. (1990), ‘Ockham on Mental Language’, in J.C. Smith (ed.), The Historical Foundations of Cognitive Science, Dordrecht: Reidel, 53–70. Pasnau, R. (1997), Theories of Cognition in the Later Middle Ages, Cambridge: Cambridge University Press. Read, S. (1977), ‘The Objective Being of Ockham’s Ficta’, Philosophical Quarterly, 27, 14–31. Sellars, W. (1960), ‘Being and Being Known’, The Proceedings of The American Catholic Philosophical Association, 34, 28–49; reprinted in Science, Perception, and Reality, London: Routledge & Kegan Paul, 41–59. Spruit, L. (1994), Species intelligibilis: From Perception to Knowledge, vol. 1: Classical Roots and Medieval Discussions; vol. 2: Renaissance Controversies, Later Scholasticism, and the Elimination of the Intelligible Species in Modern Philosophy, Leiden: Brill. Stump, E. (1999), ‘The Mechanisms of Cognition: Ockham on Mediating Species’, in Paul Spade (ed.), The Cambridge Companion to William of Ockham, Cambridge: Cambridge University Press, 168–203. Tabarroni, A. (1989), ‘Mental Signs and the Theory of Representation in Ockham’ in Umberto Eco and Costantino Marmo (eds), On the Medieval Theory of Signs, (Foundations of Semiotics, vol. 21), John Benjamins, 195–224. Tachau, K. (1988), Vision and Certitude in the Age of Ockham: Optics, Epistemology, and the Foundations of Semantics 1250–1345, (Studien und Texte zur Geistesgeschichte des Mittelalters, Band 22), Leiden: Brill. Tweedale, M. (1990), ‘Mental Representations in Later Scholasticism’, in J.C. Smith (ed.), The Historical Foundations of Cognitive Science, Dordrecht: Reidel, 35–51.

Chapter VI

William Ockham and Mental Language Mikko Yrjönsuuri

In many different connections throughout his whole oeuvre, William Ockham puts forward the program of an ideal mental language. This language is presented as a universal representative system in which all thinking takes place and which lies at the background of all communication. In addition to people, it is used by all possible thinking beings, angels as well as the God. Most obviously, Ockham would have supposed machines to use this language as well, if he had presented fantasies of thinking machines. Also, Ockham clearly thought that everything that can be expressed in any language could also be expressed in the mental language. This is because all expressions of the spoken languages are subordinate to expressions of mental language, so that all meanings of spoken words are derivative upon those of mental terms. My claim in this chapter is that not only everything that there is in the world, but also everything true or false that can be said about the world, can be expressed in this language. In the twentieth-century debates, Ockham’s theory of mental language has often been interpreted in relation to the concept of logical form. In the background, there has been the idea that the mental language reveals with complete transparence the true content of the linguistic expression, unlike the sentences of spoken language. From such a point of view, Ockham’s theory appears primarily as a theory that explains the deductive relations of sentences and gives the keys to define the truthvalues of sentences. Thus, the theory appears to belong to the field of logic in the twentieth-century sense. The recent interpretive debate of Ockham’s theory has led to many problems. It seems to me that they largely arise from leaning on a comparison with logical analysis in the interpretation. This strategy doesn’t do justice to Ockham’s intentions in drafting his theory of mental language. In this chapter, I try to examine Ockham’s theory from another kind of perspective – from the perspective of a universal language. I want to sketch Ockham’s mental language as an ideal representational system, and look how the program would work for that purpose. In his book Cartesian Linguistics (1966), Noam Chomsky discusses seventeenthcentury theories of language. Chomsky’s main aim is to point out in what way the forerunners to his own notions of the universal deep structure of all languages can be found in the seventeenth-century linguistic theories, which he calls Cartesian – though acknowledging the problems of that label. In this chapter, my aim is to

102

Representation and Objects of Thought in Medieval Philosophy

compare Chomsky’s picture of Cartesian linguistics to Ockham’s sketchy outline of the mental language. I believe that we can get a better picture of Ockham’s theory if we consider it with linguistic models in view, and not just from the logical point of view. I will not evaluate Chomsky’s interpretation of the seventeenth century at all but instead use the picture drawn by him as a heuristic background to my own interpretation of Ockham’s theory. Chomsky’s Picture of the Seventeenth-Century Linguistics and Ockham’s Theory In his story, Chomsky poses as the leading principle of seventeenth-century linguistics the idea of language as enabling the creativity belonging specifically to human life in distinction from the rest of the animals. Chomsky refers (3–6) to a passage in Descartes’ Discourse on the Method concerning the differences between humans and the other animals. According to Descartes, the metaphysically significant difference is, of course, that humans have minds whereas the beasts only have bodies. As a practical criterion for the presence of the mind in an external being – human, animal or machine – Descartes presents rational use of language. Descartes thinks that machines or beasts can be trained to repeat sentences but not to use language rationally in varying situations. Chomsky links this rational use of language with creativity. Chomsky interprets his Cartesian starting point so that there appears the idea of language as the universal medium of thought, or as the condition for the occurrence of rational thinking. According to Chomsky, seventeenth-century philosophers thought that language offers the medium, which enables creative thought. The special structure of language makes it possible to construct from a limited number of elements, i.e. words, an unlimited number of sentences – or thoughts. In other words, language gives humans their specific ability of self-expression and a creative potential that differs from the capacities of the beasts. Admittedly, Ockham does not develop the idea of the creative power of language. Still, it is clear that he regards language precisely as such a universal medium of thought as Chomsky saw the seventeenth-century philosophers speaking about. In Chomsky’s words, ‘human language, in its normal use, is free from the control of independently identifiable external stimuli or internal states and is not restricted to any practical communicative function’ (Chomsky, 1966, 29). Ockham, too, treats mental language – the exemplar of all spoken languages – as an uncontrollable universal system that governs all language use, and specifically the internal discourse that may or may not be expressed in external speech. Non-communicative use of mental language is clearly visible in question V, 5 of Ockham’s Quodlibetal Questions, where he treats beliefs on the one hand as structures expressed in mental language, and on the other as propositional attitudes to such linguistic structures as formulated or represented in one’s own mind. When shaping one’s beliefs, in Ockham’s opinion, humans formulate them as linguistic

William Ockham and Mental Language

103

structures in the mind. In such cases, the language is not used communicatively – it is used as a medium of representation. According to Ockham, judgement or belief is to be seen as a process directed by the will. The mind decides to take some sentences formulated in the mental language as true and others as false. Both in Chomsky’s discussion of seventeenth-century linguistics and in Ockham’s treatment of mental language, the main focus is precisely on the deep structures that appear in the mental language. Chomsky refers to the Cartesian metaphysical foundation: the dualism of mind and body. In Chomsky’s mind, the language expressed this dualism to the seventeenth-century thinkers as well, for there are linguistic structures on the level of thinking as well as of matter. (I’ll soon come back to the fact that Ockham distinguishes as many as three levels of language: mental, spoken and written, though it is of course clear that spoken and written language are much closer to each other than to the mental.) The external, corporeally conditioned use of language is not as ideally structured as the internal use of language on the mental level. It is possible to talk about the surface structure of language, which appears in spoken language and the ideal deep structure, which appears in thinking. At this point, Chomsky refers to the seventeenth-century enterprise to formulate a universal grammar (grammaire générale). This grammar would describe the grammatical structures that all languages necessarily contain, for it would concentrate on the necessary linguistic conditions of thinking. This universal grammar would contrast to the particular grammars of different spoken languages. These particular grammars, of course, contain a variety of arbitrary structures. The universal and necessary basic structures of language – if there are such – must belong innately to the human mind, Chomsky argues. From this basis, he examines the famous seventeenth-century debates concerning ‘Innate Ideas’ and ‘Common Notions’. For Chomsky the main issue is, above all, that the linguistic faculties must to a large extent be innate in humans, and actually only the use of external, spoken language requires a certain process of learning. Such a view of the innate character of language is perhaps the clearest difference between Ockham’s theory and the Cartesian linguistics described by Chomsky. Ockham takes the firm position that the mental language has to be learned. Humans do not possess it innately – or at least any of its vocabulary. Ockham did not believe in ‘innate ideas’ if understood in the Chomskian way. The case of ‘common notions’ is more complex, and indeed one way of putting my discussion of Ockham is to take it as addressing the problem of whether and how we learn the universal grammar of mental language. For the purposes of this paper, I am putting aside most of the problem of how one conceptual sign or word represents particular things, and considering the issue of how these signs are organized into a full representational system – into a language with capabilities for making statements and other kinds of speech acts.

104

Representation and Objects of Thought in Medieval Philosophy

The Grammar of Mental Language Ockham’s most detailed version of the grammar of the mental language can be found in the beginning of Summa logicae. This location has of course suited perfectly the interpretation that the theory really belongs to the field of logic in the twentiethcentury sense. But it has to be remembered that in the fourteenth century the concept of logic (logica) differed considerably from the current concept. At that time, it also covered many themes that nowadays would be clearly connected to philosophical semantics or even linguistics. In the first chapter of the first book of Summa logicae, we read as follows (trans. Loux, 1974, 49): As Boethius points out in his Commentary on the first book of the De Interpretatione, discourse is of three types – the written, the spoken, and the conceptual (this last existing only in the mind). In the same way there are three sorts of terms – written, spoken and conceptual. The written term is a part of a proposition which has been inscribed on something material and is capable of being seen by the bodily eye. The spoken term is a part of a proposition which has been uttered aloud and is capable of being heard with the bodily ear.

Ockham’s starting point is that mental language is essentially parallel with spoken and written language. As the sentences of spoken Latin are composed of words, in the mind the sentences of mental language are composed of words as well. An essential difference between these levels appears, when we read Ockham’s description of the mental terms. The text continues as follows (trans. Loux, 1974, 49–50): The conceptual term is an intention or impression of the soul which signifies or consignifies something naturally and is capable of being a part of mental proposition and of suppositing in such a proposition for the thing it signifies. Thus, these conceptual terms and the propositions composed of them are the mental words which, according to St Augustine in chapter 15 of De Trinitate, belong to no language. They reside in the intellect alone and are incapable of being uttered aloud, although the spoken words which are subordinated to them as signs are uttered aloud.

After about a half-page, Ockham further clarifies the naturalness of the semantic relations of mental language (trans. Loux, 1974, 50): Now, there are certain differences among these three kinds of terms. For one thing the concept or impression of the soul signifies naturally; whereas the spoken or written term signifies only conventionally. This difference gives rise to a further difference. We can decide to alter the signification of a spoken or written term, but no decision or agreement on the part of anyone can have the effect of altering the signification of a conceptual term.

All in all, Ockham makes it very clear that mental language is the level that explains the meaningfulness of language on the other levels. Without a connection to the

William Ockham and Mental Language

105

mental language, written or uttered words would be mere material phenomena without any linguistic meaning. On the other hand, the terms of mental language mean naturally what they mean. There are of course numerous problems with this thesis, but in the texts cited above Ockham at least makes it clear that he thought that each term of the mental language is a natural sign. This kind of ‘intention of the soul’ is by its essence such that it means something, and according to Ockham’s definition of signification (significatio) it makes the subject think about the thing whose sign it is (cf. Summa logicae, I, 33; Loux, 1974, 113–14). Ockham does not claim that the naturalness of the signs of the mental language would suffice to show that the perceiving of the mental term would necessarily mean understanding what is being thought. We can see this from his discussion of the communication between angels, which he explains with an analogy to mindreading. He ends up claiming that a mind-reading angel would not understand the read thoughts unless he has at some time got acquainted with the things that the terms of the read thoughts mean. In order to know that the other angel is thinking about, say, cats, the mind-reading angel would have to be acquainted with cats. In other words, a sentence of mental language is intelligible only to a subject that can itself produce all the parts of the mental sentence as its own thoughts. In normal human thought this condition is of course always fulfilled – everybody is of course able to produce his or her own thoughts – but in an imaginary mind-reading situation this condition may not be fulfilled. Ockham’s primary means to characterize the structures of mental language is to examine features of Latin in respect to whether they have to have an equivalent in the mental language. The starting point is that the mental language has its own grammar in the same way as Latin has. Ockham writes (Summa logicae, I, 3; trans. Loux, 1974, 52): In the case of spoken and written language terms are either names, verbs or other parts of speech (i.e., pronouns, participles, adverbs, conjunctions, prepositions); likewise, the intentions of the soul are either names, verbs or other parts of speech (i.e., pronouns, adverbs, conjunctions, prepositions).

Most obviously, it was this text that made Peter Geach claim in his famous book Mental Acts that Ockham simply copied random structures of Latin to necessary structures of the mental language. As John Trentman had already remarked in 1970, Ockham in fact did not make a mistake like this. The fact that Ockham believed that mental language had to have a grammar, and that this grammar was easiest to sketch in relation to Latin, did not mean that Ockham directly copied the Latin grammar as the grammar of mental language. In fact, Ockham had in his mind a distinction fairly similar to the one that Chomsky discussed as a distinction between particular and universal grammar. According to the criterion used by Ockham, the grammar of mental language must be able to express all the structural differences between Latin expressions that cause differences in truth-values. Thus, anything having an effect on the truth-value

106

Representation and Objects of Thought in Medieval Philosophy

must have an equivalent in the grammar of the mental language. On the contrary, features of Latin that are insignificant in respect to truth-values are not found in the mental language. For example, there is no grammatical gender. Nouns do not belong to different declinations. In order to see the criterion in real work, let us pay attention to a text where Ockham considering whether Latin grammatical distinctions must exist in mental language (Summa logicae, I, 3; trans. Loux, 1974, 53): Mental and spoken names, on the other hand, differ in that although all of the grammatical features of mental names belong to spoken names, the reverse is not true; whereas some grammatical features belong to both mental and spoken names, others are peculiar to spoken and written names (the grammatical features of these two kinds of names being always the same).

Ockham brings up case and number as examples of grammatical features that in his opinion have to have equivalents in mental language. Behind these examples, we most obviously find the intention to point out cases that are not self-evident, but for which Ockham can prove to find good grounds. He treats them as follows (Summa logicae, I, 3; trans. Loux, 1974, 53): Thus, just as the spoken propositions ‘Man is an animal’ and ‘Man is not the animals’ have distinct predicates, one of which is singular and the other, plural, so it is with the corresponding mental propositions which the mind asserts before any word is uttered: the predicate of the one is singular; that of the other, plural. Further, the spoken propositions ‘Man is a man’ and ‘Man is not man’s’ have predicates which differ in case; the same holds true of the corresponding propositions in the mind.

In Ockham’s examples, number and case have an effect on the truth-value of the sentence. When the term ‘animal’ is in the singular, we can predicate it affirmatively of the subject term ‘man’, but when it is in the plural, the proposition must be negative to be true. In the same way, if the predicate ‘man’ is put in the genitive case, it produces a false affirmation. These examples give a good picture of Ockham’s way of applying the principle of truth-value. He is cautious enough not to call the predicate term inflected in the mental language, or to give any other statement of how the mental language applies the distinctions of these grammatical categories. Ockham only notes that these distinctions can also be found in mental language. Ockham uses only a few pages in the beginning of Summa logicae to sketch out the grammar of the mental language. He clearly found it important to only draw the guidelines with which this kind of grammar could in principle be formulated. Most obviously he did not find carrying out this kind of project as urgent. Modern interpretations have, however, paid some attention to the extent to which the deep structure of Ockham’s mental language could be useful regarding the needs of deductive logic. This interest has, it seems to me, been based on approaching Ockham’s thought from the viewpoint of modern logical analysis. For Ockham, the

William Ockham and Mental Language

107

reduction of sentences to the ideal form of the mental language was not, however, the kind of project as the enthusiasm of the twentieth century to formalize propositions in order to see in that way their true logical form. Ockham did not find it necessary to think through what signs exactly belong to the structures of the mental language. Therefore, neither was characterizing the deep structure or the logical form any overarching aim for him. Ockham used this technique contextually in specific situations rather than as a universal philosophical method. Ockham made a firm distinction between categorematic and syncategorematic terms. This would have made possible a clear distinction between logical form and content too. According to Ockham’s characterization, categorematic terms supposit for (supponit pro; denote or refer to) something outside the sentence. They form a bond between the sentence and the things talked about. Syncategorematic terms, on the contrary, do not signify anything in this sense. In Ockham’s words, ‘none of these expressions has a definite and determinate signification, nor does any of them signify anything distinct from what is signified by categorematic terms.’ In Ockham’s opinion, the function of these words is to perform different operations in relation to the categorematical terms to which they are combined. Ockham continues: ‘it makes that categorematic expression signify something or supposit for something in a determinate manner, or it performs some other function with regard to the relevant categorematic term.’ (Summa logicae, I, c. 4; trans. Loux, 1974, 55.) After Ockham, John Buridan used the idea of mental language and especially the distinction of categorematic and syncategorematic terms in defining a formal validity in a deduction. According to Buridan’s definition, deduction is formally valid if the categorematic terms in it can be substituted with any other terms and the deduction is still valid. This substitutional way of defining the formal validity of deduction does not seem to have been Ockham’s choice. On the contrary, Ockham seems to have thought that deduction can also be formally valid in certain cases in which validity depends on the meanings of categorematic terms used in the sentence. From this perspective, it seems impossible to think that Ockham would have thought of logic in the now current sense of the word when spelling out the deep structure of an ambiguous sentence in terms of the mental sentence behind it. We should also note that Ockham mentions that prepositions are syncategorematic terms. Their role is not very useful for the needs of deductive logic. In fact, when treating sentential truth in general Ockham is forced to consider exceptions to account for the effects of prepositions and cases. He has to give up the universal principle according to which truth would mean in an affirmative sentence that the subject and the predicate refer to the same thing and in a negative sentence to different things. For example, the affirmative sentence ‘the donkey belongs to a man’ (asinus est hominis) specifically requires that the subject (‘donkey’) and the predicate (‘man’ in the genitive case) refer to different things – although this condition does not guarantee the truth of the sentence (see Summa logicae, II, c. 4). The genitive as a linguistic structure is in this sentence used to represent ownership – and as we shall soon see, such features of the external world may not be expressible with mere identity-predications. However, identity predications were in Ockham’s eyes

108

Representation and Objects of Thought in Medieval Philosophy

crucial for the functionality and generalizability of his main system of inference, Aristotelian predicate logic. Thus, admitting genitive case in the basic structure of the mental language causes problems in systems of deductive logic. In a word, Ockham’s syllogistic logic works only for certain kinds of mental expressions – not for all of them. To my understanding, this feature of Ockham’s theory shows that his theory of mental language was not intended to support logic (in the sense of the word which refers to deductive systems). Rather, its purpose was to provide a theory of language as a universal representational system. The Words of Mental Language and Their Signification So far, we have seen that Ockham’s discussion of the grammar of mental language shows that Ockham’s intentions were not the same as, for example, Bertrand Russell’s intentions in the analysis of the sentence ‘The king of France is bald’. Let us now look at what Ockham says about the actual glossary of this language, categorematic terms and their meanings. Ockham calls ideally behaving categorematic terms ‘absolute terms’. In principle, all terms meaning so-called natural kinds belong to this group. Ockham’s prime examples include ‘man’, ‘donkey’, ‘heat’, and ‘whiteness’. Each term of this kind means everything it means in the same way, unlike so-called connotative terms, which directly represent certain things while connoting something altogether different – I will soon come back to this group of terms. A closer look at Ockham’s explanation of the absolute terms shows that they behave in the mental language in a rather extensional way, i.e., almost as the logical analysis of the twentieth century would require. The function of an absolute term in its context is almost purely to refer to the things signified by it. Let us, therefore, accept such an extensional reading of the meaning relation as a kind of intuitive starting point when we turn to looking at the ways in which terms, in Ockham’s theory of mental language, signify the things that they make the subject think of. Ockham accepted, at least in some sense, two customary medieval ways of explaining why a certain ‘intention of the soul’ means exactly what it means. These two explanations were, on one hand, similarity, and on the other, causal connection. Ockham thinks that it is acceptable to say that a word of the mental language is in some relevant sense similar to the thing signified by it. He also accepts the principle of causality. Absolute terms of mental language derive their signification from their first appearance in the mind, and specifically its causal background. So, when I see a donkey, this perception can be described as the following kind of causal chain: There’s a donkey in front of me and my eyes are open and directed to the donkey. The donkey causes a certain occurrence in the mind, a mental act. This mental act is in some awkwardly specifiable sense similar to the donkey. Whatever this similarity boils down to, because it was originally a donkey that caused the mental act, I thereby think of a donkey. At the same time, I get an ability (habitus) to produce new similar mental acts. These repeated mental acts – the tokens of the term ‘donkey’ in mental

William Ockham and Mental Language

109

language – are similar to the original act caused by the donkey, and therefore mean donkeys, even when there are no donkeys present. In this story there are many gaps, if its intention is to tell what the semantic relation of signification really is about. Presumably, Ockham was very aware of this; probably he rather thought that no specific explanation is needed than that this story would have been a sufficient explanation. All in all, the semantic relation of signification is for Ockham essentially a natural phenomenon, something that is connected to every thought by its nature as a mental act. As far as Ockham’s mental language is concerned, his nominalism has often been interpreted as inconsistent with the idea that the terms of mental language have intensional content. It has been thought that it is impossible to say anything more about a term of Ockham’s mental language than a list of the actual and potential individuals which the term signifies or can signify. It seems to me that this is a wrong line of interpretation. Rather, Ockham thought that learning the vocabulary of the mental language is equivalent to learning the real (individual) essential forms of the things signified by the terms. It has to be remembered that he explicitly denied that the terms of mental language could mean whatever their user desires. Learning concepts is learning to classify the world in a manner that is not dependent on the classifying mind. If, for example, Socrates, Plato and a line are compared, no understanding can come to any other conclusion than putting Plato and Socrates under the same concept, in distinction from the line. Plato and Socrates are essentially similar to each other and different from the line. That is why they fit under a concept under which the line does not fit. (Ordinatio, I, q. 6; OTh, II, 211–12) That the terms of mental language have intensional content is perhaps most clearly visible in examining truths like the sentence ‘Every man is an animal’. Ockham thinks that this sentence is per se known to be true. We do not need to examine the complete list of human beings to show that every man is an animal, because anyone who masters these concepts, sees by mere examination of the terms that it is true (if we allow for the existential presupposition). That there cannot be humans who would not be animals can be seen by anyone who masters the terms. When Ockham distinguishes formally and materially valid deductions he even regards deduction leaning on this kind of truth as formally valid (for example ‘Socrates is a man; therefore, Socrates is an animal’). Ockham is known as a nominalist. Still, he seems to have believed that the conceptual structure of the mental language corresponds to the structure of reality, and in some sense reflects it. It seems that his view that the vocabulary of mental language as essentially learned, not innate, has to do exactly with this view of the relation between the structures of the world and of thinking. Ockham did not believe that human understanding could learn to know the world’s metaphysical structure – i.e. natures of the individuals themselves – in any direct way independent from perceiving the individuals. If platonic ideas existed, such an explanation would have been possible, but Ockham could not accept that. Ockham was a firm Aristotelian believing that the terms of the mental language show what the individuals are really like, and not only classify them arbitrarily to distinct groups.

110

Representation and Objects of Thought in Medieval Philosophy

In his ontological discussions, Ockham often remarks that there are no other real things than substances and qualities. This appears to deny a real status from certain essential features of the world, like relations and e.g. time. For Ockham, they are not things (res). On the other hand, it is clear that Ockham refuses even in these cases to admit that the understanding could construe for example temporal order or various relations between different things arbitrarily. We learn these things through experience, and so the assignment of truth-values to the corresponding sentences of mental language must be learned through experience. We cannot decide whether Rome was founded before Christ was born. Rather, we have to learn it. But now, how do we represent such structure in our thinking? What are the structures of mental language needed for such sentences? Do we have to learn them, or do we have them naturally? And if we have to learn what time is, e.g., how can that happen if it is not a real thing that we could encounter in any way? The simplest example of a non-existing but important metaphysical structure in the Ockhamist context is the relation of inherence between a certain quality and the substance to which it belongs. Let us consider a white cat. Ockham clearly takes the ontological commitment that we are here dealing with two real individual things, a certain cat and a certain whiteness. He rejects the view that there could be a third thing, a relation of inherence between these two. The fact that this whiteness is specifically the whiteness of this cat is a relation between two individual things, but Ockham very clearly comes to the conclusion that it does not exist in the world as a third individual thing. But if the relation of inherence obtaining between the substance and a quality does not exist, how do we explain that the two are bound together? Why should we think in our example that the whiteness qualifies precisely the cat? Ockham clearly believes that we are not free to make up our minds whatever way, but that the world determines whether the mental sentence ‘the cat has whiteness’ is true or not. In case of uncertainty about the color of the cat, we should go to the cat in good daylight and look. But the real mystery is that we still will not be able to see any inherence between the whiteness and the cat. We just see (propositionally) that the cat is white – or is not. Somehow, our representation of the situation is accurate only if it represents the whiteness as inhering in the cat; if the two things are represented as existing separately, the representation is inaccurate. And if we represent the two things as identical, the representation is even worse, since the cat is not whiteness, it is white. Thus, in our representational system we must be capable of representing inherence, even if there is no such thing in the world. How can we do that? Did Ockham think that we innately master the syncategorematic structure connecting the concept of a cat and the concept of a whiteness to a predication expressing inherence? And are empirical observations simply in some inexplicable way such that predications like this turn out to be sometimes true and sometimes false although there is nothing in the world that could correspond to a crucial part of the affirmative predication? Or, should we think that we learn by observing the world, in addition to the concepts of ‘cat’ and ‘whiteness’, a way to somehow connect these concepts

William Ockham and Mental Language

111

to a predication expressing inherence? Do we learn to use the claim ‘the cat has whiteness’ as an accurate representation of the world when ‘the cat does not have whiteness’ will not do? Ockham does not properly answer these kinds of questions. However, it seems already quite remarkable that he manages to distinguish these two approaches to our representational system. On the one hand, we are capable of representing individual things. But that does not suffice to give an accurate representation of the world. In addition, we must be able to represent something more. We ought not to blame Ockham if he did not manage to get a clear picture of what this something more is and how we learn to represent it. But we may want to turn to his one-time secretary Adam Wodeham, and the theory of complexa significabilia, which was to come a major topic of semantic theorizing in the fourteenth century. If looked at from the perspective we have been building now, it seems natural for Wodeham to take this step and to claim that in addition to the things represented by the categorematic terms, a proposition signifies something real, which is not a thing. Needless to say, the ontological status of such ‘states of affairs’ was obscure, and was questioned by the contemporaries. Nevertheless, late fourteenth-century medieval linguists challenging their status had to provide their alternative story of what is the content of a proposition over and above the real things signified by its terms. Let me connect this issue to another issue, that of connotative terms, which has caused a lot of confusion in the modern interpretations. Ockham calls ‘connotative’ those terms which have two distinct ways of meaning. For example, in the sentence ‘The cat is white’ the word ‘white’ (in Ockham’s Latin, albus) means on one hand the cat (the thing which is white), and on the other the whiteness (the color which the cat has). Furthermore, it means these two distinct things in a quite different way, since in the sentence it refers to the cat (the subject where the whiteness inheres), but not to the whiteness itself. If we wanted to refer to the color without speaking of the cat we would have to use the word ‘whiteness’ (in Ockham’s Latin, albedo). So, the word ‘white’ is a connotative term, and the word ‘whiteness’ correspondingly an absolute term. When we look at the term ‘white’ it appears relatively naturally that an ideal mental language would not need this term as an irreducible part of the language. Instead of saying, ‘the cat is white’ we could and perhaps should say, ‘whiteness inheres in the cat’. It seems that the mental language could manage with the concept of whiteness and the syncategorematic structures with which it is possible to express that whiteness inheres in some substance. Such reduction would be rather useful for the purposes of an ideal language, because then only one type of semantic relation between the language and the world would be needed. If the mental language contains irreducible connotative terms, it has to lean on a multitude of types of semantic relations – even though the ideal language should be simple. Modern scholars have studied the reduction of connotative terms mainly appealing to a principle given by Ockham, which says that all connotative terms must have a nominal definition explaining their meaning. They cannot have a real definition which would tell what the things denoted by the term essentially are. In

112

Representation and Objects of Thought in Medieval Philosophy

sentential contexts, connotative terms do not usually denote things that are relevant for its meaning. (For example for the word ‘busy’ it is useless to define the essence of man, although ‘busy’ typically refers to men.) Rather, connotative terms need a nominal definition that tells what the word means or how it is used. However, in the modern interpretive debate it has turned out that, at least for the connotative terms expressing relations, it is from Ockham’s perspective impossible to give such nominal definitions which could be substituted for the terms in different contexts. The modern interpretive debate has drawn from this (and some other problems) the conclusion that Ockham in fact accepted irreducible connotative terms as a part of mental language. In my opinion, the modern debate has not sufficiently paid attention to the fact that then the relations between linguistic units and the real things become considerably more complicated. It is questionable whether Ockham’s theory of mental language can be considered as a theory of ideal language at all, if we accept different connotative semantic relations as irreducible primitive structures in it. The modern debate has not sufficiently studied the role, which the syncategorematic terms have in Ockham’s mental language, either. I think it is clear that connotative terms cannot be substituted by their nominal definitions – absolute terms with some syncategorematic structures – if the analysis has to be carried out step by step so that we get simple affirmative or negative predications as a result. As it seems to me, this restriction is too hard on Ockham’s theory and underestimates the complexity of mental language as a representational system. According to Ockham’s theory, the syncategorematic and grammatical structures of the mental language are very numerous and complex; they cannot even be compared to what modern logic calls logical constants. It is important to remember that Ockham built his theory of mental language using the spoken language as the starting point, instead of formulating a logical construction. Nevertheless, I think that even in analyzing a complex relative term (like ‘father’) is useful to take as an Ockhamist starting point the heuristic division between direct and simple representations of things and representational structures built from such simple representations. As I already noted, Ockham never completed his research program on the structure of the mental language, and we have no clear answer from him to the question of irreducible connotative terms. But as it seems to me, he thought that allowing one type of relation of signification should suffice for explaining all linguistic meaning – if but only if syncategorematic and grammatical structures can be fully utilized in the explications. Recent research has shown that Ockham’s connotative terms cannot be substituted with analyzed terms containing only absolute and syncategorematic terms, no matter how complex they are allowed to be – and Ockham probably knew that. However, to my knowledge no one has tried out what happens if we tried analysis in terms of non-asserted propositional content. And after all, this would be the most natural line to take given that apparently all interesting examples of nominal definitions of connotative terms include nonasserted propositional content. Thus, ‘father’ is defined as someone having a son, and ‘white’ as something having whiteness. ‘White’ does not only signify primarily

William Ockham and Mental Language

113

some subject and secondarily a specific quality, but also somehow includes in its meaning the standard copula of an inherence-predication. As it seems to me, Ockham suggested that such propositionality is somehow included in the connotative terms, but the topic remains largely unstudied. Couldn’t we think that the connotative terms like ‘white’ are built from absolute and syncategorematic terms in grammatical constructs not only representing things in the world but also having representational structure of a propositional type? Conclusion Ockham’s theory of mental language was a theory of ideal language. However, during his time in the academia he did not complete a systematic research program on the topic, and could thus not offer clear answers to many natural questions about the properties of this language. Some basic principles are anyway clear. The mental language is common to everyone, and it rises from what the real structure of the world is like. Learning to think means learning the universal language. Thus, it seems that despite his nominalism, Ockham rather profoundly accepts the Aristotelian basic conviction that our understanding can meet the metaphysical basic structure of the world – that the world is intelligibile. In his opinion, the language in which we think is universal because in the final analysis both its glossary and its grammar are defined by the world itself, not by our own minds. In fact, the representatives of the seventeenth-century Cartesian linguistics as described by Chomsky admitted a similar basic conviction. Already the idea that some universal language is common to all thinking things implies the condition that no private person or certain culture defines this structure. Ockham, however, emphasizes that the mental language must be learned. We learn it by observing the world and finding out what kinds of things there are and what is true about them. Ockham does not think that the glossary of the mental language, or the relations between terms could rise from the mind’s own nature. For him, the universal mental language is much more clearly a representation of the basic metaphysical structure of the world than for the seventeenth-century theorists who think that it springs out from the structure of the mind itself. From the perspective of seventeenth-century thinkers, it becomes a natural epistemological problem to ask why the mental language represents the world accurately. Because Ockham held that the structure of the mental language springs from the structure of the world, he had no need to answer to such epistemological challenges. I have tried to argue here that Ockham thought that in order to gain an adequate representation of the world, it is not sufficient that everything in the world has its own representative in the mind. These representations must form a system. To represent everything true or false that can be said or believed about the world, one needs to be able to construct propositions. And Ockham seems to have been aware that propositions and even many terms mean more than just the things referred to. But what this something could be, remains obscure.

114

Representation and Objects of Thought in Medieval Philosophy

Bibliography Alféri, P. (1989), Guillaume d’Ockham: Le Singulier, Paris: Les Éditions de Minuit. Biard, J. (1989), Logique et Théorie du Signe au XIVe Siècle, (Études de Philosophie Médiévale, LXIV), Paris: Vrin. Brinkley, Richard, Theory of Sentential Reference, (ed. and trans.) M.J. Fitzgerald, Leiden: Brill 1987. Chalmers, D. (1999), ‘Is There Synonymy in Ockham’s Mental Language?’, in Paul Vincent Spade (ed.), The Cambridge Companion to Ockham, Cambridge: Cambridge University Press, 76–99. Chomsky, N. (1966), Cartesian Linguistics: A Chapter in the History of Rationalist Thought, New York: Harper & Row. Freddoso, A.J. and Schuurman, H. (1980), Ockham’s Theory of Propositions: Part II of the Summa logicae, Notre Dame: University of Notre Dame Press. Goddu, A. (1993), ‘Connotative Concepts and Mathematics in Ockham’s Natural Philosophy’, Vivarium, XXXI, 1, 106–39. Karger, E. (1994), ‘Theories de la pensée, de ses objects et de son discours chez Guillaume d’Occam’, Dialogue, XXXIII, 437–56. King, P. (1985), Jean Buridan’s Logic: The Treatise on Supposition and The Treatise on Consequences, Dordrecht: Reidel. Loux, M.J. (1974), Ockham’s Theory of Terms: Part I of the Summa logicae, Notre Dame: University of Notre Dame Press. Normore, C. (1990), ‘Ockham on Mental Language’, in J-C. Smith (ed.), Historical Foundations of Cognitive Science, (Philosophical Studies Series 46), Dordrecht: Kluwer. Ockham, William, Summa logicae, (Opera philosophica et theologica, Opera philosophica, I), Institutum Franciscanum, St Bonaventure, NY 1974. ——, Quodlibeta septem, (Opera philosophica et theologica, Opera theologica, IX), Institutum Franciscanum, St Bonaventure, NY 1980. ——, Quodlibetal Questions, vols 1–2, trans. Alfred J. Freddoso and Francis E. Kelley, (Yale Library of Medieval Philosophy), New Haven & London: Yale University Press 1991. Panaccio, C. (1990), ‘Connotative Terms in Ockham’s Mental Language’, Cahiers d’épistémologie, No. 9016, Montréal: Université du Québec. —— (1991), Les Mots, les Concepts et les Choses: La sémantique de Guillaume d’Occam et le nominalisme d’aujourd’hui, (Collection Analytiques, 3), Paris: Vrin. —— (1992), ‘Intuition, abstraction et langage mental dans la théorie occamiste de la connaissance’, Revue de Métaphysique et de Morale 97, 61–81. —— (1999), ‘Semantics and Mental Language’, in Paul Vincent Spade (ed.), The Cambridge Companion to Ockham, Cambridge: Cambridge University Press, 53–75.

William Ockham and Mental Language

115

Schaeffer, F. (1987), ‘Syntax and Semantics in Supposition Theory’, in Ockham and Ockhamism, Nijmegen: Ingenium, 63–9. Spade, P. (1975), ‘Ockham’s Distinction between Absolute and Connotative Terms’, Vivarium, XIII, 55–76. —— (1988), ‘The Logic of the Categorical: The Medieval Theory of Descent and Ascent’, in N. Kretzmann (ed.), Meaning and Inference in Medieval Philosophy, Dordrecht: Kluwer, 187–224. Trentman, J. (1970), ‘Ockham on Mental’, Mind, 79, 586–90. Yrjönsuuri, M. (1997), ‘Supposition and Truth in Ockham’s Mental Language’, Topoi, 16, 15–25.

This page intentionally left blank

Chapter VII

The Matter of Thought Calvin G. Normore

The Mind is like a mirror, it collects dust. (attributed to the Fourth Patriarch of Chinese Buddhism)

A doctrine very widely, though not universally, held by philosophers within the Aristotelian tradition is that the intellect is immaterial. Indeed it used be said that John Locke had been the first philosopher since antiquity to broach the possibility that matter might think. This is of course far from the truth. The view that the intellect was material was associated in both the Arabic and Latin Middle Ages with Alexander of Aphrodisias and it was widely and seriously discussed. In many minds the doctrine of the immateriality of the intellect was closely associated with that of the immortality of the intellect and even of the immortality of the soul, and, perhaps, since those who rejected the immortality of intellect (whether it be an individual intellect or an intellect for the entire human species) were few, they were also few who asserted that the intellect was material. Yet the possibility intrigued thinkers throughout the medieval period. Matter is introduced by Aristotle, in Physics I to account for change. It is the ‘subject’ from which something comes to be. Understood in this way there is no very tight conceptual connection between being a material object – that is one composed of matter – and a physical object in any of the senses in which that phrase is understood nowadays. Some medieval writers, most famously Ibn Gabirol, held that purely spiritual beings like angels were material and almost everyone held that objects occupied space and had the other properties we now associate with physical objects by virtue of their forms. The view that matter thinks has become a kind of orthodoxy. It is very widely held that to suppose otherwise is to be unnatural or at least unnaturalist, to monger mysteries, to reject evolution and to despair of the possibility of a unified science. In these discussions ‘material’ and ‘physical’ are usually used interchangeably – and rather loosely. Noam Chomsky has commented in response to queries about whether his psycholinguistics entailed dualism that when we came to understand something sufficiently well we would call it physical and the history of science provides some support for his view. The forces and even the particles of contemporary microphysics would likely not have been regarded as matter by the architects of the mechanical philosophy.

118

Representation and Objects of Thought in Medieval Philosophy

But leaving aside worries about what it is for something to be material or physical one might wonder, as Chomsky seems to wonder, why there has been so much fuss about whether thought is a physical process or about whether physical or material objects can think. It is not an automatic consequence of a worry about immortality. After all without any thought as to whether it was physical Cebes and Simmias had already taxed Socrates with the worry that the soul might be like a cloak which could outlive many a body but would finally perish, and there is no conceptual reason why a physical system might not last forever. If one turns from worries to arguments one notices, I think, that the arguments about the immateriality of thought and thinkers have for a long time rested on a perceived mismatch between the scope of thought and the powers of physical or material objects. There is still debate about whether the human capacity to do mathematics shows that no human mind could be modeled as an axiomatic system.1 Arguments that the powers of the mind outstrip those of any possible physical system are not new. Indeed they are the central arguments employed in debates about the immateriality of the intellect within the Aristotelian tradition. My concern here is to relate them to another debate – that about the ontological status of objects of thought. My focus will be on thinkers in the Middle Ages loosely conceived, but I hope there to be in the discussion morals for our time.

Among twentieth-century philosophers of mind there were broadly speaking, two kinds of theories of what it is to think about something. One group, the relational theories, held that to think about something is for a subject to bear a relation to an object. The other group, the adverbial or qualitative theories, held that to think about something is to be in a mental state which is identifiable independently of any external object.2 Within the English-speaking world there was a time when 1 The central argument for the views seems to be that since Gödel has shown that for any axiomatic system strong enough to include the axioms of arithmetic there are mathematical truths which could not be proved within it, and that since there are no mathematical truths which a human being could not in principle prove, given time and thought enough, then a human mind could not be isomorphic to any axiomatic system – and so not to any Turing machine or computer either. To this argument it was quickly replied by David Lewis and others, that we had no good reason to accept the second premise – the premise that there were no mathematical truths a human, given time and thought enough, could not prove. Indeed it was suggested, there was evidence to the contrary. A more recent argument along different but related lines has been advanced by James Ross (Ross, 1992). Ross begins from Kripke’s work on Wittgenstein’s discussions of rule-following. Ross argues that Kripke has shown that the ability to follow a rule is more fine-grained than any physical disposition could be. Hence that ability cannot be or supervene on any such physical disposition. Therefore the mind has non–physical properties at least. 2 Those who think (with Jerry Fodor) that psychology is concerned with ‘narrow content’ I would number with the qualitative theorist. Those who focus on ‘wide content’ fall into the relational side of my dichotomy.

The Matter of Thought

119

Roderick Chisholm stood as the champion of the relational view and Wilfrid Sellars as champion of the adverbial or qualitative view. In an odd way these views are complementary. The relational view has no trouble explaining why a thought is of one thing rather than another. In the simplest form of relational theory the explanation is that the thinker bears the relation ‘thinking of’ to the one thing and not the other. The price such a theory pays for this simplicity is having to choose between claiming that we can only think about what exists and admitting that when we think about objects which don’t exist we are related to objects which don’t exist. The qualitative view need have no truck with such shadowy entities, holding rather that to think about this chair is just to be in the thinking-about-this-chair mental state. The price paid for this ontological economy is that if one asks what it is which distinguishes the thinking-about-this-chair state from the thinking-about-this-table state, satisfactory replies are hard to find.3 Time has passed since the days of Chisholm and Sellars and one might understand some recent externalist theories as efforts to fuse their two pictures. On some such externalist pictures for example, to think about Vienna is to be in a certain physical state – say a certain brain state – which bears a certain relation to – say is caused by – Vienna. On this picture the brain state is an ordinary physical state and the relation between it and the object of thought is an ordinary physical relation. What, though, is the thought of Vienna? Two answers to this question suggest themselves. One, which I think of as a reductionist answer, is that the thought just is the brain state though the brain state is only a thought of an object when it bears the appropriate relation to that object, just as Paul Martin is essentially a man but only Prime Minister of Canada when he bears the appropriately complex relation to the Queen. On this picture the thought is the brain state in appropriate circumstances. There is no new ontology. The second answer, perhaps encoded in Putnam’s remarks that thoughts are not inside the head, is that the thought of Vienna is a complex object consisting of the brain state, Vienna, and, perhaps, some relations as well. This picture does seem to introduce new items into the ontology and it is unclear to me whether we should count those as physical. Suppose, for example, that my thought of Vienna is the ordered triple . Is that triple a physical object? Both the relational and the adverbial views are medieval and so, I think, is their more recent hybrid. Indeed it seems likely that all have a more or less continuous history from the Middle Ages to our century, which history accounts for their being the options which still confront us. Of course the scientific context in which they were discussed is much changed. Throughout much of the Middle Ages astronomy was the most developed science of material objects, but it was not considered to be a proper part of Physics. Of the properly Physical sciences the most developed was Optics and the most powerful metaphor for the mind was not the windmill or the computer but the mirror. Despite the popularity of this image viewing the mind as a 3

Perhaps the most sustained recent effort to work out this view is Brandom (1994).

120

Representation and Objects of Thought in Medieval Philosophy

mirror was by no means the only available medieval option. In particular there was the contrasting tradition of viewing thinking as a kind of activity whose elements were not representations in any sense but rather acts. Here I wish to explore some of the medieval aspects of each of these views.4 In the first set of Objections to Descartes’s Meditations Caterus writes: My question is this: what sort of cause does an idea need? Indeed, what is an idea? It is the thing that is thought of, in so far at it has objective being in the intellect. But what is ‘objective being in the intellect’? According to what I was taught, this is simply the determination of an act of the intellect by means of an object. And this is merely an extraneous label which adds nothing to the thing itself. Just as ‘being seen’ is nothing other than an act of vision attributable to myself, so ‘being thought of’, or having objective being in the intellect, is simply a thought of the mind which stops and terminates in the mind. And this can occur without any movement or change in the thing itself, and indeed without the thing in question existing at all. So why should I look for a cause of something which is not actual, and which is simply an empty label, a non-entity? ‘Nevertheless’, says our ingenious author, ‘in order for a given idea to contain such and such objective reality it must surely derive it from some cause.’ On the contrary, this requires no cause; for objective reality is a pure label, not anything actual. A cause imparts some real and actual influence; but what does not actually exist cannot take on anything, and so does not receive or require any actual causal influence. Hence, though I have ideas, there is no cause for these ideas, let alone some cause which is greater than I am, or which is infinite. 5

Caterus’s puzzlement was not idiosyncratic. Arnauld later raises the same objection. Descartes’s reply is to insist that Caterus (and Arnauld) are wrong – the idea of the sun is not of the sun by an extrinsic denomination but just is the sun itself existing objectively in the intellect: But if the question is about what the idea of the sun is, and we answer that it is the thing which is thought of, in so far as it has objective being in the intellect, no one will take this to be the sun itself with this extraneous label applied to it. ‘Objective being in the intellect’ will not here mean ‘the determination of an act of the intellect by means of an object’, but will signify the object’s being in the intellect in the way in which its objects are normally there. By this I mean that the idea of the sun is the sun itself existing in the intellect – not of course formally existing, as it does in the heavens, but objectively existing, i.e. in the way in which objects normally are in the intellect. Now this mode of being is of course much less perfect than that possessed by things which exist outside the intellect; but, as I did explain, it is not therefore simply nothing. 6

4 This section is heavily indebted to the work of David Lindberg and his ‘school’ and particularly to Katherine Tachau’s pioneering study; Tachau (1988). 5 AT, VII, 92–3; CSM, II, 66–7. 6 AT, VII, 102–3; CSM, II, 75.

The Matter of Thought

121

I have argued elsewhere that Descartes is quite traditional in thinking that the sun existing objectively in the intellect requires a cause.7 The tradition he follows here is not one to which Caterus and Arnauld subscribe, but it is a tradition nonetheless. It is, I think, one wing of the Scotist tradition. If this is correct then what is novel about Descartes’s view of objective reality is not that it requires a cause but that the very same item, the idea, requires two causes – one of its formal reality and one of its objective reality. Why such causal profligacy? The answer is I think quite complex but one important strand, at least, lies precisely in the history of argument about the immateriality of the intellect. In De Anima, III, 4 (429a19–21) Aristotle claims that: since everything is an object of thought, mind, in order, as Anaxagoras says, to dominate, that is, to know, must be pure from all admixture, for the co-presence of what is alien to its nature is a hindrance and a block: it follows that it can have no nature of its own, other than that of having a certain capacity.

From the earliest extant commentaries this passage has been understood as an argument against the materiality of the intellect – but it is on its face a very obscure argument. The discussion of intellect in De Anima is closely connected with discussions of perception and conclusions about intellect are often based, at least in part, on analogies with perception. To understand how and why the mind must be ‘pure from all admixture’ the commentary tradition turned to two particular analogies. The first is with sight. At 418b27 Aristotle claims that ‘what is capable of taking on color is what in itself is colorless’. The second is with taste. At 422b1ff. He claims that ‘since what can be tasted is liquid, the organ for its perception cannot be either actually liquid or incapable of becoming liquid. Tasting is being affected by what can be tasted as such; hence the organ of taste must be liquefied, and so to start with must be non-liquid but capable of liquefaction without loss of its distinctive nature. This is confirmed by the fact that the tongue cannot taste either when it is too dry or when it is too moist; in the latter case there is contact with the pre-existent moisture, as when after a foretaste of some strong flavor we try to taste another flavor; it is in this way that sick persons find everything they taste bitter, viz. Because, when they taste, their tongues are overflowing with bitter moisture’. These analogies suggest that for sensation to be possible the sense organ must lack what is sensed and by analogy that for intellection to be possible the intellect must lack whatever is understood. The argument would then proceed from the premise that everything can be understood (is an object of thought) to the conclusion that the intellect cannot of its nature be anything. So far so good; but suppose for a moment that the intellect were actually something, what exactly would then fail to be an object of thought, that is what exactly would we be unable to understand? Two hypotheses suggest themselves (a) that what could not be understood is what the mind itself is and (b) that what could not be understood 7

See Normore (1989).

122

Representation and Objects of Thought in Medieval Philosophy

is anything which cannot be co-present with what the mind is. If we suppose both a and b then the analogies with perception would lead us to conclude that Aristotle or his commentators thought that were the eye naturally colored we would not see at all and were the tongue naturally liquefied we would not taste at all. This doesn’t seem plausible as an interpretation either of Aristotle or of the tradition so it seems that a and b are genuine alternatives between which a philosopher committed to the argument must choose. Something hangs on the choice. The model of perception by analogy with which the intellect is understood within the Aristotelian tradition has it that perception involves the relevant sense organ taking on, in some way, the form of the proper object of the sense. As we shall see, exactly how the form is received is a matter of some controversy but the use of perception in the argument that the intellect is fundamentally a capacity requires that, however it is understood, the reception be such that it makes sense to suppose that the presence of a form in the sense organ in the usual way (the eye being colored green for example) would interfere with the reception of the form in thought. If our understanding of the argument involves (a) then we would expect it to rely on an appeal to the principle that there cannot be two instances of specifically the same form in the same proper subject. On the other hand if our understanding involves (b) we would expect instead appeal to some principle which explains when forms of different kinds cannot coexist in the same subject. It seems that the early tradition, at least was divided. For example Alexander of Aphrodisias seems to have understood the argument in the first way. In his De Intellectu he writes: For what is destined to grasp (antiambanesthai) everything should not actually be any one of them in its own nature. For the intrusion (paremphainomenon) of its form into the grasping of things lying outside would impede thoughts about them. For neither do the sense grasp those things in which their being consists. Hence for this reason the organ through which there is the sight which is capable of grasping colors and through which there is the grasping is colorless. (For water is colorless in its own color). Furthermore the sense of smell which is capable of grasping odors is from air which is odorless. Moreover touch does not perceive things which are hot or cold or rough or smooth to the same degree as it itself is but rather things which differ from it to greater or lesser degree. In summary then, just as in the case of the senses it is impossible for a sense which possess to be capable of grasping and discerning that which it possesses, so also since intellect is an act of grasping and discerning objects of thought it is in no way possible for it to be among the objects discerned by it.8

Themistius, on the other hand, seems to have understood it in the second way: Anaxagoras was not, then, mistakenly dreaming in making the soul unmixed and of a nature different from everything that it becomes acquainted with. For in that way it would very easily become acquainted with [all things] since there would be nothing of its own 8 See De Intellectu 106.27–107.7 (translation after Schroeder in Schroeder and Todd, 1990).

The Matter of Thought

123

to intrude, i.e. coexist with it. The form inherent to it, that is, would exclude and obstruct the others as though they belonged to other things. An intellect of this sort must, therefore, have no nature or form of its own except its capacity to comprehend the natures and forms that belong to other things.9

So far as I can tell this division about how to read the argument persists at least until Aquinas – who for the most part sides with Themistius. For example, in his argument for the immortality of the soul in Summa theologiae, I, q. 75, he takes a Themistian stance. Aquinas there begins with an investigation of what a soul is. He concludes that it is the ‘first principle’ of life in a body. (Thus angels should not be thought of as souls since they have no bodies). He argues that the first principle of life in a body cannot itself be a body (since there are lots of non-living bodies it cannot be as a body that something is alive) but must be the activation of a body – the ‘act’ of a body. He then turns to the question whether the act of a human body is ‘something subsistent’ that is something which we can properly call a thing. He considers the objection that it is not because it is the complete human being and not merely its soul which is properly a thing and the objection that it is not the soul which acts but the whole human being. In his resolution of the question he writes: I answer that, It must necessarily be allowed that the principle of intellectual operation which we call the soul, is a principle both incorporeal and subsistent. For it is clear that by means of the intellect man can have knowledge of all corporeal things. Now whatever knows certain things cannot have any of them in its own nature; because that which is in it naturally would impede the knowledge of anything else. Thus we observe that a sick man’s tongue being vitiated by a feverish and bitter humor, is insensible to anything sweet, and everything seems bitter to it. Therefore, if the intellectual principle contained the nature of a body it would be unable to know all bodies. Now every body has its own determinate nature. Therefore it is impossible for the intellectual principle to be a body. It is likewise impossible for it to understand by means of a bodily organ; since the determinate nature of that organ would impede knowledge of all bodies; as when a certain determinate color is not only in the pupil of the eye, but also in a glass vase, the liquid in the vase seems to be of that same color.10

Suppose one sided with Themistius and Thomas and concluded that the intellect could not, for example, be both (entirely) made of bronze and grasp the form of water in thought. Why not? If the form of water would have to be present in such an intellect in the same way that the form of bronze is present then one could argue that a brazen intellect could not think about water because a body cannot be both (entirely) made of bronze and made of water and the intellect would be if it were informed by both the from of bronze and that of water. The trouble with this argument is that it suggests that the intellect becomes literally whatever it thinks and it seems silly to suppose that 9 See Themistius, Paraphrase of De Anima, 3.4–8. Translation by R.B. Todd (slightly modified from Schroeder and Todd 1990). 10 See Thomas Aquinas, Summa theologiae, I, q. 75.

124

Representation and Objects of Thought in Medieval Philosophy

anyone who thinks of an elephant literally becomes one. We may be well-advised to guard our thoughts but not for that reason! One could reply that while it is of course true that someone who thinks of an elephant does not become one this is only because what is special about the intellect is that unlike everything else it when it takes on the form of X it does not become X. If the intellect were made of bronze, this reply might continue, then it would not be the special thing it is and so could not take on the form of water without becoming water. This line of thought has its own charms but in it the analogy with perception which seemed to the commentary tradition to be basic to Aristotle’s argument plays no role. In the analogy with perception we reason from the premise that a sense organ can sense every form in a certain range to the conclusion that it does not actually have in its constitution or nature any form in that range. We do not conclude that it is a peculiar kind of stuff which could not have in its nature any form in that range. Indeed as the tongue of the sick person illustrates the sense organ typically could have a form in that range and were it to would malfunction. Within the Aristotelian tradition perception is itself understood in part by analogy with the way in which a form is present in an ordinary object. The sense organ becomes what is sensed and this is understood by analogy with matter being informed in more ordinary cases – a copper statue becoming green, or a lump of copper becoming a statue. Against this background it may not be totally implausible in the case of the proper object of vision, color, to suppose that the sense organ perceives by literally becoming colored in the same way the bronze statue becomes green but it is very hard to make this out for all the senses: a nose smelling rancid meat does not seem to itself begin to smell rancid. Moreover in cases in which there is a medium between the sense organ and what is sensed it would seem that the form which informs the sense organ is somehow transmitted through the medium. A little reflection convinced the commentary tradition that if this too was to be modeled as the form successively informing different parts of the medium then there were grounds for insisting that a form could be present in a way other than the way in which green is present in the weathered copper statue. All of this suggests that in any analysis of thought which preserves the point of the analogy with perception we will have to suppose that the form being thought of is present in the intellect in a way different from the way it is present in the thing(s) whose form it normally is. This leaves us with a latent problem which surfaces from time to time during the Middle Ages as we shall see below: if in thought a form is present in the intellect in some special way (the tradition drew the analogy to the way in which the color of a reflected object is present in a mirror) then it is less than clear why a form couldn’t be present in one way as the form of the intellect and in another way as the form in the intellect. During late antiquity and the Middle Ages reflection on the ways in which forms could be present to sense organs and intellects became complexly intertwined with reflection on the ways in which forms could be present in the media between object and perceivers. This is not surprising. Aristotle had maintained that perception involved the passage of a form from the object perceived to the perceiver and its

The Matter of Thought

125

presence in some way in the medium between them. Just as in the cases of sense organs and intellects it was natural to ask whether forms were present in the medium in the way they were present in the object. One powerful reason to think they were not was provided by the observation that the paths between different perceivers and the objects they perceive intersect. If you and I see each other there must be forms passing from you to me and from me to you in the same space between us. Some of these forms will, for example be color forms and, if we are wearing different colors will be different color forms. These seem to be co-present in the medium without interfering with each other. The color I see you as wearing does not (normally) vary with the color I am wearing myself. It is perhaps worth pointing out that the medieval conception of colors in the medium really is a conception of something not easily assimilated to either a crude materialism or a crude version of the view that colors are mind-dependent. Against the materialist interpretation is the fact that in the same air there can be contrary colors. Two people looking at each other may perceive very different colors. If these colors are ‘in’ the space between them at the same time then distinct and even ‘contrary’ objects are in the same place at the same time. Thus it seems that colors cannot be present in a medium in the way they are in colored objects and if they are corporeal forms when in the medium we run afoul of the very powerful intuition that no two distinct corporeal bodies or forms of the same kind can be in the same place at the same time. On the other hand optics requires that the colors in the medium be in particular places, pass through (or be ‘multiplied’ over) particular rectilinear paths etc. This is not the behavior of a mind-dependent object in the modern sense. It is thus not silly to look for a third category sharing some of the features of the corporeal and some of the mind-dependent. 11 The concept of esse spirituale was introduced to label precisely this intermediate status. It offered offered medieval thinkers the prospect of a unified theory of optics and psychology and, in particular of a unified science of Alhazen’s optics and Avicenna’s psychology which had arrived in the Latin West more or less together. Such a synthesis was aided by the fact that both Alhazen and Avicenna use the Arabic ‘ma’na’ (pl. ma’ani) for key concepts in their theories. Alhazen uses it for an object of perception – that is, on one of the properties of objects which can be perceived. On his account each of these is a form (sura) which is present in the object and which is perceived by means of a form (sura) which is transmitted through the medium to the sense organ. It is this form which is reflected from a mirror. Avicenna seems to reserve ‘ma’na’ by itself not for the objects of perception for which Alhazen used it but precisely for properties which could not be sensed – like the hostility of the wolf which is in some sense percived though not sensed by the sheep. ‘Intentio’ is the Latin word translaters use to translate ma’na. The translators 11 Notice, by the way, that our own (or the nineteenth century) story about forces has many of the same embarrassments. Forces too can occupy the same place at the same time but forces too do not seem mind-dependent. It is perhaps not without reason that Descartes compares the action of the soul to gravitational attraction.

126

Representation and Objects of Thought in Medieval Philosophy

also used ‘intentio’ to translate ‘ma’qul’ (formed from ‘aqala’, the verb to think and, as Kwame Gyekye has shown, ‘intentio’ is also used to translate various forms made from the verb ‘qasada’ (like the Greek ‘egomai’), many of which would be better translated simply as ‘principally’ or ‘primarily’. Thus ‘intentio’ as it appears in Latin texts is deeply ambiguous. The Latin translation of Avicenna distinguishes first and second intentions seeming to mean by first intentions something like characteristics of objects, and by second intentions something like characteristics of first intentions, but it also has Avicenna speaking of intentions as items imparted to perceivers by objects along with their forms. The Latin text of Avicenna also insists that intentions have less being than the objects which intend them and that second intentions have less being than first intentions. It does not say what kind of being this is. The result of this melange was a tradition which assigned intentions a role in logic (which Avicenna had described as the science of second intentions, a role in the theory of sensation as the items which passed from the sensed to the sensor, and a role in the transmission of insensible but in some way perceptible properties. The upshot is that it would seem to a studious reader of the Latin Avicenna that there were items which could be in the same way in the medium and the soul. As the research group round David Lindberg (especially Katherine Tachau) has made clear Bacon tried heroically to devise a theory in which the same item could fill all these roles. He proposed that objects had the power to generate (multiply) species which passed through the media between perceiver and perceived according to the principles of Alhazen’s optics. These sensible species are also intentions and likenesses of things. They are also natural signs of the things which give rise to them. Words are imposed to signify conventionally the same things which these species signify naturally. With the concept of intentio as lynchpin Bacon thus proposed a single account which would explain both the physics and the psychology of perception. Bacon’s view of representations is part of his metaphysics of light. He claims that light, for example the light of the sun, acts on the objects around it by means of a power which we can call ‘similitudo’, ‘imago’ or ‘species’.12 This power operates through the multipication (or more properly successive generation) of these likeness or species. Each species is or is made up of intentions. These species behaved according to the laws of optics and the also informed the medium through which they passed – but in such a way that species of qualities in the same range, species of red and green for example, could be copresent in the same place at the same time. Bacon himself insisted on the corporeality of the visible species. He did so because he thought that only something with corporeal being (esse corporale) could be similar to a corporeal object, and only something corporeal could be expected to follow the laws of optics.13 In so insisting he took a stand not only against such thinkers as Albertus Magnus who cites favorably the view that colour in the medium 12 See Roger Bacon, Opus Maius, pt. 4,d. 2,c. 1 [I:111] cf. Lindberg, 1976, 113. 13 ‘Et hoc iterum patet quoniam species est similitudo rei corporalis et non spiritualis, ergo habebit esse corporale.’ (Roger Bacon, Opus Majus, pt. 5; pt. 1, d. 6 c. 4.).

The Matter of Thought

127

has spiritual being but also against the translators of Averroes who had translated the Averroes as claiming that while colors have corporeal being in colored objects they have spiritual being in the diaphomous medium. Against such authority Bacon’s insistence on the corporeality of visible species had little effect. The sort of being which colors have in a medium continued to be called esse spirituale throughout the medieval period. Esse spirituale was contrasted with various other notions – esse naturale, esse formale, esse corporale were all employed as contrasts at different times and in different contexts. The terminology was a recipe for trouble. That spiritual beings cannot as such be sensed was, I think, regarded as a conceptual truth in the later Middle Ages. This did not, however, entail that spiritual beings were immaterial. Many believed that corporeality was necessary for being sensed and that corporeality was a form which inhered in substances while matter was a principle of potentiality in things. On this view God, being all that He can Be, is the only immaterial being, but the angels, for example, are incorporeal beings. Exactly how the incorporeal, the immaterial and the spiritual fit together has never been as clear as one might wish and it was not in the High Middle Ages either. To get a sense of how esse spirituale could function as ‘in between’ natural being and something purely mind-dependent consider the picture as we find it in Nicholas Oresme’s Question 21 of Book III of the Questiones super Meteororum which concerns ‘Whether the rainbow may be a real form (forma realis) impressed on a cloud or may be only and imaginary form (forma ymaginaria)?’ After posing a number of problems for the view that the rainbow is a real form Oresme begins his response to the question by distinguishing first between ‘real’ as contrasted with mind-dependent or ‘imaginary’ forms and second within real forms in this sense between those which cannot co-exist with contrary forms and those which can. It is these latter non-imaginary forms which can coexist with others in the same subject and be unaffected by them which he calls ‘spiritual’ the paradigm of spiritual forms for Oresme are intelligible species and they are contrasted on the one hand with such forms as whiteness or sweetness or heat which Oresme thinks to be unqualifiedly real and on the other with phenomena like the fiery circle there appears to be when one swings a burning stick in a circle of the colors that appear from a distance in a dove’s neck. Having drawn his three-fold distinction of forms into ‘real’, ‘spiritual’ and ‘imaginary’, Oresme proceeds to employ it. He argues that the rainbow is a real form in the first sense in which it is opposed to ‘imaginary form’ but not in the second sense in which it is opposed to something ‘spiritual’. Second conclusion: that a rainbow is not a real corporeal form, taking a real form in the second manner, such as heat or some such. This is proven since it is a species or similitude of a thing; therefore, etc. The consequence is valid since the rainbow is an image, that is an apparition, of the sun, and from another [argument] since the rainbow is exactly like an image in a mirror, which is a species as is evident. And that the rainbow may be an image is evident in the text because it is said there that it is necessary to imagine mirrors for the

128

Representation and Objects of Thought in Medieval Philosophy generation of a rainbow just as for the generation of the halo which would not be unless the rainbow appeared by mirrors.14

So rainbows and mirror images are ‘spiritual’ forms but not imaginary forms. They are real in the first sense. They are not real in the second sense because they do not inhere in their objects in such a way as to exclude other forms. ‘Esse spirituale’ was never a term reserved for mirror images and rainbows. Oresme himself takes intelligible species to be the paradigms of things with esse spirituale and this use for the kind of being an object has as thought seems to be a basic use at least from the first half of the thirteenth century. We find it for example in the Speculum Animae of Roger Bacon’s contemporary Richard Rufus of Cornwall. The Speculum Animae is divided into five questions which culminate in the fifth question ‘concerning the cause of the immortality of the soul’. After distinguishing various senses in which a soul might be said to ‘die’ Richard turns to the question of whether it could be corrupted either per se or per accidens. He argues that it cannot be corrupted per se because it neither has a contrary nor is a composite of contraries. He then turns to whether it could be corrupted per accidens. This, he explains, is the way a material form ‘necessarily needing a subject and constituted in being through a subject’ (necessario indigens subiecto et consitutua in esse per subiectum) might be corrupted – and so his argument turns to whether the soul could be or could depend on a material form. Rufus attributes to Averroes the basic argument that the soul is not a material form. He writes: You should also know that by the moderns this consideration of Averroes’s is correctly accepted, namely that this intellect receives all material forms and ‘everything which receives is stripped of what has been received or of what is to be received, therefore this intellect is none of these material forms. The rational soul is not a material form therefore it is neither divisible through division of a subject nor constituted in being through a subject.15

But Rufus goes on at one to raise two cavillationes. The first is that since the intellect can understand immaterial things as well as material things the argument proves too much; if sound it proves that the soul cannot be an immaterial form either! The second is that strictly speaking the soul does not receive any material form because as the intellect receives forms they are ‘species intelligibiles, spirituales ideas’ (SA, q. 5, 41) received ‘secundum esse abstractionis a materia et ab esse materiali’ (SA, q. 5, 40).

14 See McCluskey, 1974, 288–9. 15 ‘Scis etiam quod a MODERNIS valde accepta sit haec ratio ipsius Averrois, scilicet quod quia iste intellectus recipit omnes formas materiales et omne recipiens denudtum est a recepto seu a recipiendo ergo iste intellectus nulla est istarum formarum materialium. Non est anima rationalis forma materialis, ergo nec est divisibilis per divisionem subiecti, nec constitute in esse per subiectum.’ (SA, q. 5, 40).

The Matter of Thought

129

Rufus in a way endorses the second cavillatio. He too accepts that the soul is immaterial but not because if it were material it would be unable to understand material things. His argument is, rather, that because it receives forms as ‘species intelligibiles [et] spirituales ideas’ which are ‘abstractum ab omni situ et corporeitate’ it must itself be ‘non situalis’. Since everything material is in situ the soul is not material. Rufus’s way with the first cavillatio is to reject the confusion to which the terminology of esse spirituale is prey, the confusion of the mode of existence of thought-objects with the mode of existence of spirits. The soul, he claims, is an immaterial form and receives immaterial forms – but it does not receive them (any more than it does material forms) according to their esse naturale. In his terminology it receives them as formas species not as formas naturas. This distinction, paralleling the distinction between esse naturale and esse spirituale which Rufus takes up in the first question of the treatise enables him to grant the principle that ‘omnis recipiens denudatum est a recipiente’ while still maintaining consistently that the soul is an immaterial form which can grasp itself. Rufus’s position here is a model of consistency. He rejects Averroes’s argument and agrees with the objector that if it were good a parallel argument would force the conclusion that the intellect and hence the soul is not an immaterial form either. In order to preserve the claim that anything could be an object of thought for the intellect while rejecting Averroes’s argument he clearly distinguishes between the esse naturale and the esse spirituale of a form and identifies the form with the latter with what he calls species forma. Rufus has put his finger on the problem to which the tradition of discussion of Aristotle’s argument gives rise and which we left dangling above. Why can an intellect not have a given form and still be able to think both that form and those which could not co-exist with it? If we suppose that the form in the intellect is the very same form we find in the world present in the same way and we suppose that to think is to take on a form then the answer is plain: one cannot take on a form one already has. If we suppose that the intellect is the very same form we find in the world and we suppose that to think is to have a form present in the intellect the answer is also plain: one cannot take on a form incompatible with a form one has. Thus, depending on the answer we give to what thinking is we get the Alexandrian and the Themistian ways of taking Aristotle’s argument. But once the notion of esse spirituale has been introduced the answer is much less plain. If the form in the intellect is not present in the same way it is in the world but rather in the way it is thought to be present in the medium between perceiver and perceived then its ‘spiritual’ presence in the intellect should no more rule out the ‘natural’ presence of a form there than would the presence of one color in the medium between perceiver and perceived rule out the presence of another. It is the hallmark of ‘spiritual’ forms that they can co-exist with each other – why can they not co-exist with natural forms? Had Rufus’s way of framing the problem become standard this problem in the form in which I just presented it might have led to the abandonment of Aristotle’s argument for the immateriality of the intellect. But Rufus’s framing of the problem

130

Representation and Objects of Thought in Medieval Philosophy

did not become standard, what did was a very different picture worked out by Aquinas and others in which esse spirituale came to be conceived as the status a form derived from a material object in the outside word came to have after a process of dematerialization. In Aquinas’s view the process begins with the emission of sensible species from objects and their multiplication in the media between perceiver and perceived. Aquinas holds that it is the form of the perceived thing itself which is carried by the sensible species and he accepts the tripartite division of the cognitive faculties into sensation, imagination and intellect. In his view when the sensible species is impressed on the sense organ there arise other species that when propagated to the imagination give rise to a phantasm which, though still particular, is fitted to be acted upon in such a way as to give rise to intelligible species. These intelligible species are the form of the thing perceived but now stripped of the matter in which it had been present in the individual sensed. At this stage Aquinas supposes a rather complex further process in which intelligible species give rise to one another. The last of these intelligible species in the process is the so-called species impressa which Aquinas describes as a conceptus and as an imago rei. For the ontological status of items fairly far along in this process Aquinas is content to use the terms esse intentionale or esse spirituale. In Aquinas’s picture esse spirituale is not really a different intrinsic mode of existence of a form. It is rather presence of a form without the matter which usually accompanies it. It is thus only the forms of material objects which can be present either naturally or spiritually. Angelic forms would have only one mode of presence, call it what you will. Within this framework the original Aristotelian argument makes more sense. If a form is present then its presence would prevent both its becoming present again and the presence of incompatible forms. That it is present dematerialized should have little to do with it. Thus if one takes esse spirituale the Thomist way Aristotle’s argument presents no mystery – but the co-presence of incompatible forms (of color for example) in the same space between perceivers and perceived should. Esse spirituale was not the only term used for kind of being an item had indifferently either in a medium or in a mind. We can see this by seeing how at the end of the thirteenth century Peter John Olivi conducts his argument against ‘species’. Olivi is interested in showing that in the most straightforward cases there is no need to posit anything besides the perceiver/knower and the object perceived/ known to explain what is going on in perception/knowledge. Olivi bolsters his own account with a critique of the multiplication of species view. The core of this critique is a dilemma posed to the believer in species about the ontological status of such species. Either they are corporeal with ‘esse naturale et sensibile’ or they are merely intentional with ‘esse intentionale et spirituale et simplex’. Olivi argues that on the first alternative each species in the chain of species between perceiver and perceived should represent that which immediately preceded it, while on the second alternative the propagation of such species from the corporeal object to the corporeal sense

The Matter of Thought

131

organ is utterly mysterious since a merely intentional object cannot pass across space or alter a physical object like the air or eye.16 As interesting for my purpose as Olivi’s crticisms is his terminology. Of particular interest is his explicit conflation of esse intentionale and esse spirituale. Here we see exactly what these terms originally meant. Esse intentionale is the esse that an intentio has. It is esse spirituale for the reasons of which Roger Bacon complained. Intentional being then began as the being which intentions have. Given the many burdens intentions have to bear it would be no surprise if it turned out that there was no single kind of esse which could bear them all. Olivi’s criticisms of the theory of species are at least a charge that the defender of such a theory has tried to invent an entity which is both a physical object and so able to enter causal relations like other physical objects, and yet enough of a nonphysical object so that it is able to represent in a more robust sense than simply be the effect of. He thinks that nothing can fill both these roles and so the theory of species is deeply confused. This was the background against which Descartes worked out his theory of ideas. As is relatively well known the idea of idea as an item present in a mind was old – deriving from late Platonic attempts to identify the forms with the products of the mind of the Demiurgus. A striking formulation can be found as early as Philo of Alexandria’s De Opificio Mundi. Less well known is that, as we saw above, the idea of ideas as items in a human mind is present as early as Richard Rufus. It appears again in Durandus of St Pourcain and it seems to be through him that it is transmitted to sixteenth-century scholasticism. To what extent Descartes was aware of this tradition I don’t know. Descartes never suggests he thought of his use of ‘idea’ as original. In any case we can now see him as trying to solve a problem bequeathed him by scholastic Aristotelianism – the problems with which I began. Aristotle had claimed that an intellect could only understand X if it were itself not-X. Taken together with the premises that the intellect could understand every corporeal thing his suggested that the intellect was incorporeal and taken together in turn with the premises that every material thing was corporeal that the intellect was immaterial. But, as Rufus pointed out, this proved too much. It seemed, for example to rule out self-knowledge altogether. Moreover there seemed a competing line of thought which compared items in the intellect to images in a mirror – whose status seemed to be that of incorporeal and so immaterial yet natural objects. That these were two distinct ways of thinking about items in the intellect was masked to some extent by the terminology of esse spirituale and its conflation with esse 16 Olivi raises the problem of why on the species account the species doesn’t represent the species which gives rise to it in II Sent., q. 73 quoted in Tachau (1988), 44: ‘Item ad primum, species influxa prius et fortius et magis proprie et conformius repraesentabit speciem a qua immediate gignitur quam aliam. Sed ab obiecto distanti non potest fieri species in visu nostro nisi per aliquam genitam in medio. Ergo prius et conformius at magis proprie repraesentabit illam speciem quam illud obiectum.’

132

Representation and Objects of Thought in Medieval Philosophy

intentionale and other terms like esse obiectivum. Olivi’s critique of species focused on this conflation. Later writers like Buridan, rejected the idea that intellectual items had to have a special ontological status but their alternatives did not win universal acceptance and by Descartes’s time the Thomist synthesis was again on the table in a number of different forms. Seen against this background Descartes’s formulation of the theory of ideas is striking. He baldly insists that ideas have both natural (formal) and objective (intentional) being. Considered naturally they are modes of mind. Considered objectively or intentionally they are objects or natures. This picture of ideas raises a number of difficult metaphysical questions about the kinds of distinction there can be between the mind and its ideas and between objects with formal reality and those objects with objective reality. But though it raises problems it solves some too. By separating sharply the formal and objective reality of an idea Descartes was able to keep his optics purely mechanical. All optical phenomena are physical phenomena in the strongest sense. At the same time he was able to avoid the perils of the naturalization of intentionality which the model of esse spirituale had served. As Deborah Brown has argued it enables Descartes to be at once a direct realist and a representationalist.17 If the story presented here is accurate much of the history of theories of thought has been driven by efforts to make sense of and to work out the consequences of Aristotle’s mysterious argument for the conclusion that the mind cannot be of a sort it can understand. The aftermath of these efforts is with us still in the guises of the adverbial and relational theories of mind. I am tempted to think that the history should give pause to anyone who thinks that thought can either be modeled on perception or can be in any interesting sense ‘naturalized’. But, if I can be forgiven a last Latinate pun, that is just speculation! Bibliography Bacon, Roger, Opus Majus, ed. J.H. Bridges, London, 1900. Brandom, R. (1994), Making It Explicit: Reasoning, Representing, and Discursive Commitment, Cambridge, MA: Harvard University Press. Lindberg, D. (1976), Theories of Vision from al-Kindi to Kepler, Chicago: University of Chicago Press. McCluskey, S.C. (Jr) (1974), Nicole Oresme On Light, Color, and the Rainbow: An Edition and Translation, with Introduction and Critical Notes, of Part of Book Three of His ‘Questiones super quatuor libros Meteororum’, PhD dissertation, The University of Wisconsin: Madison. Normore, C. (1986), ‘Meaning and Objective Being: Descartes and His Sources’, in A.O. Rorty (ed.), Essays on Descartes Meditations, Berkeley: California University Press.

17 Cf. Deborah Brown’s chapter in this book.

The Matter of Thought

133

Richard Rufus of Cornwall, Speculum animae, ed. R. Wood (forthcoming) [SA]. Ross, J. (1992) ‘Immaterial Aspects of Thought’, Journal of Philosophy, 89(3): 136–50. Schroeder, F.M. and Todd, R.B. (1990), Two Greek Aristotelian Commentators on the Intellect: The De intellectu attributed to Alexander of Aphrodisias and Themistius’ Paraphrase of Aristotle, De anima 3.4–8/Introduction, Translation, Commentary and Notes, Toronto: Pontifical Institute of Mediaeval Studies. Tachau, K. (1988), Vision and Certitude in the Age of Ockham, Brill: Leiden.

This page intentionally left blank

Chapter VIII

Objective Being in Descartes: That Which We Know or That By Which We Know?* Deborah Brown

Oh incompetence! My dreams never seem to engender the creature I so hunger for. The tiger does appear, but it is all dried up, or it’s flimsy-looking, or it has impure vagaries of shape or an unacceptable size, or it’s altogether too ephemeral, or it looks more like a dog or bird than like a tiger. (Jorge Luis Borges, ‘Dreamtigers’, 294)

What is it for an idea to be of something, a tiger say, and yet neither resemble a tiger nor represent it as it is? What would make it the idea it is, namely, the idea of a tiger? Answering this question on behalf of Descartes would, I think, tell us much about how he understood the relationship between the mind and its objects. It would also tell us much about how he conceived of the relationship between the knower and the known. There is a certain picture of Descartes’s theory of ideas which is standard and which I would like to join the challenge against. It has been challenged before by Brian O’Neil,1 Calvin Normore2 and Lilli Alanen3 among others4 and much of what I have to say is an extension of their ideas. The picture we all reject is this one: Rather than securing a firm foundation for knowledge, Descartes erects between * I would like to thank my friends in the Inter-Nordic community of scholars working on medieval and early modern conceptions of mind (and honorary members of this community from other continents) for stimulating discussions about this topic. In particular, I am grateful to Lilli Alanen and Calvin Normore and to the participants and organizers of the conference Intellect, Knowledge and the Object of Thought from 1200 to 1700 Oslo, November 24–6, 2000 and to the editor of this volume, Henrik Lagerlund for all assistance and helpful conversations. 1 O’Neil (1974) directs our attention to the strong form of direct realism in the Regulae and the influence of Thomism throughout Descartes’ works despite his rejection of substantial forms. See also Kemp Smith (1952), 51–2, and Beck (1952), 72–4. 2 See Normore (1986). 3 See Alanen (1990). 4 See also Nadler (1989), Monte Cook (1987), Yolton (1975) and Lennon (1974). For a more sceptical attitude towards this trend towards reading Descartes as a direct realist, see Hoffman (2002).

136

Representation and Objects of Thought in Medieval Philosophy

the knower and the known a ‘veil of ideas’ or intermediate objects of thought and perception. Barry Stroud paints the picture aptly when he describes Descartes’s sceptical conclusion in the First Meditation ... as implying that we are permanently sealed off from a world we can never reach. We are restricted to the passing show on the veil of perception, with no possibility of extending our knowledge to the world beyond. We are confined to appearances we can never know to match or deviate from the imperceptible reality that is forever denied us.5

This way of framing the problem invites a number of cheap responses and discounts Descartes’s own anti-sceptical arguments. Stroud’s targets are the cheap responses: the attempts ‘to minimize the seriousness of the predicament, to try to settle for what is undeniably available to us, or perhaps even to argue that nothing that concerns us or makes human life worthwhile has been left out.’6 But few who take seriously the sceptical challenge would also take seriously Descartes’s claims in the Second Meditation to have found a paradigm of certainty in the idea of the self. No amount of attention to the ‘internal marks’ of indubitability in an idea, clarity and distinctness, can, in the words of one scholar, address the ‘fatal objection’ that the possession of an idea does not in itself justify our claims to know about such matters as the existence of God and the nature of substance.7 What is evident from such objections is the assumption that at the end of the day Descartes is a representational realist. Although there may be variations on any theme, one way to characterize the representational realist is as one who holds to the following three propositions: (i) The only immediate objects of knowledge are the mind’s ideas or internal states (concepts or percepts). (ii) The mind knows indubitably its own ideas or internal states.8 (iii) Knowledge of extramental reality depends on an inference from knowledge claims about the content of one’s ideas to claims about what those ideas purport to represent.9

5 See Stroud (1984), 33–4. 6 Ibid., 34. 7 See Ashworth (1972), 105. See also Gibson (1932) on the ‘nemesis’ of representative perception: ‘for if we know only through the medium of representative ideas, how do we know there is anything except ideas?’ (79) 8 Perhaps it is not necessary for a representationalist to claim that the mind has indubitable knowledge of its own ideas or objects which are somehow internal to consciousness but it is unclear to me why anyone would bother with a representationalist epistemology were there not some such guarantee regarding the inner sanctum. 9 Differences within representationalism may depend upon whether or not ideas themselves are regarded as the intermediate objects of perception and knowledge or as acts directed at distinct objects (e.g., sense data). See Hoffman (2002) for five ways of drawing

Objective Being in Descartes

137

On the representationalist reading Descartes is seen as subscribing to these basic tenets of representationalism. He is thus regarded as having regressed away from the direct realism of those Scholastics who, like Aquinas, espoused a non-inferential immediate awareness of extramental natures. I suspect that there is very little evidence to support the reading of Descartes as a representationalist. In fact, there is very little reason to believe that Descartes subscribed to these three propositions in their present form or, if he did, that his endorsement supports a reading of him as a representational realist. I shall not here attempt to evaluate the success of Descartes’s arguments against skepticism – that is well beyond the scope of this paper. My concern is with the form of his theory of our epistemic access to the world. The charge that Descartes is a representational realist has (at least) two sources: the Demon thought experiment and Descartes’s representational theory of mind. Although my concern is primarily with the latter, what I have to say about it bears on the former. The Demon hypothesis seems to support the interpretation of Descartes as a representationalist insofar as it is construed in the following way. Descartes appears to present two possible worlds: the Demon world in which all our ideas of the external world misrepresent it (either by representing it as existing when it doesn’t or by representing it as other than it is) and the world Descartes sees himself as proving is the actual one, a world in which at least an important subset of our ideas, upon critical reflection, can be known to correspond to the way things are. What is important for this reading of Descartes is that in both scenarios the ideas are the same. Thus what we know directly in either scenario is the same – namely, ideas – and the task for the meditator is to figure out what can be inferred from those ideas about the external world. That the Demon hypothesis supports a representationalist reading of Descartes thus depends on a certain view of Cartesian ideas as being objects the identity of which is independent of the way the world actually is. But to draw any conclusions from the Demon hypothesis about the kind of realism to which Descartes subscribes, one has first to establish that there is a logical connection between representationalism as a theory of mind and representationalism as a theory of knowledge. The representational theory of mind is generally understood as defining thought as the processing of symbols or mental representations – that is, as involving some kind of commitment to realism about mental representations. That is a description sufficient to distinguish representationalism from behaviorism and eliminativism but leaves it an open question precisely what the nature of thought is – i.e., whether it is best understood by analogy with language or on some other model is up for grabs. It is also a further question whether the representational theory of mind entails representational realism. If there is an entailment between the two theories then those who believe that Scholastics like Aquinas are direct realists are in trouble for the Scholastics were as much committed to the representational theory of mind as Descartes was. Paul Hoffman has for this reason recently suggested the distinction between direct and representational realism. But I assume that any theory of representationalism will be committed to something like these three propositions.

138

Representation and Objects of Thought in Medieval Philosophy

that Aquinas might also be a representational realist.10 The fact that, according to Aquinas, ‘the intelligible species is not that what is understood, but that by which the intellect understands’11 does not rule out, on a reading such as Hoffman’s, that the intellect knows the natures of material things by first cognizing their likenesses (similitudines).12 But why think that representationalism about the mind does entail a representationalist epistemology? Let us look at the question from the other direction. What exactly is it about the representational theory of mind that seems to conflict with direct realism? One likely answer is that the link is forged by a certain picture of what ideas are. Descartes follows a tradition of thought which characterises intentionality in terms of the objective existence of things. On this picture there are what we might refer to loosely as two modes of being: the being of a thing as represented by the mind (objective being) and the non-representational being something has as either a mode or a substance (formal being). It is the objective being of ideas which, Descartes claims in the Third Meditation, accounts for the content of ideas and thus distinguishes one from the other: In so far as ideas are simply modes of thinking, there is no recognizable inequality among them and all appear to proceed from me in the same way. But in so far as one represents one thing, another another thing, it is clear that they differ from each other greatly. (AT, VII, 40)

My question is whether this picture of representation is incompatible with direct realism. An incompatibilist might reason thus: On the objective existence theory, thought is a relation between the thinker and things which have objective being in the mind. The mind is only directly aware of things which appear on its private stage; thoughts about extramental objects are thus mediated by inference from thoughts about objective existents. But this shows us that the link between representationalism as a theory of mind and representationalism as a theory of knowledge is mediated by a certain understanding of the former. It is not really representationalism per se which compromises direct realism but objective existence versions of it in conjunction with the idea that what the mind knows primarily and indubitably are only things with objective reality. We are now at the heart of the matter. It is supposedly objective existence versions of the representational theory of mind which are incompatible with direct realism. It is my view that there is nothing inherent in an objective existence version of the representational theory of mind, provided it is understood a certain way, that entails representational realism. To get our bearings on this topic it will be useful to take a short detour through the major accounts of objective reality in the middle ages.

10 See Hoffman (2002), 165–7. For a detailed account of Hoffman’s Thomistic reading of Descartes Hoffman (1996). 11 See ST, I, q. 85, a. 2. 12 Ibid., I, q. 85, a. 2.

Objective Being in Descartes

139

Esse Objectivum Before Descartes The history of the notion of objective being has been canvassed by others and I do not propose to repeat the story here.13 My purpose is to draw attention simply to those aspects of the tradition prior to Descartes which test the credibility of a necessary conceptual link between the theory of objective existence and representational realism. The concept of esse obiectivum figured in debates about the objects of visual illusions and in debates over the status of universals during the Middle Ages. On socalled ‘perspectivist’ theories such as Alhazen’s, Roger Bacon’s and Peter Aureol’s when one is subject to a sensory illusion what is perceived is something with ‘diminished’, ‘apparent’ or ‘objective’ being. The following example from Aureol illustrates nicely one problem objective being was intended to solve. When one is carried on the water, the trees existing on the shore appear to move. This motion, therefore, which is objectively in the eye (in oculo obiective) cannot be posited to be vision itself; otherwise vision would be the object seen, and a vision would have been seen, and vision would be a reflective power. Nor can it be posited to be really in the trees or in the shore, because then they would really have moved. Nor can it be posited to be in the air because it is not attributed to the air but to the trees. Therefore, it is only intentionally (tantum intentionaliter), not really, in seen being and in judged being. 14

Aureol reasons thus: Since the motion which is seen cannot reside in the trees themselves, nor the air, nor do we perceive what takes place in the eyes, it must reside in something else: esse viso iudicato et apparenti. In contemporary theories of perception, visual illusions have provided the best argument for representational realism. If your awareness is the same in both cases of veridical and non-veridical perception, then, so the argument goes, what you must be aware of in both cases is the same. It follows that in the normal (i.e., veridical) case what you are aware of is the sensory state itself or sense data. I doubt that this argument would have persuaded many philosophers of the Middle Ages. Objective entia were not the percepts or concepts or properties of these things but objects, albeit with diminished, apparent, judged or objective being. For perspectivists such a Aureol they had a kind of ‘third realm’ status: neither mind-independent extramental objects nor intramental objects. From an ontological perspective positing objective beings incurs no greater cost in terms of additional beings than sense data views. There is no obvious reason to favor (and perhaps every reason to avoid) the idea that what I see when I see a bent stick in water and what I see when I see a stick which is really bent are properties of the mental or visual event itself over the idea that what I see are two different orders of object. The latter idea at least has all the advantages of being compatible with direct realism. 13 See Normore (1986), Tachau (1988) and Read (1977). 14 Peter Aureol, Scriptum in I Sentarium, lat. 329. d. 3, s. 14, a.1; II:696. See Tachau’s discussion (1988), 89–100.

140

Representation and Objects of Thought in Medieval Philosophy

What then of the argument that universals are objective entia? Although the notion of objective being was not confined to nominalist theories (some argued that universals had subjective being in singulars and objective being as objects of thought) it was enormously convenient for nominalists who wanted to avail themselves of universals to account for general thought without admitting them into the ontology in any serious way. In Ockham’s early works, a universal is a fictum – ‘that which immediately terminates the act of thinking when no singular thing is thought of.’15 When I think abstractly of the kind horse without thinking of any particular horse what I think of is neither a quality of mind nor something with formal being but something with objective being.16 Is there anything in this picture which compromises direct realism? Certainly, critics of the nominalist view such as Walter Chatton thought so. Chatton argued that things with objective being must be either really distinct from mental acts or not. If not, then for every thought of the same type there is a distinct fictum, which is a violation of Ockham’s razor. If, however, an act and its fictum are really distinct then the one can exist without the other ‘but then there would be objective and intellective being without any intellection, which is a plain contradiction.’17 Immunity against this argument could have been bought, as it seems to have been by Aureol, but only at the cost of admitting objective beings to a third realm. For one could just as easily come down on the side of the object rather than the mental act at this point. How objective beings could be mind-dependent yet have neither qualities of the mind nor of extramental reality is a mystery to be sure but not obviously contradictory. Alternatively, one could accept the identity of act and object but still find some use for thinking of mental acts in objective terms, that is, as defined by the directedness of the act towards an object. In any event, there is nothing essentially representationalist about this theory of universals. Ficta are not mental representations which are known prior to singulars but objects acquired through the same process in which singulars are known and through which singulars are known in a certain way. Persuaded by his confrere’s arguments, Ockham, however, abandoned the concept of ficta altogether and the act-object distinction it seemed to presuppose. In the later intellectio or act-only theory, the process of acquiring a general concept by abstraction from the concept for a particular human is described thus: First a human is apprehended (cognoscitur) by some particular sense, then that same human is apprehended by the intellect and when (the human being) has been conceived, a general notion common to all humans is formed. This apprehension (cognitio) is called a concept, intention or passion which is a common concept to all humans and when it exists 15 William of Ockham, Ordinatio, 274 16 Stephen Read notes that once the notion of objective existence caught on it dominated the discussion about universals for two decades and had the additional benefit of providing a point of reference for the idea of ‘existing in the understanding’ in Anselm’s Ontological Argument. See Read (1977), 20. 17 Walter Chatton, Lectura, I, d. 3, a. 2, quoted in Gal (1967), 202–3.

Objective Being in Descartes

141

in the intellect, the intellect immediately knows that a human is something without any process of reasoning (sine discursu).18

What is lost when one gives up the concept of objective existence? Not much on Ockham’s account and to some extent I am inclined to agree. But there is something unsatisfying about the act-only theory of representation akin to what has struck some philosophers as so deeply unsatisfying about causal theories of reference. It is hard to give up the intuition that ideas don’t merely represent their causes. Borges’s idea represents the tiger as more like a bird or as having impure vagaries of shape or as having an unacceptable size. The attempt to build these features of ideas into the description of the act – the bizarre constructions of adverbial theories (e.g., thinkingtiger-as-a-bird-ly) – is clumsy and has little economical advantage over retaining objects of thought. Act-only theories also do little to reduce the unease produced by thinking that in our first encounters with objects we cannot be in error. On Ockham’s later view, our first experience of a tiger should be sufficient for acquiring the concept of a tiger regardless of whatever dreamlike qualities it may be infected with. Which brings me to Aquinas, the last stop in this all-too-brief tour of the history of esse obiectivum and an odd place to stop since Aquinas, to my knowledge, never used the terminology nor, in using the notion of intelligible species, had in mind anything like Ockham’s ficta or Aureol’s apparent beings. It might seem more natural to stop the tour with Suarez, the one whose doctrine of objective being Caterus had in mind when he objected to Descartes that he was unjustifiably reifying the notion of objective being in his proof of God’s existence. (AT, VII, 92–3.) But Descartes’s own view has less in common with Suarez’ than Aquinas’s doctrine of intelligible species for Suarez thinks of objective being as simply a way of talking about extramental objects insofar as they are thought about or ‘denominated’ (esse cognitum quoad denominationem)19 whereas Descartes makes it clear in his reply to Caterus that objective being is for him being ‘in the intellect in the way objects usually are there.’ (AT, VII, 102). Unlike Suarez, moreover, Aquinas has a role to play in explaining our contemporary use of the notion of intentional object through his influence on Brentano although this too is a peculiar piece of philosophical history.20 Brentano 18 Ockham, Summa logicae, III, 2, c. 29, lines 14–22: ‘Non quod isti conceptus praecedant notitiam intuitivam hominis, sed iste est processus quod primo homo cognoscitur aliquo sensu particulari, deinde ille idem homo cognoscitur ab intellectu, quo cognito habetur una notitia generalis et communis omni homini. Et ista cognitio vocatur conceptus, intentio, passio, qui conceptus communis est omni homini; quo existente in intellectu statim intellectus scit quod homo est aliquid, sine discursu. Deinde apprehenso alio animali ab homine vel aliis animalibus, elicitur una notitia generalis omni animali, et illa notitia generalis omni animali vocatur passio seu intentio animae sive conceptus communis omni animali.’ 19 Francisco Suarez, Disputationes metaphysicae, 25, 1, 32. 20 Peculiar because, again, Aquinas does not use the terminology which suggests that despite Brentano’s famous assertion that he derives the notion from the Scholastics, it was Descartes’s use (on which Brentano frequently lectured) which was more influential on Brentano and thus the modern tradition. See Brown (2000).

142

Representation and Objects of Thought in Medieval Philosophy

himself denied that the intentional ‘in-existence’ of objects, as he put it, was anything but a way of speaking synsemantically, (mitbedeutende) about the thinker thinking and carried no ontological weight.21 Aquinas is important for our purposes, however, as an example of someone who professes direct realism but who also subscribes to two modes of being of objects. What are Aquinas’s objects of thought if not objective entia? They are intelligible species or the forms of extramental objects. The account of concept acquisition begins with the acquisition of sensible species, the accidental forms of external objects received in a spiritual mode into the matter of the sense organs. Species thus reach the intellect through increasing levels of abstraction: first from the matter of their original objects, then from the matter of the sense organs. By means of this process, the mind has puportedly direct access to the natures of material things because the species just are the forms of extramental things existing in a spiritual or intentional mode of being in the human soul. It is the formal identity of the species both within and outside the mind which is supposed to save the account from a representationalist epistemology. At least this was the hope. Hoffman has raised doubts about the rights of Thomas to draw this conclusion and before Hoffman, A. Boyce Gibson put a similar objection delightfully thus: The medieval theory of perception was realistic; the senses are the open gates thronged by the ‘species’ which emanate by effluence from the actual object, and passing into the mind nevertheless remain what they were outside it. But if perception is representative, the external world, on its entrance to the mind, passes, as it were, through a toll-gate of unreality, and its bewildered ghost wanders about its new home, for ever doubtful of its own identity.22

The bewildered ghosts or species of material objects are the source of the doubt that Aquinas is a direct realist. The intellect is aware primarily of its own states or how it is modified (by these natures in a spiritual and universal mode) and not of how these natures are in the world, namely as individualized forms of matter. Hoffman thinks that the very fact that the mind is able to know external natures because the intelligible species resembles extramental objects is indicative of a representationalist strand in Aquinas’s realism.23 But resemblance is not a primitive concept in the theory but one analysed in terms of the formal identity of species and external objects and there is no suggestion in Aquinas’s account that one must first establish that the resemblance holds in order to know external natures. The success of the theory depends on the formal identity of the intelligible species and the forms of external things, an idea that taxes the modern mind too highly, but which does not on that account make the theory a form of representational realism. No inference from knowledge of intelligible species to extramental natures is suggested or required by the theory. 21 See Brentano (1874/1973), 332. 22 See Gibson (1932), 79. 23 See Hoffman (2002), 176.

Objective Being in Descartes

143

An inference is however required in obtaining knowledge of particular things on Aquinas’s account. Singular knowledge is indirect and inferential. In this way, therefore, (the intellect) understands the universal itself directly through an intelligible species; it understands singulars however indirectly, of which there are phantasms.24

Aquinas’s treatment of singular knowledge is fraught with difficulty. The intellect trades in universals. Knowledge of singulars is relegated to the province of the senses. The intellect must ‘revert to the phantasms’ (convertendo se ad phantasmata) to be said to know singulars. Strictly speaking, the senses do not know anything. At De Veritate, I, q. 2.6, Aquinas asserts that it is the whole human being who knows singulars, not the intellect and not the senses. How the whole human being can be said to know that which no part of it knows is not, however, further clarified. In this domain there is direct perception by the senses but not direct knowledge and all because of what seems to be an excessive attachment to the generality of thought hypothesis. The fact that we do not know directly individual things or matter (hence not the whole of extramental reality) does not, however, detract from the claim that the mind knows directly the natures of external things. After all, they are in some sense the same thing as intelligible species. Objective Reality or Perception in Descartes What, besides the terminology, does Descartes’s notion of objective being have in common with the notion as it appears in these traditions? Descartes rejects the notions of substantial forms and species in either the medium or the mind. Rather than conflating concepts with species, Descartes adopts the Platonic terminology of ‘ideas’ but, as he is commonly understood, drops the association between ideas and ‘exemplars’, copies of the Divine ideas, or models for creation. Descartes’s ideas are dependent upon their causes both for their formal reality and, importantly for our purposes, their content. (AT, VII, 41–2.) Descartes also modifies the long-standing assumption that ideas represent by virtue of being similar to what they represent. Aquinas’s insistence that species are likenesses (similitudines) of external things may not be sufficient to make him a representational realist but it certainly compromises his realism. If relations of similarity are to be understood in terms of shared properties of the relata, it is difficult to see what properties material objects could share with their ‘bewildered ghosts’ in a dematerialized intellect. Moreover, as Descartes points out in connection with his discussion of the two ideas of the sun, there is often a great disparity between the idea and its object. The idea of the sun which emanates more directly from the sun itself and, one would think, puts us more directly in contact with the sun, the idea of the sun as a small yellow disk, is the one which bears the least resemblance to its 24 See ST, I, q. 86, a.1

144

Representation and Objects of Thought in Medieval Philosophy

object (AT, VII, 39). It is possible, therefore, on Descartes’s account, to have an idea which bears little, if any, resemblance to that of which it is an idea. It is also possible to have contradictory ideas of the same thing, both of which cannot be of the thing by virtue of resembling it (AT, VII, 39). Descartes’s account of objective being commits him only to the claim that things can have two modes of being, formal and objective, to the identity of the object outside the intellect and inside it. There is nothing in his use of this notion which commits him to the claim that ideas must resemble that of which they are ideas or that ideas and their formally existing objects have any properties in common. Descartes tells us instead at the beginning of the Third Meditation that ideas are ‘just as if certain images of things’ (veluti quasdam imagines) which may not seem to advance the debate much at all (AT, VII, 42; also, 36–7). Are not images (at least in his time) representational by virtue of resembling their res representata? This is certainly how Descartes is often interpreted: as arguing that there is both a causal or referential and a resemblance or non-referential constraint on true representation. I doubt that this is his point in stating that ideas are ‘as if images’ of things. Descartes refers to his habit of judging that the ideas he finds within himself resemble reality to be the greatest source of his previous errors (AT, VII, 37). When he later reintroduces the analogy between ideas and images, rather than relying on any claims about the power of images to resemble reality his point seems more to establish that ideas are like images or copies in the sense of not being able to have a greater degree of perfection than their causes (AT, VII, 42). Thus it is clear to me, by the natural light, that the ideas in me are like images which can easily fall short of the perfection of the things from which they are taken, but which cannot contain anything greater or more perfect. (AT, VII, 42; CSM, II, 29)

The picture analogy serves to make, however, one useful point. The representationality of an idea cannot be reduced to the formal or ‘subjective’ properties of the idea – the properties an idea has by virtue of being a mode of mind any more than, as Descartes explains to Regius, one could expect to paint pictures like Apelles by arranging patterns of paint on a canvas in the same way as Apelles. (June, 1642; AT, III, 566–7.) The problem of representation Descartes is sketching here would not go away with a materialist theory of mind, a point with which many contemporary externalists about mental content concur. Denying that Descartes’s theory of ideas is in any straightforward way a resemblance theory is not to say that there is no non-referential component to ideas as Calvin Normore has pointed out.25 Indeed, ideas could not provide occasions for error if they did not present their objects in certain ways. It is this aspect of ideas which Margaret Wilson, careful to avoid resemblance talk in her early discussion of the notion of objective reality, designated the ‘representational character’ of Cartesian ideas.26 Although this notion only shifts the question of how to understand the notion 25 See Normore (1986). 26 See Wilson (1978), 102.

Objective Being in Descartes

145

of representation for Descartes to the question of how to understand the notion of ‘representational character,’ Wilson is right to have marked it as distinct from the notion of resemblance. Wilson’s representational character is close to what I think of as the non-referential component of ideas but whereas for her the representational character of an idea is distinct from the objective reality of the idea, and hence an ‘embarassment’ for Descartes, for me the non-referential component is a component of objective reality.27 I shall return to this point below in discussing the notion of ‘material falsity.’ As Descartes explains perhaps unhelpfully in the Replies to the First Objections, the objective being of ideas ‘signifies nothing other than being in the intellect in the way in which objects usually exist in the intellect’ (AT, VII, 102). The distinction between formal and objective being was not new but what does seem to be new is Descartes’s further claim that both the formal and objective aspects of ideas need a cause. For if we suppose that something is discovered in an idea which cannot be in its cause, it must therefore have this from nothing; nevertheless, however much imperfection is in that mode of being, by which the thing exists objectively in the intellect through the idea, plainly it is not really nothing, nor consequently can it exist from nothing. (AT, VII, 41)

The causal principle Descartes uses to argue for the existence of God is an extension of the Scholastic principle that there cannot be more reality in an effect than is contained formally or eminently in its cause. For Descartes an idea can have less objective reality than its cause has formal reality but it cannot have more which is why the idea of an intricate machine cannot have been caused by anything less than a machine with that much intricacy formally or a knowledge of engineering (AT, VII, 103–4.) The distinction between formal and eminent containment can be characterized in the following way. If an idea, A, represents an object X as F, then either F-ness is contained formally in its cause or, if the cause of A’s representing X as F does not contain F formally, it must have a greater degree of reality than the content of A. It might be thought that eminent containment is added to account only for the idea of God. Since God has no modes it cannot be that God has formally what our idea of God contains objectively. But eminent containment may also be useful to explain ideas which misrepresent their objects. For example, our non-astronomical idea of the sun contains features which are not contained formally in the sun but the sun has a greater degree of formal reality than this idea has objective reality and so can be its cause. The referential component of an idea is given by the identity of the object existing objectively in the intellect, as either an essence or nature, for example, triangularity, or an existing particular such as the sun (AT, III, 350). How then are we to understand the non-referential component of objective being in Descartes? Calvin Normore and I have recently argued in relation to Descartes’s account of the passions and sensations that Descartes’s notion of representation is primarily 27 Ibid., 106.

146

Representation and Objects of Thought in Medieval Philosophy

representation as.28 I want to suggest here that the objective being of a thing should not merely be thought of as the thing itself but as the thing under a certain mode of presentation (though I am not suggesting in addition the Fregean ideas that the mode of presentation is really distinct from the idea of the object or that it determines the reference of an idea). Hence, in the case of the two ideas of the sun, both are ideas of the sun but they differ in how the sun is objectively presented in each idea. It follows from this view that there can be a difference in the objective reality of ideas which is not a difference in the identity of the object represented. Descartes says as much himself in the letter to *** (1645 or 1646) when he writes that a thought of the essence of a triangle, and a thought about the existence of the same triangle ‘even understood objectively differ modally in the strict sense of the term “mode.”’ even though the essence and existence of the triangle itself are not distinct. (AT, III, 350.) Notice that on Wilson’s view these are not differences in the objective reality of ideas but a difference between their representational characters. Why does Wilson argue for this distinction and why do I want to resist it? Wilson proposes that it is necessary to separate the representational character and objective reality of ideas in order to make sense of Descartes’s claims about ‘materially false’ ideas. When an idea is confused and obscure it can provide material for error to an unwary mind by representing ‘non-things as things.’ For example, it is impossible to tell from our idea of cold whether cold is a positive entity or merely the absence of heat (AT, VII, 43–4). The problem with this claim, as Arnauld rightly observed, is that Descartes’s very theory of representation would seem to require that what is conceived must be some thing, if only an objectively existing thing. To Arnauld’s objection that if cold is a privation it can no more exist objectively in the mind than formally exist, Descartes replies somewhat obscurely: I think that a distinction is necessary: for it often happens in confused and obscure ideas, among which those of heat and cold are numbered, that they are referred to a thing other than that of which they are ideas. Thus, if cold is only a privation, the idea of cold is not cold itself, as it were objectively in the intellect, but another thing which I take wrongly for that privation; truly, it is a sensation which has no being outside the intellect. (AT, VII, 233)

Wilson takes Descartes’s notion of material falsity point to be incompatible with the claim that materially false sensations have objective being.29 In so far as sensations are representations their intentionality and distinctiveness should be due to some other feature, which Wilson labels their representational character. I used to hold views like this myself but I now think that it is deeply mistaken. For one thing, it would make Descartes’s texts stupidly inconsistent. (I’m prepared to admit they may be inconsistent but not stupidly so.) The objective mode of being is not an inessential feature but belongs, Descartes writes, to ideas ‘by their very nature’ (AT, VII, 42). There is, moreover, a problem regarding the two ideas of the 28 See Brown and Normore (2002). 29 See Wilson (1978), 111.

Objective Being in Descartes

147

sun on Wilson’s reading. The two ideas of the sun would have the same objective being but differ in representational character. Here Wilson’s view implies that if ideas were just characterized by their objective being, these two ideas would be indistinguishable. But this is precisely the mistake Descartes accused Gassendi of when he, Descartes, complained that two ideas are not the same for having the same subject (AT, VII, 363). It is not at all obvious that he would have regarded the two ideas of the sun as having the same objective reality simply because they each contain the same thing objectively. Lastly, and perhaps most importantly for our purposes, if Wilson is right, then Descartes’ theory of representation would run perilously close to entailing representational realism. For Wilson’s representational character of an idea does seem to work like a Fregean mode of presentation in fixing reference by whatever fits the representational character of the idea (or not as in the case of materially false ideas) and in being that which one has primary epistemic access to. Since the representational character determines the content of a materially false idea, not its objective reality, what else could it represent than whatever resembles the representational character of the idea? Hence, Wilson argues, the whole account is deeply embarrassing because it raises an obvious objection to Descartes’s proof of the existence of God in that: ... it entails that the objective reality of an idea is not something the idea wears on its face. Descartes would have it otherwise: in his initial exposition of the concept of objective reality he seems to indicate that an idea’s objective reality is transparent, deriving directly from its representative character ...30

But if the content of an idea is determined by its representational character, Descartes could not claim that the two ideas of the sun were ideas of the same thing. The actual sun fits the representational character of the astronomical idea; the referent of the other idea should, if any such thing exists, be some yellow disk the size of a coin.31 None of this rules out the possibility that Wilson is right and Descartes’s account is simply incoherent. We should not underestimate the difficulty of reconciling Descartes’s remarks about material falsity with his objective existence theory of content. But I think we can read the exchange with Arnauld in a way that does not compromise his theory of objective being or his direct realism. In the passage where Descartes replies that the mind mistakes a sensation for cold, Descartes seems to be suggesting that it is the sensation itself which is objectively present in the idea of cold. What if the mode of presentation of the sensation were presenting not the features properly predicable of a sensation but features properly predicable, if they exist, of bodies? If there are no such corresponding features of body – for example, if cold is a privation – then the mode of presentation of the sensation will represent a nonthing as a thing. It cannot be that cold is a feature of the objective reality of the idea for the reasons Arnauld gives – if cold were a feature of the objective reality of the 30 Ibid., 112. 31 O’Neil raises this as a general worry for objective existence theories but does not explore how it might be solved within such theories. O’Neil (1974), 10.

148

Representation and Objects of Thought in Medieval Philosophy

idea, then it would have some degree of being and so cold would not be a privation and the idea of cold would not be false. But it can be a feature of the objective reality of a sensation that it is presented as a mode of body or, in Descartes’s terminology, is ‘referred’ to some external body, and then from an intentional perspective it will represent bodies as having some property as much as if the corresponding property of bodies did exist and was objectively present in the idea.32 This reading is different from standard projectionist accounts since the claim is not that either a sensation or a feature of sensation, coldness, is projected by the mind onto bodies. The point is rather that it is part of the mode of presentation of materially false sensations that they are appear as modes of body and thus are necessarily false.33 This mode of presentation of the idea of cold, like the false idea of the sun, is, however, nothing distinct from the objective reality of the idea. In what sense, then, does a materially false idea represent a non-thing as a thing? Since a sensation is something positive, the account just given suggests that materially false sensations represent some thing as it is not rather than a non-thing as a thing. But I think the latter is just a way of thinking of the representational relation involved in material falsity from the point of view of the object rather than from the point of view of the perceiver. By representing a sensation as a mode of the body, at the same time and by the same process, we represent bodies as having modes that they of metaphysical necessity lack.34 Descartes’s general view of sensation is that because the mind is aware of its sensations and because sensations present as they do that it is natural to refer them to external objects. The notion of ‘referring’ introduced in the replies to Arnauld is thus crucial to Descartes’s theory of perceptual representation. Some ideas spring from one source (e.g., the body) and are referred to another. In Les Passions de l’Ame, this idea becomes part of the very definition of passions and sensations.35 Pain is a mode of the soul but is predicated by the soul of some part of the body like the foot or the hand; a sensation of green is referred to the grass and anger to the soul but all are caused by proximal movements of the animal spirits and pineal gland (aa. 22–9). What one is aware of when an idea is false is not some intermediate object but the sensation itself, on account of the nature of which or the way it presents itself, one refers it outside the mind. Referring, however, is not inferring . One does not know 32 See Descartes’s Les Passions de l’Ame, a. 23; AT, XI, 346. 33 Compare Nadler’s projectionist reading of Descartes and Arnauld (1989), 125–6. 34 One might wonder whether it follows that all sensations are materially false. To this it might be objected that Descartes seems to allow for the possibility of materially true sensations, for example, the idea of heat if cold turns out to be a privation (AT, VII, 44). Descartes has, however, no way of ruling out that all sensations are materially false – at least nothing about a sensation tells us whether it is true or false – but if some are true, their truth could be accounted for by the presence of the material quality (e.g., heat) objectively in the idea. The epistemic shortcoming of confused and obscure ideas is not, as Wilson suggests, that they fail to wear their objective reality on their face but that they fail to wear their material falsity on their face. 35 Ibid., a. 27; AT, XI, 349.

Objective Being in Descartes

149

first some feature of the sensation and then infer to some property of bodies; if one knew that much, the idea would not be false. But this picture raises some skeptical concerns. Our referring our sensations outside the mind does not entail the existence of anything outside the mind. But this is, for Descartes unlike the Scholastics, what we should expect for the answer to the skeptic is not going to be given simply from within the theory of perception alone. When we look, therefore, for a cause of the objective reality of a sensory idea, we need look no further than human nature or the union of mind and body. For if a sensory idea is false it stems from an imperfection in my nature and, if it is true and the reality it presents is indistinguishable from a non-thing, in either case I would be right to conclude that the idea originates wholly from within me. (AT, VII, 44)

Descartes as a Direct Realist Is there anything in Descartes’s account of objective existence and material falsity which commits him to representational realism? Let us begin by re-examining the evidence in favour of three central tenets of representationalism. First: are the immediate objects of awareness ideas rather than things our ideas are about? It is true that Descartes sometimes uses representationalist language:‘... I understand by the term idea that form of any given thought through the immediate perception of which I am conscious of that very thought’ (AT, VII, 160). But given Descartes’s commitment to the objective reality of ideas as determining the representational content of ideas, being immediately aware of one’s ideas is not incompatible with being aware directly of external objects, for the objective reality and the formal reality of the object are not really distinct. The second tenet – that the mind knows indubitably its ideas – seems uncontentious but even this proposition has to be modified in light of the discussion of material falsity. What I know indubitably is the formal reality of my sensory ideas – that I am experiencing a pain, a desire, a fit of pique – but uncovering the objective reality of the idea may require some theory (Principles, I, 45–6; AT, VIIIa, 22). In cases of material falsity what I am thinking of turns out to be nothing but the modification of mind which has been caused by some external process but then what I am conscious of directly is not merely the idea but also the object of the idea, although not under that description. The third tenet is particularly important for determining whether Descartes is a representational realist or not. Does knowledge of extramental reality depend, according to Descartes, on an inference from one’s knowledge of one’s own ideas? Here are three reasons why someone may be tempted to answer this last question affirmatively. First, Descartes’s method is to proceed from an examination of the attributes of ideas (clarity and distinctness) to judgements about the natures of things. The fourth postulate of the Geometrical Exposition of the arguments of the Meditations in the Second Set of Replies states:

150

Representation and Objects of Thought in Medieval Philosophy I ask that [the readers] examine the ideas of natures in which a complex of many attributes together is contained, such as the nature of a triangle, the nature of a square, or of other figures; likewise, the nature of mind, the nature of body, and above all the nature of God. And they should notice that all those things which we perceive to be contained in those ideas can be truly affirmed of those things. (AT, VII, 163)

Second, the method succeeds only because of these internal marks of the truth of ideas: clarity and distinctness. When an idea is both clear and distinct, that is, both present and accessible to the mind and so sharply distinguished from all other perceptions that it contains within itself only what is clear, Descartes claims that it can serve as the basis for a certain and indubitable judgement (AT, VIIIA, 22). A clear and distinct idea makes perspicuous the essential features and existence of an object thereby distinguishing it from every other thing. Finally, it may be thought that Descartes’s use of the causal principle to secure knowledge of the causes of ideas introduces an inferential aspect to knowledge. We could not judge on the basis of this idea that the sky exists unless because every idea must have as the cause of its objective reality a really existing thing; which cause we judge [in this case] to be the sky itself ... . (AT, VII, 165)

Is it not the case that Descartes uses the causal principle that an idea must have a cause which contains formally or eminently all the objective reality present in the idea to establish that certain features of an idea (those about which we have a clear and distinct perception) must be contained in the cause? (AT, VII, 41). And does he not also use it in the Sixth Meditation to defeat the supposition of the dreaming and Demon hypotheses that he is either alone in the world or stuck in very bad company? (AT, VII, 79–80). In my view, the above discussion of objective reality shows that these aspects of Descartes’ epistemology do not determine that the form of Descartes’ realism is representationalist. Descartes’ claim that an idea is what the mind perceives directly does not preclude its being true that what the mind thereby perceives directly is some true and immutable nature, triangularity, or an actual existing particular such as the piece of wax before him or the sun. When, for example, I imagine a triangle, even if perhaps a figure of no such kind exists outside my thought, or anywhere, there is nevertheless its determinate nature, or essence, or form, immutable and eternal, which is not produced by me nor dependent upon my mind ... . (AT, VII, 64)

We might say, therefore, that Descartes has something like a direct reference theory of ideas: ideas represent directly the objects with which they are in some sense identical. But does the fact that Descartes on my reading holds that ideas represent objects as this or that require an inference to draw conclusions about the degree to which ideas correspond to reality? Cartesian ideas are clearly not the same as the ‘intuitive cognitions’ of Scotus and Ockham: they require some kind of analysis or test (for clarity and distinctness) for us to be certain that they correspond to

Objective Being in Descartes

151

some thing which actually exists. It is true that Descartes’s method requires that we examine ideas for clarity and distinctness to determine whether they represent some being. Is it also part of our reasoning about the correspondence of ideas that we consider the truth of the causal principle? It is important to distinguish what in Descartes’s epistemology is part of the theory of why we know what we know and what is part of the process of knowing what we know. The clarity and distinctness of an idea explains why we know what we know but entails nothing about whether what we know directly is some idea or some external object. When Descartes describes the process by which we come to assent to clear and distinct perceptions, there is no evidence of a gap between the perception and the assent to be bridged by an inference: And even if this is proved by no reason, it is impressed upon the minds of all by nature that whenever we perceive something clearly, we assent to it willingly, and in no way are able to doubt but that it is true. (AT, VIIIA, 21)

Descartes’s clear and distinct ideas are thus in one way like Ockham’s intuitive cognitions: no further inference or reasoning is required to know that the idea corresponds to something. The causal principle is also part of the theory of why we know what we know. When Descartes rejects in the Second Replies that one knows that the sky exists because one ‘sees’ it and offers instead the justification that one knows the sky exists because seeing it affects the mind in such a way as to produce an idea which ‘must have a really existing cause of its objective reality’ and ‘thus we judge that the cause is the sky itself’ (AT, VII, 165) he should not be read here as contrasting direct and representational realism and defending the latter over the former. The sky existing in the intellect is the very sky that exists in the heavens so what one knows when one knows one’s idea just is the sky itself. The causal principle is required as part of the explanation of how one’s knowledge claims can be justified and is not proposed as a intermediary in the cognitive relationship between the knower and the known. We should, therefore, be wary of concluding that Descartes’s speaking of ideas as the basis for judgement commits him to anything like our third tenet of representational realism. Being the basis for judgement does not mean that ideas are the basis for any inference from indubitably given internal objects to the existence of external objects and their properties. To return to our earlier question, it should also be clear that there is, therefore, no intrinsic connection between representationalism as a theory of mind and representational realism. If I am right about Descartes’s account of objective being, the tendency to regard Descartes as a representational realist thus represents a significant misunderstanding of his theory of ideas as well as his epistemology.

152

Representation and Objects of Thought in Medieval Philosophy

Bibliography Alanen, L. (1990), ‘Cartesian Ideas and Intentionality’, in L. Haaparanta, M. Kusch, I. Niiniluoto (eds), Language, Knowledge and Intentionality, Helsinki: Acta Philosophica Fennica, 344–69. Ashworth, E.J. (1972), ‘Descartes’s Theory of Clear and Distinct Ideas’, in R.J. Butler (ed.), Cartesian Studies, Oxford: Blackwell. Aureol, Peter, Scriptum in I Sentarium, Vatican, Biblioteca Apostolica, Borghese. Beck, J.L. (1952), The Method of Descartes: A Study of the Regulae, Oxford: Clarendon Press. Borges, Jorge Luis, ‘Dreamtigers’, in Collected Fictions, trans. Andrew Hurley, New York: Penguin, 1998. Boyce Gibson, A. (1932), The Philosophy of Descartes, London: Methuen. Brentano, Franz, Psychology From An Empirical Standpoint, ed., Oscar Kraus; trans. A.C. Rancurello, D.B. Terrell and L.L. McAlister, London: Routledge, 1874/1973. Brown, D. (2000), ‘Immanence and Individuation: Brentano and the Scholastics on Knowledge of Singulars’, The Monist, 83.1, 22–46. Brown, D. and Normore, C. (2002), ‘Traces of the Body: Descartes on Passions of the Soul’, in Byron Williston (ed.), Passion and Virtue in Descartes, Prometheus. Cook, M. (1987), ‘Descartes’s Alleged Representationalism’, History of Philosophy Quarterly, 4, 179–95. Descartes, René, Oeuvres de Descartes, vols I–XII, (eds) Charles Adam and Paul Tannery, Paris: Leopold Cerf, 1897–1913 [AT]. ———, The Philosophical Writings of Descartes, vols I & II, trans. John Cottingham, Robert Stoothoff and Dugald Murdoch, Cambridge: Cambridge University Press, 1985 [CSM]. Gal, G. (1967), ‘Gualteri de Chatton et Guillelmi de Ockham Controversia de natura conceptus universalis’, Franciscan Studies, 27, 191–212. Hoffman, P. (1996), ‘Descartes on Misrepresentation’, Journal of the History of Philosophy, 34, 357–81. ——— (2002), ‘Direct Realism, Intentionality and the Objective Being of Ideas’, Pacific Philosophical Quarterly, 83, 163–79. Kemp Smith, N. (1952), New Studies in the Philosophy of Descartes, London: Macmillan. Lennon, T. (1974), ‘The Inherence Pattern and Descartes’s Idea’, Journal of the History of Philosophy, 12, 43–52. Nadler, S. (1998), Arnauld and the Cartesian Philosophy of Ideas, Manchester: Manchester University Press. Normore, C. (1986), ‘Meaning and Objective Being: Descartes and His Sources’, in Amelie Oksenberg Rorty (ed.), Essays on Descartes’s Meditations, Berkeley: University of California Press, 223–41. O’Neil, B.E. (1974), Epistemological Direct Realism in Descartes’s Philosophy, Albuquerque: University of New Mexico Press.

Objective Being in Descartes

153

Read, S. (1977), ‘The Objective Being of Ockham’s Ficta’, Philosophical Quarterly, 27, 14–31. Stroud, B. (1984), The Significance of Philosophical Scepticism, Oxford: Clarendon. Suarez, Francisco, ‘Disputationes metaphysicae’, Opera Omnia, XXV, 908, Georg Olms Verlagsbuchhandlung Hildesheim, 1965. Tachau, K. (1988), Vision and Certitude in The Age of Ockham, Leiden: Brill. Thomas Aquinas, Summa theologiae, Cambridge: Blackfriars ; New York: McGrawHill, 1964–81 [ST]. William of Ockham, Ordinatio, in Guillelmi de Ockham Opera philosophica et theologica ad fidem codicum manuscriptorum edita. Opera Theologica. Scriptum in Librum Primum Sententiarum I–IV, eds G. Gál, S. Brown, G. Etzkorn, and F. Kelly, St Bonaventure, NY, 1967–79. Wilson, M.D. (1978), Descartes, London: Routledge & Kegan Paul. Yolton, J. (1975), ‘Ideas and Knowledge in Seventeenth Century Philosophy’, Journal of the History of Philosophy, 13, 145–66.

This page intentionally left blank

Index of Names

Abelard, Pierre 6, 14, 25–6, 74 Alanen, L. 1, 135 Albert of Saxony 27 Albert the Great 7, 69–71, 126 d’Alès, A. 16 Alexander of Aphrodisias 7, 64, 66, 117, 122, 129 Al-farabi 3, 11 Alhazen 125–6, 139 Anselm of Canterbury 6, 18, 26 Aquinas, Thomas 4, 6–7, 24, 33–59, 69–72, 84–6, 91, 123, 130, 138, 142–3 Arnauld, Antoine 120–21, 146–7 Ariew, R. 3 Aristotle 2–3, 5, 7, 11, 13, 15, 33, 38, 63–4, 67–9, 81, 84, 86, 117–18, 121, 124, 129–30 Augustine 3, 17 Aureol, Peter 74, 87, 139–40 Averroes 2, 7, 67, 127–8 Avicenna 2–3, 6–7, 11, 19–25, 29–30, 66–7, 70–71, 125–6 Bacon, Roger 92, 126–8, 139 Berkeley, George 43 Blund, John 19, 24 Boethius 17 Bonaventure 82 Borges, J.L. 135, 141 Brentano, Franz 2–3, 11, 141 Brown, D. 5, 8, 132 Burgundio of Pisa 19 Buridan, John 21, 107, 132 Caston, V. 3 Caterus 120–21, 141 Chatton, Walter 87, 140 Chisholm, R. 119 Chomsky, N. 8, 101–3, 105, 113, 117–18 Cranz, F.E. 2 Crathorn, William 89–90

Descartes, René 1, 3–5, 8, 12, 98, 101–3, 113, 120–21, 132, 135–51 Duns Scotus, John 7, 27, 34–5, 73, 75, 77–8, 85, 87 Durandus of St Pourcain 4, 71 Fodor, J. 82 Frege, Gottlob 146–7 Garlandus Compotista 25–6 Gassendi, Pierre 147 Geach, P. 43–4, 105 Gerard of Cremona 20 Giacomodi di Ascoli 87 Gibson, A.B. 142 Gundisalvi of Gudussalinas 20 Gyckye, K. 126 Haugeland, J. 82 Henry of Ghent 33–4, 83 Hervaeus Natalis 87 Hobbes, Thomas 98 Hoffman, P. 137–8, 142 Holcot, Robert 89 Ibn Gabriol 117 Isidore of Seville 57 James of Venice 17 Kenny, A. 4–5 King, P. 7 Lagerlund, H. 3–4, 6 Lindberg, D. 126 Locke, John 3, 98, 117 Loux, M. 104–7 Mair, John 21 Matthew of Acquasparta 83 Nemesius of Emesa 19

156

Representation and Objects of Thought in Medieval Philosophy

Normore, C. 3–4, 8, 135, 144–5 O’Neil, B. 135 Olivi, Peter John 72, 130–32 Oresme, Nicole 127 Pasnau, R. 7, 71, 89 Peter of Ailly 4, 27–9 Peter of Spain 57 Philo of Alexandria 131 Philoponus, John 65–6, 71 Plato 3, 14, 29, 55, 109, 131, 143 Putnam, H. 119 Quintilian 6, 14–16 Regius, Henricus 144 Richard Rufus of Cornwall 3, 82, 128–9, 131 Rorty, R. 4–5 Ryle, G. 4

Sellars, W. 119 Sextus Empericus 19 Stroud, B. 4–5, 136 Suarez, Francisco 141 Tachau, K. 126 Tertullian 16, 19 Themistius 7, 64–6, 122–3, 129 Trentman, J. 105 Tweedale, M. 7 William of Alnwick 7, 75–8 William of Ockham 8, 27, 72, 74, 87, 93–8, 101–13, 140–41, 150 William of Sherwood 57 Wilson, M. 144–7 Wittgenstein, L. 4 Wodeham, Adam 111 Wood, R. 3 Yrjönsuuri, M. 8