Ockham on Concepts ASHGATE ASHGATE STUDIES IN MEDIEVAL PHILOSOPHY. CLAUDE PANACCIO University ofquebec in Montreal, Canada

Transcription

1 ASHGATE STUDIES IN MEDIEVAL PHILOSOPHY Ockham on Concepts Series Editors John Marenbon, Trinity College, Cambridge, UK Scott MacDonald, Cornell University, USA Christopher J. Martin, University ofauckland, New Zealand Simo Knuuttila, Academy offinland and the University ofhelsinki, Finland The study ofmedieval 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 of1teir 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. CLAUDE PANACCIO University ofquebec in Montreal, Canada Other titles in the series: Theology at Paris, Peter Auriol and the Problem of Divine Foreknowledge and Future Contingents Chris Schabel ISBN Medieval Modal Systems Problems and Concepts Paul Thom ISBN ASHGATE

2 .. 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. Claude Panaccio 2004 The author has asserted his moral right under the Copyright, Designs and Patents Act, 1988, to be identified as the author of this work. Published by Ashgate Publishing Limited Ashgate Publishing Company Gower House Suite 420 Croft Road, Aldershot 101 Cherry Street Hampshire GUll 3HR Burlington, VT England USA IAshgate website: I British Library Cataloguing in Publication Data Panaccio, Claude, Ockham on concepts. - (Ashgate studies in medieval philosophy) 1. William of Ockham, ca ca Concepts 1. Title Library of Congress Cataloging-in-Publication Data Panaccio, Claude, Ockham on concepts / Claude Panaccio. p. cm. - (Ashgate studies in medieval philosophy) Includes bibliographical references and index. ISBN (alk. paper) 1. William, of Ockham, ca ca Philosophy, Medieval. 3. Concepts. 1. Title. II. Series. B P '.4-dc22 ISBN Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire Typeset by DC Graphic Design Ltd, Swanley, Kent Contents Acknowledgements Abbreviations Introduction Notes 1: Intuition, Abstraction and Mental Language 1..Intuitive and abstractive cognitions 2. Mental language 3. Abstraction and universals 4. Intuition and singular terms 5. Mixed cognitions and singular terms of the third type 6. Ockham's representationalism Notes 2: Intellectual Acts 1. Mental acts and habitus 2. From the fictum to the actus 3. The rejection of intelligible species 3.1 Species as preconditions for intellectual acts 3.2 The razor argument against intelligible species 4. Combining acts 4.1 Simple and complex units 4.2 Propositional acts 4.3 Iudicative acts Notes ix xi

3 ,.-. vi vii 3: Concepts as Signs 1. The problem: how can a concept ever be a sign? 2. The two meanings of signum 3. Conceptual roles 4. Atomism or propositionalism? 5. Types and tokens : Concepts as Similitudes 1. Similitude sustained 2. Acts and similarities 3. Varieties of conceptual representation 3.1 Absolute specific quality concepts 3.2 Specific concepts of substances Notes 4: Connotative Terms in Mental Language 1. Connotative terms 2. Mental connotation 3. Synonymy and nominal definitions 4. Connotative terms and exponible propositions Appendix: A reply to Richard Gaskin Notes 5: The Role of Nominal Definitions 1. Four theses about nominal definitions 2. What defining amounts to: a reconstruction 3. Some consequences 3.1 Definitions and abbreviations 3.2 Possession of concepts and knowledge of definitions Simple connotative concepts 3.4 Simple generic concepts 4. Two problems about absolute concepts 4.1 Klima's objection 4.2 Brown's puzzle Notes H: Logical Concepts I. The earlier theory: logical words internalized 2. Logical constants in the actus-theory 3. Prepositions and non-standard copulas Notes l): The Meaning of Words I. Subordination 2. Types and tokens again Real orderings Notes 6: Cognition and Connotation 1. Spade's questions 2. The acquisition of simple connotative concepts 3. The adequacy of nominal definitions Reverse subordination? The instructive case of proper names Notes Conclusion Notes Bihliography Index Ockham and the Classical View 113 Notes 116

4 Acknowledgements Most of this book was written between the winter of 2002 and the fall of 2003, as I was released from my regular teaching load; first by a sabbatical leave from the University of Quebec in Trois-Riyieres, and then by a Killam Scholarship from the Canadian. Council for the Arts. The Social Sciences and Humanities Research Council of Canada also greatly contributed to the fulfilmel)t of the project by awarding it a three-year research grant from 2000 to I wish to express my sincere gratitude to all three organizations for their generous support. It should be mentioned that some of the developments to be read here partly correspond to previously published papers of mine: Chapter 1 is closely based on an article originally written in French (,Intuition, abstractionet langage mental dans la theorie occamiste de la connaissance', Revue de Metaphysique et de Morale, 97/1,1992, pp ). A preliminary version of chapter 4 has circulated for quite some time among specialists in an informal 'prepublication' format ('Connotative terms in Ockham's mental language', Cahiers d' epistemologie, no. 9016, Montreal, Universite du Quebec amontreal, 1990, p. 21), and a slightly different one later appeared in French 'in an Italian journal ('Guillaume d'ockham, les connotatifs et Ie langage mental', Documenti e studi sulla tradizionejilosojka medievale, 11,2000, pp ). Chapter 5 freely makes use ofideas about nominal definitions that I originally presented at the XIIIth European Symposium for Medieval Logic and Semantics held in Avignon in June 2000, in a contribution that was subsequently published in the acts ('Connotative concepts and their definitions in Ockham's nominalism', in La tradition medievale des Categories, Joel Biard and Irene Rosier-Catach, eds, Leuwen: Peeters, 2003, pp ). Section 1 of chapter 9 is a revised version of a recent paper in French (,Guillaume d'ockham et les syncategoremes mentaux: la premiere theorie', Histoire, Epistemologie, Langage, 25/2, 2003, pp ). I am grateful to the editoj;s and publishers involved for their role in the original dissemination of this material. The fact is that this book found its motivation in the extraordinarily stimulating - if sometimes critical - reactions these publications, especially the first two ones, elicited, both in print and in private conversations, from a number of outstanding scholars whose names will be found again and again in the following pages. Letme single out at this pointthose ofjoelbiard, ElizabethKarger, Calvin Normore and Paul Vincent Spade in particular, to whom this whole work, while disagreeing with them on certain points (and sometimes crucial ones), remains in the end deeply indebted, Several portions of the nook have neen orally presented to various audiences in France. Italy. Greece. Ukraine. the US and Canada. and have greatly nenefited from

5 x 'the numerous remarks, questions and objections that were raised on these occasions by colleagues and students. Chapters 3, 5, 6 and 7, most saliently, are elaborations on the Pierre-Abelard Lectures I gave at the University ofparis 4-Sorbonne in March Special thanks are due, in this respect, to Cyrille Michon, who not only organized these lectures, but who also provided useful and penetrating commentaries on each one of them when they were delivered. Most of all, finally, I should like to thank Cecile Juneau for her remarkably patient and professional secretarial work upon the whole manuscript over the years, and my beloved companion Claude-Elizabeth Perreault for her constant and affectionate support day after day whilei was - often single-mindedly! - working at it. Exp. in Perih. Exp. in Phys. Exp. in Porph. Exp. in Praedic. Exp. sup. Elench. OPh Ord. OTh Quaest. in Phys. Quaest. Var. Quodl. Rep. SL Summ. Phil. Nat. Tract. de Corp. Chr. Tract. de Quant. Abbreviations Expositio in Librum Perihermenias Aristotelis Expositio in Libros Physicorum Aristotelis Expositio in Librum Porphyrii de Praedicabilibus Expositio in Librum Praedicamentorum Aristotelis Expositio super Libros Elenchorum Guillelmi de Ockham Opera Philosophica Ordinatio. Scriptum in Librum Primum Sententiarum Guillelmi de Ockham Opera Theologica Quaestiones in Libros Physicorum Aristotelis Quaestiones Variae Quodlibeta Septem Reportatio. Quaestiones in Libros II, III, IV Sententiarum Summa Logicae Summula Philosophiae Naturalis Tractatus de Corpore Christi Tractatus de Quantitate

6 Introduction This book is an exercise in interpretation. My background conviction is that William of Ockham's nominalism, even if elaborated in the d,istant context of the early fourteenth century, still provides a challenging and fruitful body of theory to be in dialogue with today. In philosophy of mind and language, in particular, Ockham's obstinate refusal to countenance anything but concrete singular beings in the real world, and the way he knits together on this basis an intriguing array of theses and arguments about mental language, intentionality, and reference, could have, I suspect, a healthy counterbalancing effect against the prevailing fondness in these fields for abstract entities of all sorts: general properties, natural kinds, linguistic types, Fregean propositions, and what not... But surely, the prerequisite for these appealing virtues to be actualized is that the theory be well understood! And there still exists, lam afraid, wide disagreements among specialists as to what exactly Ockham's nominalistic programme amounts to, especially with regards to cognition. This is what I want to address here. The focus will be on concepts - conceptus in Ockham's vocabulary :- taken by him as the basic units of mental representation. Several discussions have been going on around Ockham's theory of concepts in the last fifteen years or so, in the aftermath of the remarkable critical edition of his philosophical and theological writings completed at S1. Bonaventure University in the late 1980s, and of the publication in 1987 of Marilyn Adams's landmark synthetical study, William Ockham. My aim in the present work 'is to propose a thorough presentation and defence of how I understand Ockham's positions on the matter in the light of these recent developments. The first three chapters will deal with the basics of Ockham's approach: that concepts are the ultimate components of the language of thought, and that they are normally acquired as a result of natural causal processes triggered by direct empirical encounters with real individuals (chapter 1); that concepts are to be identified with intellectual acts, rather than with purely intentional objects (chapter 2); and that they are signs in the logical sense, and canbe studied, consequently, with the technical apparatus of grammar and logic (chapter 3). The next three chapters will be devoted to one salient and far-reaching debate in recent Ockhamistic studies, in which the very viability of the theory turned out to be at stake: is our whole stock of concepts supposed to be constructed by the mind out of a limited number of simple non-relational natural kind concepts, by assembling definittons? In other words: how reductivistic is Ockham's doctrine ofconcepts, with respectin particular to relational terms? It will be argued that contrary to a widespread opinion, Ockham l'un, and does, accept whut he calls 'connotative' terms - relational ones, especially - umong the basic units of the language of thought (chapter 4). His conception of what definitions arc expected to accomplish in such cases will he made explicit (chapter 5); how such concepts are supposed to he acquired will he explained (chapter 6). In the last chapters, finally, I will discuss three more questions that come

7 2 INTRODUCTION 3 out as crucial for Ockham's theory of concepts: in what sense could he subscribe to the traditional saying thatconcepts are similitudes ofthings (chapter7)? How are the logical constants implemented in the mind (chapter 8)? And how do spoken and written words receive their meanings (chapter 9)? The emphasis all along will be on currently controversial issues of interpretation, and the developments will be supported by careful analysis of numerous relevant passages and detailed discussions of alternative readings, with some assessment, here and there, of criticisms addressed to Ockham, in so far, at least, as they rest on what I take to be misunderstandings. My hope, in so doing, is that the book be useful to medievalists, of course, but also to philosophers at large and cognitive scientists as well, as an in-depth exploration of one of the truly great theories of intellectual representation in the history of Western thought. It so happens that the main points over which disagreements emerged in the Ockham literature of the last couple of decades have to do, ultimately, with some of the most central - and most difficult issues in the current hectic search for a satisfactory theory of the mind: what are the basic components of human thought? And how can they be representations of anything? I will not engage here in a direct philosophical discussion of Ockham's doctrine on these points, or in a systematiccomparison ofit with today's approaches; but my working assumption is that a precise and accurate understanding of this wide-ranging theory has now come to be highly desirable, and that the best way of achieving just that at this point is by critically reviewing the detailed arguments of recent Ockham scholarship. In accordance with the usual requirements in the field, the account I will propose heavily rests on text analysis, and many of the relevant passages will be quoted and scrutinized in the exposition process. I I know of no better way to proceed in such affairs: historical monographs in philosophy are, first and foremost, a matter of reading texts and understanding them. With Ockham, especially, we are lucky enough to be dealing with such a rigorous writer that most of the current interpretative disagreements about his theory of concepts can be settled, I contend, on the basis of what he actually wrote. In many cases, admittedly, the reconstruction needs to be developed beyond what Ockham himself cared to, but I tried to stick, in so doing, to what he is committed to, given what he explicitly holds; such elaborations, anyway, will be clearly indicated when they occur. The theory that will emerge is surprisingly rich and unified, but quite different on the whole from what it has often been taken to be. The prevailing dogma in the last decades was thatockham's approach is a canonical - and especially stringent variant of the so-called 'Classical' (or 'Definitional') view of concepts, according to which most of our intellectual representations are in fact complex constructs made out of a limited number of primitives; and that the mental language he postulates should be structured like a logically ideal language a la Frege-Russell. One of my main points in the book will be that this standard reading is seriously misleading on many counts. What Ockham proposes. will turn out to be closer in the end to what we call today an 'atomistic' theory ofconcepts, as characterized, for example, by the American philosopher Eric Margolis: a theory, namely, 'according to whiich what makes a concept the very concept that it is is not how it is related to certain other concepts but how it is related to the world.'2 However general they can get. simple concepts, in Ockham 's approach. ultimately depend. for what they represent, on their particular causal links with concrete episodes of direct acquaintance; this holds for (at least some) relational concepts as well as non-relational ones. Many striking similarities with recent ideas in analytic philosophy of mind and language will become more and more,apparent, as we go on, to those who are familiar with the writings of people like Saul Kripke, Hilary Putnam, or - most of all - Jerry Fodor, that is, a resolute commitment to the language of thought hypothesis and to semantical atomism, a prominent concern for direct reference, a causal approach to epistemology, a strong innatist component, and even some externalistic tendencies. Yet the really distinctive feature of Ockham's theory of concepts, as will be stressed all along, is how it manages to harmonize all these interesting trends within the rigid constraints of a radically nominalistic ontology. There lies in the end its deepest lesson. Notes I. English translations will always be provided for quotes occurring within the main body of the text, with the Latin original versions to be found in the footnotes. Unless otherwise indicated, the translations are mine. 2. Margolis 1998, p See also Fodor 1998.

8 Chapter 1 Intuition, Abstraction and Mental Language On the matter of concepts, Ockham, it is striking, has two distinct terminologies. On the one hand, he regularly resorts to the vocabulary ofcognitio or notitia, and to a distinction he adapts from John DunsScotus between cognitio intuitiva and cognitio abstractiva. On the other hand, he takes more seriously than anybody before him the idea that conceptual thought is a sort of inner.. discourse, an oratio mentalis, to which he systematically applies the semantical vocabulary of terminist logic (significatio, suppositio, connotatio, etc.), which was in use since the twelfth century for the analysis of spoken or written languages. How.are these two apparatuses linked to each other? This is, curiously enough, a point about which there is 110 v~ry elaborate discussion in Ockham's own texts. Most of the passages which make use of one of the terminologies either entirely ignore the other, or give it but a secondary place; those in which both can be found do not organize them in a very explicit theory. Yet this is something we need to be clear about if we want to understand Ockham's approach to concepts with any precision, since concepts, in his most mature theory, are simultaneously identified both with cognitive acts l and with mental signs susceptible of various referential roles in propositional contexts. 2 Chapters 2 and 3 will come back in more detail to these ideas of cognitive acts and mental signs respectively, but it will. be useful to elucidate first, in the present chapter, how the Ockhamistic doctrine can coherently equate the intuitive and abstractive acts of cognition with significant terms, capable of occurring, in person so to speak, within mental propositions. For the reader unfamiliar with the field this development will serve as a general introduction toockham's epistemology of ~oncepts, and the first two sections of the chapter will be devoted to short, non-controversial presentations of the doctrine of intuitive and abstractive cognitions (section 1) and of thedoctrine of mental language (section 2). The main goal of the chapter, however, will be to establish a number of non-trivial points about how Ockham's theory of concepts maps the cognitive states corresponding to intellectual intuitions and abstractions into logical or semantical categories (sections 3:---5). An interesting upshot of all this will be that, contrary to what several recent commentators haye suggested, (kkham's theory of concepts can legitimately be labelled as a brand of rcprcsentationalism (section 6). I. Intuitive and abstractive cognitions Ockbam's theory of intellectual acts is expounded mainly in the Prologue of his On/inatio and. in a much shorter version. in some ofhis Quodliht,ftll Qucstions.' The ~

9 6 INTUITION, ABSTRACTION AND MENTAL LANGUAGE 7 precise notion of an act which is at work there obviously needs further scrutiny and we will come back to it in chapter 2, but let us be content at this point with a preliminary understanding of mental acts, in Ockham's sense, as being those mental states that are actualized whenever an agent is actually thinking, feeling, willing or perceiving something. This is basically what contemporary philosophers of mind call 'mental episodes'. Some mental acts in this sense are acts of the will, some are acts of desire, some are perceptual acts, and so on. The ones we are interested in are the intellectual acts. Ockham distinguishes quite a number of them. The.act of judgement (actus iudicativus), by which an agent judges a certain proposition to be true or false, is contrasted with the act ofapprehension (actus apprehensivus), which is a mere intellectual grasping. The latter in turn can be either complex or incomplex. The most salient example of a complex apprehensive act is the formation of a mental proposition, on the truth value of which the agent does not yet commit himself, but which is usually presupposed by the judgement. The incomplex apprehensive act, on the other hand - which always precedes the complex apprehension and makes it possible - further subdivides into intuitive cognition and abstractive cognition; this is the distinction that will now detain us. The intuitive intellectual cognition of a thing, William explains, is this simple awareness of that thing in virtue of which the intellect can judge as evident that the thing exists, if it exists, orthat it does not exist, if indeed it does not exist. 4 More generally, an intuitive cognition of something is this awareness of that thing which, under normal circumstances, causes the agent to form true and evident presenttensed contingent judgements about that very thing: my intuitive knowledge of Mary when Imeet her, for example, makes me judge as obvious that Mary presently exists, that she is brown-haired, that she is sitting in front of me, that she is talking, and so on, none of which I could evidently know without such an immediate grasping of her. Abstractive knowledge, by contrast, does nothing of the sort: 'through abstractive cognition, Ockham writes, no contingent truth, especially in the present tense, can be evidently known.'5 Abstractive cognition is that in virtue of which it cannot evidently be known, in particular, whether the thing presently exists or not. Thinking of Mary in her absence does not normally allow me to evidently know whether she is still alive or not at this very moment, whether she is sitting or not, talking or not, and so on: this is a typical case of abstractive cognition. The situation is pretty much similar when I use a general concept to think about chickadees, let's say, or daisies, when none of them are actually there before me. Abstraction, in this sense, as Ockham makes it clear, should - at first - be understood as abstraction from existence or non-existence, rather than from singularity or non-singularity. 6 Inthe natural orderofthings, intuitive cognition always corresponds to the simple and immediate grasping of something which is actually present to the agent. If the thing is not there in front of me, my cognition of it can only be abstractive. That Ockham does not simply express himself in this way is due to the fact that he is anxious, for theological reasons, to preserve God's omnipotency, which he sees as limited only by the principle ofnon-contradiction - a merely logical rather than real limitation - and that, consequently, he wants to allow for the supernatural possibility that God, through some special intervention, should provide me with a peculiar cognition of something, in virtue of which I would evidently know that this very thing docs not presently exist. This is the famous intuition of non-existents. which has been so widely discussed among modern commentators.? Such a cognition would not abstract from existence and non-existence and would indeed cause an evident and true judgement about contingent matters; it would not, therefore, be an abstractive apprehension, in Ockham's sense, but an intuitive one. Since a miraculous intervention of this sort cannot be ruled out as intrinsically contradictory, the general characterizations of intuitive and abstractive cognitions must take it into account. Under purely natural conditions, however, such occurrences never happen and I will therefore ignore them in the following discussion, confining myself in practice to the cases where intuitive cognition is concomitant with the real presence of the intuited singular thing. 8 It must be stressed that intuitive and abstractive cognitions, as I have just characterized them, are both of them acts of the intellect. Ockham rejects the division of functions that was commonly accepted - especially by Thomists between sensation, seen as the sole mode of direct acquaintance with singular objects in this terrestrial life, and intellection, seen as monopolized in principle by generality. He does adm.it, of course, that sense perception never reaches but individual things, and he is quite happy to allow for a sensory variety of intuitive cognition. 9 His point, however, is that the intellect too can reach the individual thing in its concrete presence, and that this is precisely what happens in the intellectual intuitive cognition. lo Such an intellectual intuition is required in addition to the sensory one, in Ockham's view, in order to account for the formation of evident existential judgements, the argument for this being the following: judgements are intellectual acts and as such belong within the intellect; but they presuppose, both as their parts and as their partial cause:s, incomplex cognitions belonging to the same 'cognitive subject' as themselves. The relevant simple apprehensions capable of causing evident existential judgements, then, must also belong to the intellectual part of the mind, and not merely to the senses. I I Ockham's position here partially hangs upon his (antithomistic) doctrine ofthe plurality ofsubstantial forms: a human being for him - as for most of his Franciscan predecessors - is a hylemorphic compound of a given piece of matter with a number of different substantial forms, including at least a corporeal form, a sensitive form, and an intellective form, the last two being seen as two different 'cognitive subjects', or distinct compartments of a given mind, if one prefers. '2 Ockham's argument for intellectual intuition, however, also - and even more crucially - rests on the view that the incomplex cognition which causes an evident existential judgement must itself be a part of that very judgement. We will come back to this in some detail in section 4 below. 13 What we have at this point, to sum up, is that the overall cognitive process, as (kkham describes it, takes the form of a rather complex causal sequence. The external sensible object first causes a sensory intuition in the agent, and then, with the help of this sensation, also causes an intellectual intuitive act, which itself brings ahout (under normal circumstances) two things within the intellect: a singular and evident true existential (or at least contingent) judgement on the one hand, and a simple ahstractive act on the other hand. This latter abstractive cognition, in turn, causes the formation of what Ockham cal1s a habitus - a disposition, if one prefers - which wil1 later on enahle the agent to reactivate the abstractive act even when the ohject is not present anymore. and which thus constitutes a brand of intel1ectual memory.14 General propositions can then he formed. and science can llourish. The

10 8 INTUITION, ABSTRACTION AND MENTAL LANGUAGE 9 questioni want to raise here is the following: how and where exactly do mental signs fit into this process? 2.~entallanguage The idea of an oratio mentalis - or oratio concepta - composed of signs is another major theme in Ockham's philosophy of mind, and it is also closely associated with his views on concepts. The Summa Logicae, most notably, gives it a prominent role, but the subject is frequent in Ockham's other philosophical and theological writings as well. IS Picking up a famous distinction from Boethius, William explains in the very first chapter of the Summa that there are three sorts.of discourses,,and correspondingly three sorts ofterms: written, spoken and mental.16 The written ones are perceptible to the eye, and the spoken ones to the ear, but the mental ones are hidden to public perception. Concepts, which are the basic units of this mental language, are nowhere but in the private minds of the intelligent agents, even if they are quite similar in each of them, as Aristotle had noted in chapter 1 of the De Interpretatione. These concepts, just like the words of spoken and written languages,.are. signs, but their signification is natural and primitive with respect to that of spoken and written words, which is conventional and derivative - or 'subordinated' (subordinata). This conceptual discourse, which is equated with thought itself, is endowed with a syntactical structure, very much like that of the conventional external languages. Most of the standard grammatical categories of Priscian or Donatus are borrowed by Ockham for the analysis of inner discourse. There, he claims, we have nouns, verbs, adverbs, conjunctions and prepositions; we have singular or plural phrases, casedeclensions for nouns, tenses and modes for verbs, and so on ---' in short everything which is 'necessary for signification' (propter necessitatem significationis).'7 Even more importantly, the so-called terminist logic, which had been developed in the twelfth and thirteenth centuries, is now being applied not only to spoken or written languages, as in Peter of Spain, William of Sherwood, and the other early contributors to the logica modernorum,18 but also - and even primarily - to mental terms and propositions. Concepts are said to have a signification (sign(ficatio) or a connotation (connotatio) when they are considered in themselves,.and a referential function (or suppositio) when they occur within propositional contexts. It is one of Ockham's major innovations to have systematically transposed the terminist logic he had learned at sc]lool into a theory of discursive thought. 19 With respect to the mode of existence of conceptual signs, it is now well known that Ockham has changed his mind in the course of his career. 20 In the first redaction of his Sentence Commentary, around , he hesitantly subscribed to the socalled fictum-theory, according to which a concept in the mind is a purely ideal object, which has no reality other than that of being the intentional correlate of a simple cognitive act of apprehension. 21 After a short period of oscillation - probably around he finally adopted in his Questions on Aristotle's Physics, in the Quodlibetal Questions and in the Summa Lo~icae. a very different view, known as the actus-theory, according to which concepts, rather than oeing the oojects of certain cognitive acts, are directly identified with these very acts themselves. Ockham realized at some point - and this is a very deep intuition in his mature system - that the various functions he wanted to attribute to concepts, in.so far precisely as they were semantical functions, could adequately be fulfilled by the cognitive acts themselves: 'for an act of understanding can. signify something and can supposit for something just as well as any sign; and there is no point therefore in postulating anything over and above the act of understanding'.22 Such an act can be classified, within the framework of an Aristotelian ontology, as a quality of the mind. Reducing the concept to it allows Ockham's nominalism to dispense with the special mode of existence that was required to accommodate the fictum, and to stick, in psychology as well as in physics, to a very simple naturalistic ontology with nothing but singular substances and qualities. Here is a salient case where Ockham explicitly invoked the famous Razor Principle Abstraction and universals Restricting ourselves to Ockham'slater theory, weare now in a position to reformulate our original question as having to do only with mental acts: what, we are asking, is the relation between on the. one hand those mental acts which Ockham views as. conceptual signs and Which can occur as parts of mental propositions and have asuppositio, and on the other handthe intuitive and abstractive acts that were described in section 1 above? Let us begin with the case of abstraction. The answer here is very simple. It holds in the following identification: all simple abstractive acts of cognition are general terms in the mental language. The mental state of abstraction, in other words, is itself a categorematic conceptual term, capable in principle of being the subject or predicate of a mental proposition, and of standing for a plurality of distinct singular entities. It is, literally, a mental common noun. Concepts, which are often described by Ockham as the basic units of mental language,24 are indeed directly identified in several passages with abstractive acts.2 s We don't have, as in many medieval or modem epistemologies,two distinct entities here, one of which would be. the result or the object of the other, but only one. Once the,fictum-theory of concepts is abandoned in favour of the actus-theory, there is no point, in Ockham's eyes, in distinguishing between the concept and the abstractive act anymore. Why, however, should we say that simple abstractive acts are all general terms? Ockham, after all, distinguished very clearly in the Prologue of the Ordinatia between two senses of notitiaabstractiva: abstraction from singularity, and abstraction from existence or non-existence;26 it was; he said, in the second ofthese senses, not in the first, that he wanted to contrast abstractive cognition with intuitive cognition. Isn't it possible, we may ask, to think about somebody in her absence? And isn't such a cognition, then, both singular and abstractive in Ockham's sense? Thinking about Mary while she is not here does not allow me to evidently know whether she still exists or not, whether she is sitting or not, and so on; this is precisely what Ockham calls an abstractive cognition, isn't it? Well, Ockham's position with regard to singular ahstractive cognition varied across his writings, but his final conclusion, in the Quol/liheral QUl'stiollS in particular, is that we cannot

11 10 INTUITION, ABSTRACTION AND MENTAL LANGUAGE 11 have in this life an abstractive cognition which is both simple and properly singular. The reason for this is that a concept for Ockhamis always an intellectual image of some sort, a similitudo of what it represents. But what represents anything in virtue of being an image of it also represents to the same degree whatever is exactly - or relevantly - similar to that thing. My mental image of Mary in her absence is to the\ same degree an image ofany otherbeing who happens to looklike herin therelevant respects. In so far as it functions as an image, this representation, then, is intrinsically general: it is equally capable in principle of representing a plurality of different entities. All simple abstractive cognitions, for Ockham, are like that: they are all of them general in principle, in so far as their representational function hangs upon some sort of similarity.27 The possibility of a properly singular abstractive cognition of something is left open, nevertheless, but it can only be reached, Ockham says, through an appropriate combination of simple terms, none of which would itself be properly singular. I can think about Mary after all, rather than about any twin she might have, but the concept I should form in order to do so will have to be a mental compound rather than a simple intellectual image. 28 Properly speaking, there are no proper names among abstractive simple concepts. This amounts to saying that a simple abstractive act is always in itselfa universal. Universals in the Porphyrian sense - genus, species, differentia, proprium and accident - were traditionally defined as what can be predicated of many. But, as Ockham frequently insists, only a sign can be predicated of anything. 29 Predication occurs only within propositions, and propositions, in Ockham's mature theory, can only be composed of signs. Nothing therefore is a universal 'except in its signification, in its signifying many things'.3o This indeed is the core of Ockham's nominalism. And the signs par excellence, those that are at the very basis of the whole semiotical process, are precisely the concepts (or 'intentions of the soul', as Ockham also calls them), which, as we have just seen, are identified with abstractive acts. Thespecies 'man', for example, is nothing extramental, it is a general concept in the mind, in other words, an abstractive cognition. The terminology of the mental sign appropriately merges at this point with that of the notitia or cognition, to provide a clear-cut answer to the old problem of universals. It follows from this that the process of formation of the universal is exactly the process that leads to abstraction, as it was described above at the end ofsection 1. In the case of species, in particular, a single encounter suffices for their original formation in the mind. Suppose I have never seen a blue jay before, and I see one now underoptimal conditions; The intellegtual intuitive act, as we have seen - which is the immediate grasping of that singular bird by my intellect - generates at once a first simple abstractive cognition, the prima abstractiva, which is straightaway a general, rather than a singular, representation,3l What will it be common to? Ockham's answer is that it will indifferently represent to my mind everything in the world which is maximally similar (simillimus) to this singular bird I am now looking at. But maximal similarity for Ockham precisely is this similarity which holds between all the singular members of a given specialissima species (a species which does not naturally subdivide into further subspecies).12 This means, in the present example, that the first simple abstractive act which will be caused in my mind by this intuitive cognition ofa blue jay. will be a general concept representing all blue jays. and nothing else. In Ockham's words: 'the concept of the species can be abstracted from a single individual'.33 Its formation does not require, as in some later empiricist philosophers, that different individuals be first compared to one another and their common features extracted. What it presupposes is an innate mental mechanism of schematic representation, capable of being triggered by a single encounter. This is not so, on the other hand, for generic concepts, which, Ockham says, are 'never abstracted from just one individual'.34 I must already have at least two concepts of distinct species in order to'form on that basis the concept of the genus which includes them both. There is, in this way, a genealogical hierarchy of abstractive acts which does not correspond to the order of logical implication. The simple generic concept of bird, for instance, logically implies no particular species concept, and it is not directly implied by any of them either, but it could not grow within a particular mind unless it was preceded in it by at least a couple of distinct lower-level specific concepts such as that of chickadee, blue jay, or sparrow. Ockham's concepts - the universals - are neither mere abbreviations for lists of proper names,35 nor structured definitions. They are, in their most elementary form, simple abstractive acts, which, in virtue of abstracting from existence and nonexistence, are directly constituted into intrinsically general natural signs, representing in the agent's mind a plurality of things, all of them singular of course but nevertheless objectively similar to one another. 36 This similarity among its significates is maximal in the case of the concept of a specialissima species, and more and more relaxed as concepts get more general. In all cases, the universal thus generated - which is all at once a concept, an abstractive cognition, and a general categorematic mental sign - can supposit within mental propositions for all or any. of these singular things - real or possible - which it signifies. 4. Intuition and singular terms What about intuitive acts? Can they also be assimilated with some sort of signs and play logical and referential roles within propositions, such as those which are described by supposition-theory? Marilyn Adams, in her William Ockham, does not hesitate to answer this question in the affirmative: 'In the case just described - where [ have an intuitive cognition of Socrates and his whiteness and thereby formulate the judgement 'Socrates is white' - the intuitive cognition of Socrates and whiteness themselves serve as the subject and predicate terms of the mental proposition.'37 Since subjects and predicates always have supposition, the intuitive intellectual act should be seen, in this view, as a semantic and syntactic unit of the language of thought: a mental sign in the fullest sense. This reading, as I will argue, is basically correct. It raises however a number of difficulties and calls for some nuances. For one thing, the predicate of 'Socrates is white' being a general term, it is hardto see how itcanbe identified with an intuitive cognition. Contrary to what Adams says, the predicate of the mental proposition corresponding to 'Socrates is white' must, in any analysis, be a general concept, and therefore an abstractive act. The interesting question, however, is: what about the subject in such cases'! Can my intuitive cognition of Socrates serve as the very subject of a mental proposition such as.socrates is white'. Socrates exists', or

12 12 INTUITION, ABSTRACTION AND MENTAL LANGUAGE 13 'Socrates is here'? And a similar query arises about the predicates of propositions such as 'this is Socrates' or 'some man is Socrates'. To these questions, Adam's answer is indeed the right one, at least for Ockham's later theory. It is true that in the original redaction of the Ordinatio, Ockham readily spoke of the 'extremes' of a mental proposition (that is, its subject and predicate) as being objects of intuitions, rather than intuitions themselves. 38 But these passages were written while he was still subscribing to thefictum-theory of concepts. He would then agree with his senior colleague Walter Burley that singular mental propositions can be said to have as their subjects- and in some cases, as their predicates - the very external things they are about. 39 At some later point, however, he corrected his own text: wherever he had alluded to the intuitive cognition of the extremes of a proposition, he would now add a corrective formula such as '... or rather, of the things signified by the extremes'.40 What these corrections indicate is thatthe subject of a mental proposition like 'Socrates exists' or 'Socrates is white', as formed by the agent in the presence of Socrates, is now seen as a sign..and what else could this sign be but the intuitive cognition itself, just as Adams says? It is not the real Mister Socrates obviously, the rejection of this possibility being the very reason for the corrections, nor an abstractive cognition, the hypothesis of which would be entirely superfluous in the context. Ockham, actually, gets explicit about that in question 7 of his Questions on the Physics ofaristotle: When the intellect apprehends a singular thing by intuition, it forms in itself an intuitive cognition, which is a cognition of this singular thing only, andis capable by its very nature to supposit for this singular thing... And just as a spoken word conventionally supposits for its significate, similarly this [intuitive] intellection naturally supposits for the thing it is an intellection of,41 Here is a case where the epistemological terminology of the cognitio intuitiva is clearly connected with the semantical vocabulary of supposition-theory. And supposition, for Ockham, belongs to significant terms inso far only as they are taken as subjects or predicates within propositions. 42 It follows that the intellectual intuitive act is an integral part of the language of thought; and that it can be subject or predicate of mental propositions, where it can display s.emantical properties such as a supposition ' This is an intriguing idea, ilnd although Ockham did not work it out in detail, it deserves a bit of elaboration. The intuitive act, forockham, is a realthing (a res absoluta), namely a quality.44 This thing is a sign, in the very sense of the word which interests logicians: 'anything which brings something to mind and can supposit for that thing'.45 Just like the abstractive cognition which wiluollow upon it, the intuitive act resembles its object in some way46, but unlike abstractive cognition, it is not thus rendered general. The reason for this is that in the case of intuitive.cognition, it is causality rather than similarity which determines the signification: 47 an intuitive intellection is a cognition - and a sign - of exactly this singular thing which (partially) causes it. Such a causal account of signification was not available to Ockham in the case of abstractive cognitions, because the only true causes for him arc the immediate ones,4~ and the causal link between an abstractive act and its external object is always indirect. Only in the cilse of intuition is lhl' external object an immediate (albeit partial) cause. In so far, then, as they are parts of mental language, intuitive acts, unlike the abstractive ones, are singular terms. What kind of singular terms? Ockham, in the Summa, distinguishes three varieties of singular terms: proper names such as 'Socrates', demonstrative pronouns such as 'this', and common nouns accompanied by demonstrative pronouns such as 'this man'.49 Let us leave the third category aside for a moment - we will come back to it in the next section - and let us ask the following:. are intuitive. cognitions, seen as signs in the mental language, proper names or demonstrative pronouns? Marilyn Adams opts; without discussion, for the former. 50 This, however, meets with an important obstacle: proper names, as we know them, can very well be uttered in the absence of their referents. Intuitive acts, by contrast, always require - at least in the natural order - the actual presence of their referents,.behaving in this more like demonstratives than like proper names. On the other hand, demonstratives in the usual sense do not readily lend themselves to the role we would like to attribute to intuitive acts. I am not referring here to the fact that Ockham leans toward excluding all pronouns, whether demonstrative or not, from the mentallanguage;51 this exclusion might not have'to be taken too seriously, since, as I have tried to show elsewhere, Ockham,in these passages, widely underestimates the importance of demonstratives for his own logic.52 What I have.in mind, rather, is that a salient feature of ordinary demonstratives, as David Kaplan has insisted, is that they have, so to say, two levels of meaning, which Kaplan respectively calls a 'character' and a 'content'.53 The 'character' of a demonstrative, in this vocabulary, is its general capacity to refer to different things indifferent contexts, or - to speak like Kaplan ~,to have different 'contents' in different contexts: 'this' can refer to any objectinthe world, according to what the speaker designates at the moment of utterance. Ockham's intuitive acts, however, do not present any such 'character'; Each one of them is inescapably attached to a particular referent. A given intuition cannot, like a demonstrative, be reused againlateron by the agent to refer to II new object. Whether I use 'this' to speak about a chickadee ora car, it stilhs the sameword. My corresponding intuitive acts, by contrast, are not, in my mental language, tokens of the same sign. What is required here is a special semantical and syntactical. category, which presents strong similarities with what Bertrand Russell used to oall 'proper names in the logical sense',54 but which is even more stringent. A logically proper name, in Russell's view, is never given but toa single referent, which the speaker must at some point have been directly acquaintedwith, this episode of direct acquaintance having fixed once and for allthe referent for this particular designator. Ockham's intuitive cognitions are direct designators too: they do not have descriptive contents, any more than Russell's logically proper names do. And they also presuppose a direct acquaintance of the agent with the object. The requirement of acquaintance, however, is even stronger in the case of Ockham's intuitive cognitions, since such cognitions simply cannot occur (in the natural order) in the absence of their objects. Once the speaker has been in direct contact with a given referent, she can, in Russell's view, use its logically proper name, even when the thing is not there IInymore; but this is not so with the intuitive cognition. Ockham's intuitive acts must correspond to an even more demanding category of signs, which contemporary philosophy of language docs not seem to have cared much about.

13 14 INTUITION, ABSTRACTION AND MENTAL LANGUAGE 15 I will call them 'rigid deictics', since they literally show their objects (this is the deixis aspect) and never change them (this is the rigidity). They refer to their objects without the help of any form of description, of any general concept, or of any intermediary whatsoever. They are not repeatable at will, but they nevertheless fully belong to the language of thought in so far as they can be subjects or predicates of mental propositions, and in so far as their presence within these propositions directly influences their truth-conditions and inferential roles. They can be identified neither with ordinary proper names, nor with spoken or written demonstratives, and not even, strictly speaking, with Russell's logically proper names. Ockham's semantics tum out to be suggesting here - without the Venerabilis Inceptor being clearly conscious of ithimself - a non-trivial enrichment of the general syntax of language, a new 'part of speech', which could exist only in the language of thought. 5. Mixed cognitions and singular terms of the third type Let us come back now to the third sort ofsingular terms listed by Ockham, those that are composed of a demonstrative accompanied by a common noun, such as 'this man'.55 If, as we are now entitled to admit, the utterance of a spoken demonstrative is normally associated with an intuitive cognition in the mind of the. speaker, such complex phrases as 'this man' must also correspond to mental acts involving intuitions in one way or another. Are we to grant, then, that a general term can be somehow incorporated within an intuition, and that some intuitive cognitions can have a descriptive content after all? The question is important for a correct understanding of how intuitions and concepts are connected with each other in Ockham's epistemology. And it has to be answered in the negative: the mental cognition corresponding to a phrase like 'this man' must be a compound, just like the spoken phrase is. What it is composed of is an intuitive act on the one hand (corresponding to 'this') and an abstractive act on the other hand (corresponding to 'man'). This is what I will call a mixed cognition. Its conceptual content ~ which is its abstractive part - stays utterly external with respect to the intuitive act which constitutes its other part. Both are associated with each other in a contingent way. The intuitive cognitionin itself is, in this case as in the previous one, devoid of any generality. In order to see it, we must pay attention to Ockham's developments on the semantics ofphrases oftheform 'this F'. Thefirst conclusion to be drawnfrom them is that a complex singular term of this sort can sometimes be used although the general term which is part of it does not truly apply to the object designated by the demonstrative. 56 William mentions at one point the case of 'this man is an animal' uttered when the designated object is in fact a donkey, or that of 'this white thing is not colored' uttered when no white thing at all exists in the worldy In this latter situation, Ockham surprisingly says, the uttered sentence would be true, whatever object is designated (demonstrato quocumque). This apparently strange diagnosis which Ockham doesn't care to explain much - becomes intelligible only if we grant that the subject 'this white thing', in such a case, supposits for nothing, even though a real object is designated by the speaker with the help of the demonstrative. What we have here is a negative proposition with an empty subject. and such sentences are always considered to be true in Ockham's semantics. 58 The only other possibility available would be that 'this white thing', in the example under consideration, should supposit for the non-white thing designated by the demonstrative; if it were so, however, the sentence 'this white thingis not colored' would not be true for any designated object: it would be false, in particular, ifthe speaker was pointing at a red object, since it is plainly false that a red object is not colored! That the supposition of the complex subject-term turns out to be empty in such deviant cases does not keep the demonstrative from designating something. Ockham's use of 'quocumque demonstrato' clearly indicates that something is designated in the example he is then considering; this becomes quite explicit when he imagines someone saying 'this man is an animal' while designating a donkey. The situation, therefore, is the following: the demonstrative, in those cases as in the normal ones, does designate a singular thing, and the common noun signifies the things it applies to (if such things exist). Combined to one another, they yield a complex term, which supposits for nothing in the propositions where it occurs if the object designated by the demonstrative is not one ofthe significates ofthe noun. This is precisely what happens for all other complex terms, where each component 'restricts' the scope of the other one: 'black horse'can supposit only for those beings which are signified both by 'black' and by 'horse'; if there were none, the complex term 'black horse' would supposit for nothing. This result can be directly transposed into mental language. The singular mental term corresponding to 'this man' must be seen not as an intuitive act internally endowed with a conceptual content, but as a mixed cognition, a complex mental term composed of an intuitive cognition (corresponding to the demonstrative) and an abstractive cognition (corresponding to the common noun). Such a compound can occur within a mental proposition, either as subject or as predicate. Ifit does, it will then precisely supposit for the one singular significate of the general concept which also happens to be the object of the intuitive cognition; if there is none, its supposition will be empty. Since it is thus possible that such a complex term represents nothing in certain contexts, even in the natural order; it cannot as a whole be identified with an intuitive act: a mental proposition of the form 'this F exists' could tum out to be false! Theintuitive and the abstractive cognitions, therefore, can combine with each other to form a complex singular term in the mental language, but they do not merge into one another. Each one keeps its identity and function. Intuitions and concepts are irreducibly distinct mental signs. 6. Ockharn's representationalism The Ockhamistic mapping of semantics into epistemology thus yields the following results: ( I) All simple abstractive acts are general terms in the mental language; this is precisely what a universal amounts to. (2) Intellectual intuitive acts are singular terms in the mental language; they can be subjects or predicates of mental propositions, and they can, as such, supposit for their objects.

14 16 INTUITION, ABSTRACTION AND MENTAL LANGUAGE 17 (3) There are mixed cognitions, composed of at least one irituitive.act and one abstractive act; they are complex singular terms in the mental language. All of these are signs naturally representirig in the mind certain objects other than themselves. A cognitive act, whether intuitive, abstractive or mixed, is a distinct singular thing, namely an internal quality of the mind, which represents something in virtue of certain external relations it has with its objects: causality, in the case of intuitive acts, or similarity in the case of abstractive acts. These acts are of such a nature, moreover, that they can be parts of mental propositions and influence their truth-conditions and logic, as well as their epistemic attractiveness for the agent. There is, it seems to me, a deep wisdom in this doctrine, which contemporary philosophy of mind and language should ponder. It provides a reasonably simple picture of the cognitive process and of its relation to language, without requiring any special ontological enrichment: no real universals out there in the world, no ideal objects, and no pure consciousness either, mysteriously open to otherness. The approach is both nominalistic and basically naturalistic. It conjoins the idea of an intellectual intuition with that of a mental language, and provides intriguing suggestions for a semantical approach to our understanding of singular things and its integration within discursive thought. Ockham's epistemology, at this point, turns.out to be far less 'antirepresentationalist' than what several recent commentators have taken it to be. 59 Of course, we have to be careful with such polymorphic labels. Marilyn Adams, for example, contrasts two approaches to epistemology, which she respectively calls 'direct realism' and 'representationalism', and she unreservedly associates Ockham with the former. 6o Given what she means by those terms, I have no quarrel with that. What Adams calls 'representationalism' is the doctrine according to which the immediate object of cognition is always a representation rather than the external thing itself; and 'direct realism', in her vocabulary, simply is the negation of this thesis. It is straightforwardthatockham's later theoryofcognition is a case ofdirect realism in this sense, and not of representationalism, since both intellectual intuitions and concepts are identified, in this theory, with cognitive acts having (in most cases) external things as their immediate objects. These cognitive acts, however, are themselves seen by Ockham as representations. A mental representation in contemporary philosophy of mind is usually taken to be a symbolic mental token capable of playing a causal role within the mind and of referring to something other than itself, which it can stand for in mental computations. Cognitive acts, for Ockham, are precisely like that: both intuitive and abstractive acts are real things within the mind, both are similitudes of their objects, both are signs capable of suppositing for them in mental propositions, and both are elements of causal chains in the mental life of human agents; Ifrepresentationalism is defined - as it seems natural to do - as any epistemological theory that attributes an indispensable and crucial cognitive role to such mental representations, then Ockhamism is a full-fledged form of representationalism in this sense. Although it does not, in its mature version, postulate intermediate mental objects for cognitive acts, its account of cognition always requires, nevertheless, a third entity between the mind and the external things: the cognitive act itself, which is indeed, whether intuitive or abstractive, a semantical representation. Notes 1. See, for instance, Quaest. in Phys., q. 1-7 (rightly called questions de conceptu in the critical edition), OPh VI, pi> See, forinstance, SL I, 1, OPhI, pp See Ord. I, Prol., q. 1, OTh I, pp. 3-75; and Quod!. V,q. 5-6, OTh IX, pp ; and also Rep. II, q , OTh V, pp. 251'-310. Many commentators have ably discussed these passages; see in particular: Day 1947; Adams 1987, chap , pp ; Alferi 1989, especially pp ; Michon 1994, chap. 2~3, pp ; and, most recently, Karger 1999a, Stump Cf. Ord. I, Prol., q. 1, OTh I, p. 31: '... notitia intuitiva rei est talis notitia virtute cuius potest sciri utrum res sit vel non, ita quod si res scit, statim intellectus iudicat earn esse, nisi forte impediatur propter imperfectionem illius notitiae'; and p. 70: '... dico quod per notitiam intuitivam rei potest evidenter cognosci resnon essequando non est vel si non sit'; see also Quod!. V, q. 5, OTh IX, pp Cf. Ord. I, Prol., q, 1, OTh I, p. 31: 'Notitia autem abstractiva est ilia virtute cuius dere contingente non potest sciri evidenter utrum sit vel non sit... Similiter, per notitiam abstractivamnulla veritas contingens, maxime de praesenti, potest evidenter cognosci.' 6. Ibid., p. 32: 'Et per istum modum notitia abstractiva abstrahit ab exsistentia et non exsistentia, quia nec per ipsam potest evidenter sciri de re exsistente quod exsistit, nec de non exsistente quod non exsistit, per oppositum ad notitiam intuitivam.' 7. See in particular: Boehner 1943, 1945; Adams 1970; Streveler 1975; Tachalj 1988, pp. 113:...53; and Karger 1999a. 8. See Quod!. VI, q. 6, OTh IX, pp : 'Utrum cognitio intuitiva possit esse de obiecto non existente.' Ockham's answer is twofold: 'In ista quaestione pono duas conclusiones: prima est quod cognitio intuitiva potest esse per potentiam divinam de obiecto non existente' (p. 604); 'Secunda conclusio est quod natura1iter cognitio intuitiva non potest causari nec conservari, obiecto non existente'. (p. 606), Let me stress - because it has been the source of many misunderstandings - that the intuitive cognition of non-existent beings, forockham, is not misleading: the agent in such cases correctly judges that the things in question do not exist. This, outhe other hand, in rio way implies that Godcould not deceive us if He should choose to do so; it would simply be a different sort of case (see on the latter point Quod!. V, q. 5, pp ; along with Karger 1999a's insightful commentaries on this passage). 9. See for example Ord. I, Prol., q. 1, OTh I, pp Cf. Ord. I, Prol., q. 1, OTh I, p. 64: '... habet [intellectus] aliam notitiam per quam concernit hic et nunc...: et illa est notitia intuitiva'. Also Quod!. I, q. 15, OTh IX, pp. 83-'--6: 'Utrum intellectus noster pro statu isto cognoscat intuitive sensibilia.' On the history of singular intellection in medieval philosophy, see Berube II. This argument is spelled out, for example, in Ord. I, Prol., q. I,DTh I, pp. 25-7,and in Quod!. I, q. 15, OTh IX, pp For Ockham's theory of the plurality of substantial forms, see in particularquod!. II, q. to-ii, OTh IX, pp Adams 1987, chap. 15, provides a detailed presentation and discussion of the theory. For the historical background of the. controversy; see Zavalloni Stump 1999 holds, in a Thomistic vein, that Ockham's doubling of intuitions is superfluous and that 'the operations of the two cognitive faculties seem completely redundant' in his account (p. 194). Her brief discussion ofthepoint, however (p. 202, n. 75), disregards Ockham's requirement that the relevant intuitive act should be apart, and not merely an exlemai cause, of the existential proposition which the agent evidently judges to he true. More on this helow.

15 18 INTUITION, ABSTRACTION AND MENTAL LANGUAGE On Ockham's theory of intellectual habitus, see in particular Rep. III, q. 7, OTh VI, pp , and Quodl. III, q , OTh IX, pp ; also chapter 2, section 1 below. The theory of memory, on the other hand, has been ably studied by Adams 1987, pp , who shows, in particular, its evolution in Ockham's works. 15. For recent presentations of Ockham's theory of mental language, see, among others: Tabarroni 1989, Normore 1990, Karger 1994, Panaccio 1992a, 1992b, 1999a, 1999b. 16. SU, 1 OPh I, pp See SL 1,3, OPh I, pp ; and Quodl. V, q. 8, OTh IX, pp Ockham raises a doubt as to the presence of participles and pronouns in mental language. Interjections, on the other hand, are ignored, and grammatical genders are explicitly excluded as devoid of semantical relevance. 18. See, for example, the various treatises of the late twelfth and early thirteenth centuries edited by de Rijk See Panaccio 1999a. 20. See Boehner 1951, Leff 1975, and Adams 1987, esp. chap A good presentation of Ockham's version of this fictum-theory is provided by Karger SL I, 12, OPh I, p. 43: '... supponere pro alio et signijicare aliud potest competere actui intelligendi sicut alii signa. Igitur praeter actum intelligendi non oportet aliquid aliud ponere' (transi. Loux 1974, p. 74; the italics are mine). 23. Ibid.:'... frustra fit per plura quod potest fieri per pauciora. Omnia autem quae salvantur ponendo aliquid distinctum ab actu intelligendi possunt salvari sine tali distincto... '. On the place of the Razor Principle in Ockham's nominalism, see Maurer 1978 and See, for example, SL I, 1, OPh I, p. 7: 'Terminus conceptus est intentio seu passio animae aliquid naturaliter significans vel consignificans, nata esse pars propositionis mentalis, et pro eodem nata supponere.' 25. See, for example, Quod/. V, q. 7, OTh IX, p. 504: 'Et loquor de conceptu qui est cognitio abstractiva'; or Quodl. I, q. 13, where Ockham indifferently uses conceptus or cognitio abstractiva (OTh IX, esp. pp ). 26. See Ord. I, ProI., q. 1, OTh I, pp : 'Sciendum tamen quod notitia abstractiva potest accipi dupliciter: uno modo quia est respectu alicuius abstracti a multis singularibus... Aliter accipitur cognitio abstractiva secundum quod abstrahit ab exsistentia et non exsistentia et ab aliis condicionibus quae contingenter accidunt rei... '. 27. See Quodl. V, q. 7, OTh IX, p. 506: '... dico quod intellectus noster de nulla creatura potest habere aliquem talem conceptum simplicem proprium...; et hoc quia quaelibet talis cognitio sive conceptus aequaliter est similitudo et repraesentat omnia individua simillima, et ita non plus est conceptus proprius unius quam alterius'; and Quodl. I, q. 13, OTh IX, p. 74: '... nulla cognitio abstractiva simplex est plus similitudo unius rei singularis quam alterius sibi simillimae...; igitur nulla talis est propria singulari sed quaelibet est universalis'. On Ockham's evolution in this respect, see Adams 1987, pp , and Panaccio 1992a, pp See Quod/. I, q. 13, OTh IX, p. 77: '... habeo aliquam cognitionemabstractivam propriam, sed illa non erit simplex sed composita ex simplicibus'. 29. See SL I, 15, OPh I, p. 53: '... omne universale, secundum omnes, est de multis praedicabile; sed sola intentio animae vel signum voluntarie institutum natum est praedicari et non substantia aliqua; ergo sola intentio animae vel signum voluntarie institutum est universale.' 30. SL I, 14, OPh I, p. 48: '... non est universale nisi per significationem, quia est signum plurium' (transi. Loux 1974, p. 78). 31. See Quodl. I, q. 13, OTh IX, p. 74:'... dico quod cognitio prima abstractiva primitate generationis et simplex non est propria singulari, sed est cognitio communis aliquando, immo semper'. 32. On maximal similarity as cospecificity in Ockham, see Panaccio 1992a, pp Quodl. IV, q. 17, OTh IX, p. 385: '... conceptus speciei potest abstrahi ab uno individuo' (transl. Freddoso and Kelley 1991, p. 317). Similarly, the short summary ofthe cognitive process given in SL III-2, 29, OPh I, p. 557, requires the meeting of only one man for the original formation of the concept of man. 34. Quodl. I, q. 13, OTh IX, p. 77: '... dico quod conceptus generis numquam abstrahitur ab uno individuo' (transi. Freddoso and Kelley 1991, p. 67). See also SL III-2, 29, OPh I, p As was suggested by Kneale and Kneale 1962, pp For a detailed critical discussion of this interpretation, see Panaccio 1992a, pp See Ord. I, dist. 30, q. 5, OTh IV, p. 385: '... similitudo dicitur relatio realis propter hoc quod unum album ex natura rei est simile alteri albo, et ad hoc quod unum sit simile alteri non plus facit intellectus quam facit ad hoc quod Sortes sit albus vel quodplato sit albus'. 37. Adams 1987, p See, for example, Ord. I, ProI., q. 1, OTh I, p. 50: '... dico quod... nulla propositio contingens possit evidenter cognosci nisi ex notitia intuitiva alicuius extremi... ' (the italics are mine). 39. The point is made in detail by Karger 1994 and In the final redaction, for example, the passage quoted above in note 38 is followed by 'vel significati per extremum'. The St. Bonaventure critical edition of the Ordinatio conveniently identifies Ockham's later additions to his own text. Most of the time, they have to do with his shift from the fictum to the actus-theory of concepts. 41. Quaest. in Phys. q. 7, OPh VI, p. 411: 'intellectus apprehendens intuitione rem singularem elicit unam cognitionem intuitivam in se quae est tantum cognitio illius rei singularis, potens ex natura sua supponere pro illa re singulari '" Et ita sicut vox supponit ex institutione pro suo significato, ita ista intellectio supponit naturaliter pro re cuius est.' 42. See SL I, 63, OPh I, p. 193: '... restat dicere de suppositione, quae est proprietas conveniens termino sed numquam nisi in propositione Et sictam subiectum quam praedicatum supponit'; and SL I, 69, OPh I, p. 208: ' solum categorema, quod est extremum propositionis, significative sumptum, supponit personaliter.' 43. This is precisely why Ockham's intellectual intuition is not redundant, after all, with the sensory intuition (pace Stump 1999; see above n. 13): only a properly intellectual act could be a proper part of a mental proposition formed within the intellect. 44. See Ord. I, ProI., q. 1, OTh I, p. 39: 'Sed visio intuitiva, tam sensitiva quam intellectiva, est res absoluta, distincta loco et subiecto ab obiecto.' 45. SL I, 1, OPh I, p. 9: '... accipitur signum pro illo quod aliquid facit in cognitionem venire etnatum estpro ilio supponere' (transi. Loux 1974, p. 50). More on this definition in chapter 3 below. 46. See Rep. II, q , OTh V, p. 287: 'Dico tunc quod intellectio est similitudo obiecti.' The context shows this to apply to intuitive as well as abstractive intellection, sinc;e Ockham is then speaking ofthe intellectual apprehensionofsingularthings and replying to an objection having to do with both intuitive and abstractive cognitions. 47. Ibid., p. 288: '... licet intellectus assimiletur omnibus individuis [eiusdem speciei) aequaliter..., tamen potest unum determinate cognoscere et non aliud. Sed hoc non est propter assimilationem, sed causa est quia omnis effectus naturaliter producibilis ex natura sua determinat sibi quod producatur ab una causa efficiente et non ab alia... '. 41<. See Ord. I, dis!. 45, q. unica, OTh IV, p. 665:'... omnis causa proprie dicta est causa immediata'. 49. See SL I, 19, OPh I, p. 66.

16 See Adams 1987,p. 319, n. 9. Adams even doubts that pronouns are counted as categorematic terms by Ockham. Yet this is what he explicitly does in Quodl. II, q. 19, OTh IX, p. 195: 'Et ideo dico quod ex primaria institutione [pronomen] est vox categorematica.' See also SL III-4, 10, OPh I, p. 798, where Ockham counts pronouns among the signs having a determinate signitication, which is precisely how a categorematic term is defined when the notion is first introduced in SL I, 4, OPh I, p SeeSL I, 3, OPh I, p~ 11, where after having raised a doubt about the presence of participles in the mental language, as distinct from verbs, he laconically adds: 'Et de pronominibus posset esse consimilis dubitatio.' In the corresponding discussion of Quodl. Y, q. 8, he doesn't mention pronouns at all among the grammatical categories that are needed, according to him,.for the analysis of the language ofthought. 52. See Panaccio 1980, 1992a, pp Kaplan 1978, Russell 1918, pp. 200 et seq. 55. SDI, 19, OPh I, p Karger 1978 has judiciously called attention to cases of this sort. 57. SL III-3, 32, OPh I, p See SL II, 2, OPh I, p. 255, where the thesis is made explicit for the case of particular negative propositions. It also holds, a fortiori, for universal negative propositions, as specified in SL III~3, 32, OPh I, p For example: McGrade 1986; Tachau 1988, chap. 5; and Alferi Adams 1987, pp Chapter 2 Intellectual Acts Concepts, then, are intellectual acts in Ockham's mature theory. This is an intriguing idea. For one thing, the sense in which Ockham uses the term 'act' is an unfamiliar one for modem readers and has to be explained; this will be the object of section 1 of the present chapter. The reasons which led him to drop the fictum-theory of concepts in favour of the actus-theory will then be reviewed (section2), as well as the reasons why he rejected, from the very start, the so-called species-theory, held in particular by Thomas Aquinas (section 3). Once this is done, we will be in a position to reach a better understanding of how conceptual acts fit into human thought in Ockham's view (section 4). 1. Mental acts and habitus The first thing to notice concerning Ockham's general notion of actus is that it is very different from today's idea of action as it occurs, say, in the philosophy of action. An action in the modem sense roughly corresponds to 'someone's doing something intentionally'. I Acts in the medieval sense, by contrast, are not always done intentionally, and are not always someone's acts either. One could speak of heating as being the act of a fire, or of shining as the act of the Sun. The eruption of a volcano or the ringing of an alarm-clock would be good examples too. The relevant background here is Aristotle's idea of actuality (entelecheia), as opposed to mere potentiality. The act of something, in this vocabulary, is itsactual operation, what it does, that is, in virtue of its internal powers being somehow set into activity. Such operations can in many cases be triggered by external factors, yet they will be considered as acts in the relevant sense as long as they are typical realizations of certain internal powers the thing has in virtue of its essence, rather than merely accidental occurrences: growing, for example, is an actus of a tree, but its falling down as a result of being struck by lightning is not. [n the case of human beings, Ockham distinguishes between external and internal acts. 2 External acts are publicly observable and directly affect something outside the agent. Most intentional actions in the modem sense, such as walking, cutting a tree, or speaking aloud, are typical external human acts in this sense; but so is breathing or sweating, although these are not intentional Internal acts, on the other hand, are II/ental actualizations of various human psychological powers or faculties (potentiue): there are appetitive acts, for example, corresponding to the sensitive appetitive faculty, such as acts of desire, fear, or repulsion;3 there are volitive acts or acts of the will, such as decisions;4 and, of course, there are cognitive acts too, whether sensitive or intellectual (among which concepts). None of these, let me insist, need be intentional in the sense in which an action is usually taken to be. 21

17 22 INTELLECTUAL ACTS 23 Many intellectual acts, in particular, are not under the control of the will, according to Ockham,5 and are not, consequently, the sort of things that can be done intentionally. A further distinction which was unanimously considered in the thirteenth and fourteenth centuries as crucially relevant for a correct understanding of human mental acts is the one Aristotle had allusively drawn in Book II of the De Anima: 'But actuality; he said there, is so spoken of in two ways, first as knowledge is and second as contemplation is.'6 In the medieval vocabulary, these came to be called respectively the primary act (actus primus) and the secondary act (actus secundus),7 or again - as in Ockham - the habitus and the act tout court. Aristotle's point was that the actualization of a capacity often comes in two degrees, especially in human beings. The example he points at in the relevant passage of the De Anima is that of a given scientific truth. At first, one is utterly ignorant of it, but has the capacity to learn it: this is a case of mere potentiality. When the person has learned this particular truth, but is not currently contemplating it, this knowledge is said to be actualized in her as a primary act - or habitus. When she does actually contemplate the truth in question, it thereby gains a stronger form of actuality within her, and is said to be there as a secondary act - or act tout court. In Ockham's vocabulary, then, the habitus is an acquired disposition, while the mental act is an actual psychological episode. Ockham makes abundant use of this distinction in epistemology as well as in ethics. In the end, the general pattern of connection he establishes between mental habitus and the corresponding acts comes down to the following two principles:. (a) (b) A psychological habitus is normally caused by an original psychological act leaving a trace within the mind;8 The habitus in turn is a partial efficient cause for further acts of the same sort as the original one. 9 In order to avoid an infinite regress, this requires, as Ockham acknowledges, that not all mental acts should (even partially) be caused by a habitus. 1O In the case of knowledge, in particular, the intuitive grasping of an object is not normally caused by an acquired habitus: the J;Ilind simply has an innate capacity to grasp - sensitively and intellectually - objects that are physically present in its environment. When this is done, though, the intuitive act, as we have seen in chapter I, causes an abstractive act, which in turn causes a habitus to be implemented within the mind. This habitus will then serve, under favourable circumstances, as a partial cause for further abstractive acts, numerically distinct from the one that caused it, but of the same sort. ll This causal effect of the habitus is described by Ockham as an 'inclination'. A mental habitus, he typically says, inclines the mind towards certain sorts of acts: 12 when the other relevant factors are present, acts of this sort will tend to occur within the mind of whoever is endowed with such a habitus. And when a new act is thus engendered by a psychological habitus (whether cognitive, volitive or appetitive), a new habitus, similar to the previous one, will in turn be engendered, or - more often, presumably - the previously existing habitus will he strenghtened.1\ Acts and habitus thus form a rich cluster ofcausal chains within the mind. Ontologically speaking. all these psychological acts and hahitus are seen hy Ockham as real singular qualities of singular minds. 14 His nominalistic ontology, remember, admits of only two sorts of real things in the world: singular substances and singular qualities. Since the existence of certain psychological acts is known by experience (we know, for instance, that we are joyful or sad),15 the reality of such acts is above suspicion. Being accidental and transient states of substances (I am sometimes sad, sometimes not), they cannot be identified with the substances themselves. They have, therefore, to be qualities. If this is true of some psychological acts, such as pleasure or sadness, there is no reason why it shouldn't be true of all of them in so far as they are real. And if it is true of psychological acts in general, it must be true of psychological habitus as well. The upshot of it all for our present purpose is that concepts, if they are to be identified with the abstractive acts of human minds, are real episodic singular qualities of such minds, generated either by intuitive acts (when such acts occur) or by previously acquired cognitive dispositions or habitus, and causally producing, in turn, new cognitive dispositions within the mind or, at least, strengthening previously existing ones. 2. From the fictum to the actus But why should concepts be identified with cognitive acts? Many medieval authors, after all, thought otherwise. Thomas Aquinas, for one, had it that a concept is the object and product of the intellectual act, rather than the act itself. 16 Ockham, in his earlier teaching, had favoured a similar approach himself, known in the literature as the /ictum-theory of concepts. 17 According to this doctrine, concepts are not real entities belonging to the Aristotelian categories of substance or quality, but mere objects of thought, having no other existence than that of being cognized (which Ockham then called 'intentional being' [esse intentionale] or 'objective being' [esse ohiective]).18 The Venerabilis Inceptor probably thought for a while that the postulation of such special unreal objects was necessary to avoid admitting common natures into being, as his senior colleague Walter Burley had been led to. Burley, in effect, having rejected Aquinas's theory of the concept as the intentional object of the intellectual act, was quite happy with having ontological universals, instead, as ohjects for such acts. 19 Ockham, on the other hand, thought from the start that the admission of universals or common natures as extramental beings was one of the worst possible errors in philosophy, leading to all sorts of incongruities and inconsistencies. 20 While he was ready to admit real individual things themselves as ohjects for singular cognitive acts, he believed at first that general thoughts required some special sort of objects and he reverted, therefore, to flcta, rather than common natures, to play this role. As Elizabeth Karger ably writes: 'It seems that Ockham must have introduced /icta as substitutes for common natures... without the drawhacks of an objectionable ontology.'21 The question is: why did he change his mind? Many recent commentators have surmised that Ockham's evolution on this point was due to the ohjections raised in the early I320s against his theory of concepts as Jicta hy his Franciscan confrere Walter Chatton. 22 The matter, however. requires a closer look. In his most complete listing of reasons for dropping the jictum-theory. in question I of the Qum stion!;.\' in Lihras Physicorum Aristoteli.\, Ockham provides

18 24 INTELLECTUAL ACTS 25 seven arguments, only two of which are directly borrowed from Chatton's original critique.23 The first clear signs of Ockham's evolution on the matter, moreover, occur in the Prologue of his commentary on the Perihermeneias, which, according to the chronology proposed by the editors, was written before Chatton's intervention,24 and two of the arguments against the.fzctum-theorymentioned with approbation in the Quaestiones in Libros Physicorum do come from this earlier text. 25 It might very well be that the discussion with Chatton was the occasion for Ockham to rethink the question for himself andthat some of Chatton's points really impressed him, but on the whole, as we are about to see, his motivation for shifting from the flctum to the actus-theory of concepts was quite distinctive. 26 The key reason for his change of mind is revealed by Ockham,not so much in the arguments he offers against thefictum-theory, as in the one positive consideration he brings forward in favour of his new position, both in the Quodlibeta and in the Summa Logicae: 'whatever is preserved by appeal to a fictive entity can be preserved by appeal to an act of understanding'.27 Seeing this as the central point invites for a reorganization of Ockham's whole argumentation on the matter in the form of the following modus ponens: (1) it would be preferable to dispense with the fictum if we can; (2) it is indeed possible to dispense with the flctum,since all the functions we want to attribute to concepts can be adequately fulfilled by intellectual acts; therefore: (3) we should dispense with the.fzctum in favour of the actus. All of Ockham's particular objections against the flctum-theory, whether in his commentary on the Perihermeneias, in Quodlibeta IV, 35 or in the Quaestiones in Libros Physicorum 1, can be associated with either premiss (1) or premiss (2) of this master argument, some of them enumerating the drawbacks of an appeal to the fictum (without, however, being decisive in Ockham's eyes), while the others insist on its actual dispensability. The whole scheme can be reconstructed in the following manner. First, the drawbacks. There are five of them, the first two - in the order adopted here - having to do with complications induced by the introduction of the flctum, and the last three with conflicts it tends to generate with well-accepted theoretical principles: (1.1) The acceptance of the fictum as the object of cognition requires a counterintuitivecomplexification of the ontology, since it cannot be identified with a natural entity, whether a substance or an accident. 28 (1.2) It also induces a special complexification of the epistemological process, since the position of such aflctum is not analytically required by the idea of an intellectual act of cognition (even a general one).29 (1.3) The insertion of the fictum as the object of general cognition represents a threat to direct realism in epistemology, otherwise favoured by many medieval authors including Ockham. since it introduces an intermediate between the act of cognition and the thing itself. \() It must be stressed that. contrary to what Marilyn Adams suggests, this objection is considered relevant by Ockham only in so far as it has to do with general abstractive cognition, rather than with intuitive cognition, since he never had supposed anyway that the latter involved a fictum. Adams takes the argument to be confused because she thinks it rests on the implicit premiss that flcta, according to the flctumctheory, 'are always the immediate objects of thought and awareness',31 which, as she rightly remarks, Ockham had never accepted in the case of intuitive cognition. But Ockham, of course, knew that very well. What he must mean here is that the fictum-theory threatens the adequacy of general knowledge. (1.4) Endorsing the flctum is. not easily compatible with the idea (which was generally accepted in Ockham's times) that the representational function of general concepts rests on some similarity they are supposed to have with the things they represent: being an unnatural and unreal sort of entity, the flcta would seem to differ more than any real thing from whatever it is that they are expected to represent. 32 (1.5) The postulation of ficta, in so far as it implies the eternal and necessary existence of a realm of purely intelligible objects, runs counter to the principle of God's omnipotency and to the radical contingency of all created beings. 33 Ockham, it seems to me, must not have taken these five points as decisive by themselves against the flctum-theory. After all, he could have lived with the complications pointed out by the first two objections, had he thought them necessary for a correct account of cognition. And the introduction of ideal intermediates hetween minds and things, that gives rise to objection 1.3, would certainly have been preferable to him than the acceptance of common natures as direct external objects for scientific knowledge, even at the price of tempering his direct realism a bit. As to the last two difficulties, Ockham, no doubt, could have found ways of alleviating them. As he had already remarked in his commentary on the Perihermeneias when reacting to objection 1.4, the only sort of similarity which is relevant for an account of human cognition is the one that can hold between a mental representation whatever that is - and what it represents. In theflctum-theory, that would be the sort of similarity an intelligible object can have with real things - intentional similarity, let's call it - and there would be no point in comparing it in terms of degrees with similarities among real things: a mule might look pretty much like a horse, yet it cannot serve as a mental concept for horses, simply because it is not the sort of entity a mind can do computations withp4 The existence offlcta, finally, could have been made contingent upon the existence of the corresponding acts of thought in human minds, which would have sufficiently countered objection 1.5. The five problems mentioned up to now are not innocuous for all that - far from it - hut they could have force against thefictum-theory in Ockham's eyes only in so fllr as an acceptable alternative was available, which could avoid these complications and drawhacks without reintroducinf? common natures in the external world. a requirement Chatton's approach could hardly satisfy, being very close to that of John ()unsscotwi in metaphysics. 1\ A second argumentative step was needed in order to

19 26 INTELLECTUAL ACTS 27 show that the cognitive functions of concepts could be fulfilled by intellectual acts without undesirable ontological consequences. Some of Chatton's remarks could be revived in the process, no doubt, but the crucial point, as we shall see, was original. Three more of Ockham's arguments are relevant here: (2.1) All the uncontroversial propositions we want to accept about conceptual activity, propositions such as 'horses are conceptually cognized' or 'horse is a predicate' and so on, could very well be true even if no fictum existed at all, but only cognitive acts and real external things. If, however, the truth of a proposition requires only two things, it is inappropriate to postulate a third one on this sole basis. 36 This objection, as can readily be seen, explicitly rests on a version of the famous Razor Principle which came to be associated with Ockham's name. 3? The crux of it, nevertheless, is the assertion thatjicta are not needed to account for the sort of things we normally want to say about concepts. But this amounts to little more than to what I called premiss (2) above. How it can be sustained remains to be explained. (2.2) Concepts are expected to be subjects and predicates in mental general propositions. But this is something the cognitive act could do just.as well as the fictum. 38 This argument, closely inspired by Chatton,39 brings to the fore a particular case with respect to the previous one by stressing one particular function that concepts should fulfill, that of being the subjects or predicates of mental propositions. What it does not do, however, is to show how exactly an intellectual act could come to serve as a subject or a predicate. In Ockham's logical doctrine, what is required for anything to be the subject or predicate of a proposition is that it should be able to have a supposition (or referential function).40 The argument, then, presupposes in Ockham's hands that intellectual acts can in general supposit for something. But it does not yet show how this is possible. (2.3) An intellectual act can represent something outside the mind, it can signify such a thing, and supposit for it, just as much as afictum was expected to. This, at last, is the heart of the matter. The functions now enumeratedrepresentation, signification, and supposition - are the basic ones concepts are supposed to fulfil. If intellectual acts can do that, then ficta indeed are dispensable without loss. This general argument had first been sketched by Ockham in his commentary on the Perihermeneias as an argument in favour of positing real mental qualities within the mind, which, while distinct from the intellectual acts, could serve as objects for these acts,41 a position he never actually subscribed to. Later on, however, he realized that these functions could be directly assigned to intellectual acts. Both in the Quodlibeta and in the Summa Logicae this is his main point in favour of the actus-theory.42 The whole argumentation,. as developed by Ockham, ultimately rests on his realization that some intellectual acts can serve as mental s;~ns, be endowed, that is, with significations of their own and be capable of referential functions within mental propositions. Once this is admitted. considerations 2.2 and 2.1 directly follow. the drawbacks mentioned in can be given all their weight, and the general master argument smoothly runs through. How Ockham came to see this as possible without reintroducing common natures in the ontology is revealed, interestingly enough, by the one aspect of argument 2.3 which he substantially modified with respect to the original version of it he had given in this' commentary on the Perihermeneias. He was explicit in this earlier passage that this argument, in his view, could not be invoked in favour of identifying concepts directly with intellectual acts, because he found it difficult to see what the objects of such acts would then be, what it is, in other words, that such cognitive acts would be cognitions of. 43 What he was presupposing at the time is that every act of cognition must have one and only one object into which it 'terminates'. This is why he could see argument 2.3 as capable of beingused in favour of identifying concepts with special mental qualities serving as objects for intellectual acts, but not yet with the intellectual acts themselves. 44 Dropping this presupposition is precisely what opened the way forhis final adoption of the actus-theory. As I said earlier, it might very well be that the debate with Chatton helped him in this regard, but none of Chatton's arguments could have convinced him as long as he didn't see how he could avoid the postulation of special mental objects for intellectual acts without having to revert to common natures as what these acts should terminate in. The breakthrough must have come when he realized that simply dropping an unexamined and unwarranted assumption about cognitive acts could do the trick. Not that he came to accept that cognitive acts did not need any object at all, or did not need to 'terminate' in anything at all. What he realized instead is that each cognitive act does not need exactly one object. This must have been greatly stimulated by his coming to take with utmost seriousness the idea that a general concept is but the sign of many singular things at once, a doctrine he had already sketched in his commentary on the Sentences. 45 Clearly realizing that the identification of the cognitive act with the concept amounts to identifying it with a sign of many singular things opens the way to seeing how it can still be a cognition of something without this something being either a fictum or a common nature: if conceptual acts are seen as general signs, only their singular significata are needed to serve as their objects..t The rejection of intelligible species Another doctrine which Ockham is famous for having rejected concerning mental representation is the so-called species-theory. In its full-fledged version, as it had been developed by thirteenth-century 'perspectivists' like Roger Bacon, John Peckham and Witelo (all heavily influenced in this regard by the Arab philosopher Alhazen),46 this theory held that in order for human cognition to take place, the forms of external things had to penetrate the mind somehow. This was thought to re4uire three successive patterns of representative emanations - called species llowing out from external things: first, there was the species in medio, which was supposed to earry the thing's image through the ambient environment (called the fi/('(jium); second. a sensible.lpc'de.i' was said to be formed within the sensitive organ of the knower; and third. an intelligible,ipecic's was postulated within the intellect.

20 28 INTELLECTUAL ACTS 29 Ockham, from the very start, rejected all three of these.. I will focus here on his discussion of the intelligible species, which is the more directly relevant in the present context.47 This is a theme that has loomed large in the recent literature on fourteenth-century epistemology.48 Yet a number of misunderstandings still circulate about it. There seems to be a consensus, in particular, that Ockham misinterpreted the theory he was criticizing in this instance 49 and that his own scheme ofintellectual acts and habitus comes pretty much to the same. Which, as far as I can see, is hardly the case. What we need to do to settle the point is,first, to correctly identify what it was exactly that Ockham wanted to reject (section 3.1), and second, to recall what his main critique was (section 3.2). 3.1 Species as preconditions for intellectual acts As to the first point, everybody agrees that Ockham's main target on this was Thomas Aquinas's doctrine (or some version of it). Aquinas, of course, was far from being the sole promoter of intelligible species (and Ockham, in fact, explicitly mentions Scotus in this regard),50 but as Leen Spruit correctly points out, Aquinas had offered what 'may legitimately be called the "canonical" theory of intelligible species',5\ and Ockham was well aware of this. 52 Aquinas thought that, since universalintelligible forms couldnot be posited as actually existing outside the mind (as Aristotle was taken to have shown against Plato), they had to be actualized within the mind by a special intellectual power, called the 'agent intellect', which extracted them, so to say, from the sensible species deposited within the imagination as a result of sense perception. This process of 'abstraction', in his view, led to the formation of intellectual general representations -the intelligible species, precisely - which could then serve as the means for general knowledge. 53 What Ockham objected to in this account is not that it involved intellectual representations stored in the mind for future use - he needed that too himself in the guise of the habitus - but that the intelligible species it postulated should always be prior to the intellectual acts ofcognition, and that they were seen as presupposed by these acts: I take it that the species [according to this theory] is a precondition for the act of intellection and can remain both before and after intellections, even in the absence of the thing. And it is distinct, therefore, from the habitus, since the intellectual habitusjollows upon an act of intellection, while the species precedes both the act and the habitus. 54 Ockham's whole discussion of the matter in the Reportatio makes it clear that what he wants to dispense with is precisely this antecedence of the stored intellectual representation over the intellectual act of cognition. It has to be understood that the process of abstraction which leads to the formation of the intelligible species is not, in Aquinas's doctrine, the intellectual act of cognition itself, but a preparatory phase, postulated as a condition of possibility for intellectual acts to occur: the intelligible species resulting from this somewhat mechanical - and unconscious - process of abstraction is, in Aquinas's words, 'the source fprincipium] of the action of the intellect', and it must be carefully distinguished both from the cognitive act itself, which comes after it, and from what Aquinas calls a concept - or conception, or mental word - which he sees as the result of the cognitive act, while the intelligible species is, in all cases, supposed to be its starting point. 55 That this is where the disagreement lies between Ockham and Aquinas is something several recent commentators, even among the best, seem to have missed. Robert Pasnau, for one, bluntly asserts in the course of his discussion of Ockham's rejection of the intelligible species, in his otherwise remarkable book on Theories of Cognition in the Later Middle Ages, that 'this portion of Ockham's account... isn't different from Aquinas's', since both admit that the contact with external things puts the intellect 'into a certain state such that cognition with a certain intentional content follows',56 Ockham's point, however, is precisely that there is no such state, as distinct.from both the cognitive act and the habitus that follows upon it. Pasnau thinks that these 'states', in Aquinas's theory, are not to be reified as distinct entities with respect to the intellect itself;57 but if the intelligible species has an accidental being within the intellect for Aquinas (as Pasnau readily, and correctly, acknowledges),58 then ius certainly not to be identified with the intellect itself, even though it is not independent from it. The intelligible species is not a substance of its own, of course, in Aquinas's view; nevertheless,it is not ontologically reducible to either the intellect or the cognitive acts and habitus. 3.2 The Razor argument againstintelligible species Why Ockham rejects the postulation of such special entities is straightforward: they are,he thinks, entirely superfluous. His whole argument, here, rests upon the Razor Principle: since intellectual cognition can be accounted for with acts and habitus alone, intelligible species, seen as extra entities, are not needed. In the relevant passage of the Reportatio, he successively considers five cognitive functions one might want to attribute to the intelligible species, and purports to show that all five of them can be sufficiently fulfilled without it. 59 The first one is cognitive assimilation. It was commonly accepted in the Middle Ages that cognition, whether sensitive or intellectual, requires the knower to develop within himself some sort of similarity with the known object. Ockham agrees, but insists that the relevant sort of assimilation is sufficiently explained by assuming the act of intellection to be originally caused by the object itself. 60 It has to be stressed, pace Tachau, that this argument in no way hangs upon some anti-representationalist ttendin Ockham's thought. Tachau wrongly assumes that Ockham 'rejects the notion that intellectual cognition is a process of assimilation of intellect to object'.61 He does accept it, quite to the contrary, for every variety of simple cognitions, whether intuitive or abstractive. His point, rather,is that the existence of preparatory intelligible species is no more a necessary precondition for abstractive assimilations than it is for intuitive ones. 62 A second reason one might want intelligible species for, Ockham says, is to account for mental representation; but, once more, the object itself and the cognitive act arc sufficient for this. Ockham's starting point, here, is that the intuitive grasping of an object does not require any intervening representation between the object and the act, the direct causing of the act by the object being obviously enough in this case. h ' If this is so for intuitive. cognition. he argues, then there is no reason why it should he otherwise for the ahstrat'tive cognition that immediately follows upon it:

21 30 INTELLECTUAL ACTS 31 'just as the object sufficiently represents itself in the one cognition, it does so too in the other, which immediately follows upon the intuitive cognition'.64 Again, the idea is not to evacuate mental representation altogether. The argument, on the contrary, assumes that such a representative function is already at work at the level of intuitive cognition, from which it concludes that mental representation in general does not necessarily require a species pre-existing to the cognitive act. The causation of intellection was a third motive sometimes invoked in favour of positing intelligible species, on the basis of the alleged impossibility for material things to causally affect the spiritual realm. But if there was such an impossibility, Ockham argues, how could the formation of the intelligible species itself ever be explained without falling into an infinite regress: 'just like you accept, he says to his opponent, that a corporeal thing can be the partial cause of a species in the spiritual realm, I accept that a corporeal thing is the partial cause of the intellection in the spiritual realm.'65 This again is basically an argument from parsimony: since the causal impact of material things upon the mind has to be acknowledged anyway, the introduction of species is useless in this respect. And the argumentative pattern is the same for the last two functions Ockham discusses in this context: how the cognitive faculty is determined by its object, and how it is moved by it. 66 Nothing in his whole critique of intelligible species hangs upon either an anti-representationalist bias on Ockham's part or on his wrongly supposing that the intelligible species were intermediate objects of thought in the eyes of their supporters (which, obviously, they were not).67 Ockham,in fact, does not even stress any special drawback of accepting intelligible species (as he will do later on in the case of the fictum), except for useless complications in epistemology. Parsimony, here, is his sole manifest preoccupation. This is also the case in his short critique of any temptation to identify concepts with intelligible species, in the Prologue to his commentary on the Perihermeneias. 68 He mentions three objections there. First, such species are superfluous anyway, the reader being referred to a demonstration of the point elsewhere in Ockham's works (presumably the one we have just been discussing). Second, only acts and habitus should be admitted as real entities within the mind according to Aristotle. And third, it would follow that concepts - and the mental propositions made up from them would remain within the soul even when it is not actively thinking, just like the species are supposed to. All three arguments can be read as pointing to unnecessary epistemological complications, and have nothing to do, at any rate, with alleged threats to direct realism. Eleone Stump mentions still another argument Ockham uses against species: 'No one sees a species intuitively, and therefore experience does not lead us to this [account of cognition]';69 which she interprets as (wrongly) presupposing that in order for the species to be vindicated, its mental reception should be a conscious cognitive act in Ockham's view, as if his reasoning was the following: we have no intuitive grasping of species, therefore they don't exist. The complete argument, however, is quite different, and it is, once more, an economy argument. Ockham's point is that no natural entity should be countenanced without either a good theoretical reason or some empirical evidence. 711 The sentence Stump quotes is simply the part where he is denying that the latter holdpi in this case. It is true, as Dominik Perler has inpiisted,71 that behind the disagreement over intelligible species between Ockham and his opponents, there lay another - more fundamental - divergence. In the last analysis, the point of resorting to intelligible species - in Scotus most notably, but also in Aquinas - was that the external singular things were thought to hide some universal features within them, which had first to be isolated - or extracted - in order for general cognition to take place. Ockham, of course, energetically resisted this realist assumption and argued against it on many occasions. That he shouldn't do so in the context of his discussion of intelligible species, however, reveals that he took the postulation of such species to be superfluous even if these universal features were to be admitted: whichever way it was with ontology, his point is that if abstraction could be reached in intelligible species considered as preconditions for intellectual acts of knowledge, then there is no reason why it could not be reached directly in the intellectual acts themselves, and in the habitus that follow upon them. Whether Ockham correctly understood the theory he was criticizing there, or not, solely depends on whether this theory did in fact postulate intelligible species as prerequisites for intellectual acts of knowledge. There seems to be little doubt that it did. 4. Combining acts Both species and ficta, as we now see, were renounced by Ockham for reasons of economy. In the case of species, he was never even tempted to countenance them, because he couldn't imagine any special job they could be endowed with. Ficta were different in his eyes, because he thought for a while that, despite all their drawbacks, they provided the sole alternative to. common natures for serving as objects of general abstractive acts of cognition. This is a notion he came to abandon when he realized that the only aspect that matters for something to be a concept is its capacity to fulfil certain semiotic functions in mental computations. From then on, he was left with only acts and habitus within the intellect. Not all of them, however, were identified with concepts. Even leaving aside habitus, conceptual acts in his view are but one variety of intellectual acts. They must be distinguished, on the one hand, from intuitive acts, as was explained in chapter 1, and, on the other hand, from propositional and judicative acts, which is what I want to insist upon now. The mind, for Ockham, is basically a combinatorial device. Locating the role of concepts in human thought calls for an elucidation of how intellectual acts combine with each other into more complex units, such as propositions and judgements. I will first briefly present Ockham's distinction between simple and complex items within the mind (section 4.1), and then examine in some details the structure of mental propositional acts, of which concepts can normally be parts (section 4.2). We will see, finally, how judicative acts fit into the picture (section 4.3). 4.1 Simple and complex units Ockham frequently distinguishes between simple and complex mental items. He docs it explicitly. for example, at the outset of the Ordinatio: 'among the acts of the intellect. he says there... one is the act of apprehension, and it pertains to whatever l:an tcrminate an act of thc intcllective power. whether this should ht' ('omph'x or

22 32 INTELLECTUAL ACTS 33 incomplex'.72 This passage, admittedly, was written from the point of view of the fictum-theory of concepts, which Ockham still held at the time: the distinction drawn here between complex and incomplex items separates different kinds of possible objects for intellectual apprehensive acts. When he moved to the actus-theory, he kept the distinction, except that from then on, it was the intellectual acts themselves that he saw as either complex or incomplex: when he speaks of incomplex conceptual terms in the Summa Logicae, or of simple or composite concepts in the Quodlibeta,73 he is now referring to intellectual acts. The criterion for the distinction, nevertheless, remains the same. It corresponds to the one Ockham gives in his commentary on the Categories when explicating the strict sense of the term 'incomplex' as applied to words or signs (dictiones in this case): 'Strictly speaking, an incomplex item is a simple word, one single word, that is, without the addition of other words, such as 'man', 'runs', 'lion', 'goat'. And we call 'complex', by contrast, whatever is composed of several words.'74 A simple sign is a sign no part of which is itself a sign. A simple- or incomplex - intellectual act, by transposition, is an intellectual act no part of which is itself an intellectual act. A complex one is an intellectual act that has other intellectual acts as proper parts. 75 It must be insisted that this way of drawing the distinction does not prevent simple conceptual acts from having other kinds of complexity, as long as they don't have intellectual acts as their proper parts. 76 It does not prevent them, in particular, from having some sort of semantical complexity. Ockham, indeed, is quite ready to admit that a simple intellectual act, whether intuitive or abstractive, can have a plurality of different objects all at once,77 or that a simple concept can simultaneously signify different things in d!fferent ways.78 This is a point that will turn out to be of utmost importance in the discussion of connotative concepts that will occupy us in chapters 4 to Propositional acts Some composite mental acts are non-propositional: the combination of the concept 'white' with the concept 'horse', for example, yields a complex conceptual act which is not yet a mental proposition. Other combinations, by contrast, involve a mental copula and are - paradigmatically - true or false. These are the mental propositions proper. In their most elementary form - corresponding to what Ockham calls 'categorical' propositions, tqey are composed of a subject, a copula, and a predicate,79 each one of which being either complex or incomplex. 80 More complex propositional acts can then be produced by combining such categorematical propositions with each other with the help of connecting syncategoremata such as 'and', 'or', 'if', 'when', 'because', and so on. 81 The mind, in this doctrine, is taken to be innately endowed with a capacity for some simple acts (intuitions and simple concepts, basically), and with a further set of recursive capacities for combining in various (and non-arbitrary) ways whatever intellectual acts it produces into more complex ones. A problem that was intensely discussed in the fourteenth century about propositional acts is how they can have ordered parts. Gregory of Rimini, for one, is well known for having argued, in the early D40s, that this is simply impossible, since a mental propositional act is not supposed to be spread out in space nor in time. From which he concluded that mental propositions have to be simple acts themselves: human thought, in Gregory's view, is not intrinsically compositioha1. 82 This is a problem Ockham was well aware.of. He discussed it in some detail in his commentary on the Perihermeneias and in his Quaestiones in Libros Physicorum, some twenty years before Gregory's lecture on the Sentences. 83 In both texts - which are very dose to each other - he sees the difficulty as stemming not from the mereological structuration in itself, but from the apparent necessity for an order among the parts of a mental proposition if such parts are to be admitted. How could the mind distinguish, otherwise, between, for example, 'all men are animals' and 'all animals are men', which, apparently, have the same parts, but obviously not the same truth-conditions? A spatial ordering of parts, such as the one we find in written sentences, is excluded, of course, since the mind is not itself extended in space. And so is a temporal ordering, such as the one we have in spoken sentences, since the parts of a mental proposition, whatever they are, should be simultaneously present to the mind somehow. Ockham's solution to the riddle is radical: it amounts to renouncing linear ordering altogether within mental propositions. This, he thinks, can be done in two ways. First, we could say that a mental proposition is in fact a simple cognitive act (just as Gregory of Rimini will), but that it is equivalent to an ordered combination of terms such as the corresponding one in spoken language. 84 And second, we could admit that the parts of such non-equivalent propositions as 'all men are animals' and 'all animals are men' are not the same, since a part of the former is the complex act corresponding to 'all men', which is not a part of the. latter, and a part of the latter is the complex act corresponding to 'all animals', which is not a part of the former. 85 It might seem at first glance that these two possibilities are exclusive alternatives for Ockham, among which he simply was not ready to choose when he wrote his commentary on the Perihermeneias and his Questions on the Physics. A closer reading of the two passages, however, reveals that he took the two alternatives to be simultaneously acceptable. His point is that'some mental propositions are composed of a subject, a predicate, and a copula, while some are equivalent to such combinations'.86 The admission of the former sort of mental propositions neatly demarcates Ockham's approach from the one Rimini will later favour. Human thought, for Ockham, is basically compositional, as he explains in detail in the Summa Logicae. 87 That mental propositions, for him, can sometimes occur as simple acts within the mind is a handy - and fascinating - possibility, no doubt, but it remains a merely parasitic device: we can form, so to say, unanalysed abbreviative acts which, for some purposes, are functionally equivalent to more complex ones (in truth-conditions, for example), but the outreaching strength of human knowledge and reasoning - the possibility of a full-fledged science in particular - ultimately hangs upon this fundamental and remarkable capacity we have for combining mental acts into more complex ones in various ways. The order-problem is avoided, nevertheless, since composite mental acts do not need, in Ockham's eyes, to be linearly ordered. Various intellectual acts of the same ~eneral sort (propositional acts, for example) are differentiated from one another by the nature of their parts, not by their order. The relevant feature for Ockham's solution here is Ihal an immediate part of a complex menial aci clin itself be

23 34 INTELLECTUAL ACTS 35 complex: the complex concept 'white horse', for example, is a distinct intellectual act which can occur as such as an immediate part of a mental proposition. And the same is true when mental syncategoremataare involved. 88 All this supposes, admittedly, that the various parts of a complex mental act - and even the parts of its parts - can all be simultaneously present to the mind. But Ockham never saw that as a problem. He thought, on the contrary, that the possibility of simultaneous intellectual acts had to be acknowledged by any sound theory of mental activity,89 and that it was, at least implicitly, by Aristotle's.9o His point was simply that the complexity of mental discourse does not require a linear ordering of the sort we observe in spoken or written discourses. The relevant idea is that some parts of 'all men are animals', for example, have no equivalent in 'all animals are men'. The former, in effect, involves six parts altogether, according to Ockham's complete analysis: (1) 'all' (a simple syncategorematic act), (2) 'men' (a simple conceptual act), (3) 'all men' (a complex conceptual act having (1) and (2) as its parts), (4) 'are' (a simple copulative act),9 I (5) 'animals' (another simple conceptual act), (6) 'all-men are animals' (a complex propositional act having (3), (4) and (5) as its parts). The whole structure can be represented as a tree: (6) ffi (3) (4) (5) A (1) (2) The important thing, however, is that some of its components do differ from those of 'all animals are men', which does not have (3) as a pl;lrt, and involves, therefore, nothing like (6). The underlying thesis is that when two complex intellectual acts are of the same sort (two propositional acts, for example), their non-equivalence - if any -ultimately depends on some of their respective parts not being equivalent to each other. A non-linear arborescent order ultimately results from the very identity of the parts. For a complex arrangement of mental terms to be a proposition, it is sufficient that it be composed of at least two categorematic concepts and a copula. 92 Any such whole ipso facto has truth-conditions 93 ~ which is the distinctive mark of propositional acts - and can be connected within the mind with an act of assent - or dissent - to form a complete judgement. 4.3 Judicative acts While the main function of conceptual acts in human thought is to contribute in specific ways to the truth-conditions of mental propositional acts, the latter in tum are characterized by their functional roles with respect to mental units of a further sort, which are more immediately linked with behaviour: judicative acts. These are usually contrasted, in Ockham's texts, with merely apprehensive acts, the latter not yet involving any special commitment of the thinking agent toward the truth-value of propositions. 94 At the time. of his lectures on the Sentences, when Ockham was still supporting the fictum-theory of concepts, this distinction was drawn in a quite straightforward manner: the apprehension of a complex propositional content was identified with its very formation within the mind,95 and it was taken to be presupposed by the judicative act of assentor dissent,96 In the Quodlibeta, however, the subject gets more complicated. Ockham now distinguishes between two sorts of apprehensive acts having to do with propositions, and two sorts of judicative acts as well. The first variety of propositional apprehension is, as in the earlier approach, the mere formation of the mental proposition, in other words the propositional act itself; in the actus-theory now favoured, it usually has conceptual acts as its intrinsic parts. And the second one is the act of self-consciously considering - or cognizing - a previously formed such mental proposition;97 it is, in Ockham's vocabulary, a reflexive act, that is, a mental act with another mental act as its object,98 The first form of judicative act, on the other hand, consists in the ordinary non-reflexive acceptance that so and so is the case. Not being reflexive, it does not, properly speaking, have a propositional act as its object, but it simply follows upon the formation of some propositions, without any conscious grasping of them. 99 It can be correctly described as having a propositional content ~ since it is, indeed, an assent that so and so is the case - but not a propositional object: it is not an assent to a proposition. loo In so far as it can be said to have objects at all, these will be the cxternal things that the judgement pertains to. 101 The judicative act is a simple act in this case, but it can be said to be functionally equivalent in many respects to a complex one in that it involves somehow an apprehension of a plurality of external things. 102 Judicative acts of the second sort, by contrast, presuppose the reflexive apprehension of a propositional complex, and are reflexive therefore, in so far as their objects are mental acts (propositional acts in this case).103 The highest sort of such judicative act, for Ockham, is the act of scientific knowledge (actus scientiae or actus sciendi) in the strongest sense of the phrase, this knowledge, that is, by which some necessary truth is evidently known as such as the result of a valid inference from necessary premisses. 104 Since Oc\<ham indicates that the objects ofsuch science must be propositions,105 it must be concluded'that scientific acts of knowledge, in the strong sense, are reflexive judicative acts, according to his mature doctrine. The layman - the laicus of Quodlibeta III, can unreflexively l'ntertain some beliefs, ofcourse - and even true ones - but only the thinker who selfconsciously and reflexively considers within himself general propositions, really (;omposed of conceptual acts as their parts, can reach the highest degree of knowledge. Ockhamistic conceptual acts, in short, are real mental episodes elicited either by Ihe intuitive grasping of objects in the environment or by the activation of SOme

24 36 INTELLECTUAL ACTS 37 previously acquired dispositions - or habitus. They are either simple (if they have no other intellectual acts as their parts) or complex (if they do). Whenever they occur, they leave a trace within the mind, which is a new conceptual disposition, or the strengthening of a previously existing one. Along with intuitive and syncategorematic acts, concepts can be intrinsic components of propositional acts and be associated, through this, with judicative acts. Propositional and judicative acts, in tum, also cause the mental formation of corresponding habitus. In the case of interrelated scientific acts of knowledge, in particular, the various resulting habitus can form a cluster - or a 'collection' - within a given mind, thus constituting the possession of a science - grammar orarithmetic, for example - by this particular agent. 107 All these acts and habitus are singular qualities in Ockham's ontology. Intelligible species were discarded as superfluous, and so were, in the mature theory, the ideal objects of thought - the ficta - that had previously seemed to be required to avoid countenancing common natures. In the last analysis, the distinctive feature of conceptual acts and habitus in the mental machinery, according to Ockham, is that they function as general signs in intellectual computations. Notes 1. Hornsby 1998, p See, for example, Quaest. Var. q. 7, art. 1, OTh VIII, p See, for example, Quaest. Var. q. 6, art. 9, OTh VIII, pp , where Ockham discusses the relation of such appetitive acts with pleasure and sadness. 4. See, for example, Quaest. Var. q. 7, OTh VIII, pp (EngI. transi. in Wood 1997), where Ockham lengthily discusses the relation of virtues and vices with various sorts of volitive acts. 5. Cf. ibid., p. 364: '... actus intellectus non est primo virtuosus nec est in potestate voluntatis'. ' 6. Aristotle, De Anima II, I, 412a22-23, transi. Hamlyn 1993, p See, for example, the handbook called the Auctoritates Aristotelis, ed. Hamesse 1974, p. 177: 'Duplex est actus, scilicet primus et secundus: primus ut scientia, secundus Ut speculari secundum scientiam.' 8. See Rep. III, q. 12, art. 2, OTh VI, p. 397: '... dico quod actus est causa efficiens respectu habitus'; also Quaest. Var., q. 7, art. 1, OTh VIII, pp See Quodl. III, q. 21, OTh IX, pp.285-8: 'Utrum habitus sit causa effectiva actus.' 10. Ibid., p. 286: '... dico quod primus actus non causatur ab habitu sed,ab aliis causis'; please note that the phrase 'primus actus' here means 'original act' rather than primary act - or habitus - in the Aristotelian technical sense introduced above. 11. We can now see why Ockham postnlates the causing of an abstractive act by every intellectual iptuitive act: this is required to explain the formation of the habitus in his theoretical framework. If the intuitive act was posited as the direct cause of a habitus, this habitus would have to incline to new intuitive acts, which is contrary to experience: the habitus I acquired from seeing horses does not cause in me new intuitive graspings of horses when there are none around. See on this Rep. II, q , OTh V, pp See, for example, Rep. III, q. 12, OTh VI, p. 403: '... omnis habitus praecise inclinat immediate ad actus consimiles ex quibus generatur'; also Rep. III, q. 7, OThVl, p Rep. III, q. 7, OTh VI, p. 205: 'Unde habitus est causa actus, sed non eiusdem numero a quo causabatur, sed est causa alterius actus. Similiter, adus est causa habitus, sed non illius a quo causabatllr: sed vel alterills habitus vel alterills gradus in habitu.' 14. Ibid., pp. 197-' Ibid., p. 198: '... quilibet experitur in se quando delectatur vel tristatur.' 16. Aquinas develops in many places this theory of the concept - or mental word - as distinct from the intellectual act. See, for example, his QuaesL disp. de potentia, q. 8, art. 1, p. 252 (ed. Marietti): 'Differt autem [conceptio] ab actione intellectus; quia praedicta conceptio [N.B.: Aquinas also frequently uses 'conceptus' or 'verbum mentis' as equivalents] consideratur ut terminus actionis, et quasi quoddam per ipsam constitutum.' The doctrine has been extensively discussed both in the late Middle Ages and in modem literature; for further details and references, see Panaccio 1992b, 1999a, chap. 6 and Good expositions of thefictum-theory by Ockham himself are found in Ord. I, dist. 2, q. 8, OThII, pp , and Exp. in Perih., Prooemium, 7 and 10, OPh II, pp and For lucid modem presentations, see Adams 1987, pp , and Karger 1994, pp See Ord. I, dist. 2, q.8, OTh II, pp : 'Ideo potest aliter dici probabiliter quod universale [= conceptus]... tantum habet esse obiectivum in anima et est quoddam fictum...';or Exp.in Perih., Prooemium, 7, OPh II, p. 360: 'Potest enim [conceptus mentis] vocari intentio animae pro eo quod non est aliquid reale in anima ad modum quo habitus est aliquid reale in anima, sed habet tantum esse intentionale, scilicet esse cognitum, in anima.' 19. For more on Burley's rejection of the concept asfictum (or idolum), see Panaccio 1992b, pp , and 1999a, pp Good recent studies of Burley's ontological realism are found in Karger 1999b, Conti 2000, andcesalli Ockham's detailed critique of the various forms of realism of universals he was familiar with is found in Ord. I, dist. 2, q. 4-7, OTh II, pp. 99~266; with a cameo version in SL 1,15-17, OPh I, pp Karger 1996, p Gal 1967 showed in effect that some passages of Ockham's own critique of thefictumtheory in Quodl. IV, q. 35 are almost literally borrowed from Chatton. Pretty strong conclusions were generally drawn from this. See, for example, Adams 1987, p. 84: 'It was under the influence of criticisms from his contemporary and fellow Franciscan, Walter Chatton, that Ockham eventually abandoned the objective-existence theory...'; or Kelley 1981, p. 223: 'As we know, there is good reason to think that Ockham's final stand on the question amounted to a complete agreement with Chatton...' Chatton's original objections are listed in his Reportatio I, dist. 3, q. 2 (ed. in Gal 1967, pp and in Weyand Etzkorn 2002, pp ). Further discussions of theflctumtheory by Chatton are to be found in the Prologue, q. 2, art. 2 of his Lectura super Sententias (ed. in O'Callaghan 1955, p. 241 and in Wey 1989, p. 87), and, more extensively and more directly to the point, in Book I, dist. 3, q. 2 of the same work (ed. in Kelley 1981, pp ), but the Lectura is from according to Wey 1989, p. I, while Ockham's final adoption of the actus-theory is known to have occurred around Even ifsome parts ofchatton'slectura, orthe whole ofit, were written earlier (as suggested by Brown 1985), this would not substantially affect the main points I want to make here (although, admittedly, there are many complex connections between Ockham's and Chatton's texts which still require an in-depth examination). 2:1. Quaest. in Phys., q. I, OPh VI, pp The arguments directly borrowed from Chatton are the fourth and fifth in this list (also corresponding to the fourth and fifth arguments of a set of five put forward by Ockham in a more explicit form in Quod!. IV, q. 35. O'/'h IX, pp ). 24. See OPh II..Introduction '. p. 22*. 25. Those are arguments 6 and' 7 in QUII/'.I'I. ill Pln'.I'., q. I, OPh VI. p. WK. They had originally been proposed by Ockham in HI!'. ill Paih.. Prooemium. 7. OPh II. pp.

25 38 INTELLECTUAL ACTS It is true that he had then considered possible replies (see OPh II, p. 370: 'Et qui vellet tenere istam opinionem [that is, thefictum-theory], possetrespondere...'), but since he repeats the objections, but not the replies, in Quaest: in Phys., q. 1, he must ultimately have found these.replies unconvincing. 26. It should also be noted that Adam Wodeham, who must have personally followed the debate between Ockham and Chatton when he was a student, and who had been quite close to Ockham at the time, later held that, although the concept is indeed to be identified with a cognitive act rather than with a.fictum, none of Chatton's original arguments against Ockham's former fictum-theory was conclusive (see his Lectura Secunda in I Sent., dist. 8, q. 1, ed. Wood and Gal 1990, vol. III, pp ). 27. Quodl. IV, q. 35, OTh IX, p. 474: '... per actum potest sa1vari quidquid sa1vatur per fictum' (transl. Freddoso and Kelley.1991, vol. I, p. 390). See also SL 1,12, OPh I, p. 43: 'Omnia autem quae salvatur ponendo aliquid distinctum ab actu intelligendi possunt salvari sine tali distincto... ' 28. This is the gist of argument 6 in Quaest. in Phys. q. 1, OPh VI, p. 398: '... difficile est imaginari aliquid posse intelligi intellectione reali et tamen quod nee ipsum nee aliqua pars sui potest esse in rerum natura nee esse substantia nee accidens'. The same argument - in pretty much the same words - is the first one adduced against the fictumtheory in Exp. in Perih., Prooemium, 7, OPh II, p This is argument 5 both in Quodl. IV, q. 35 and in Quaest. in Phys., q. 1. As was usual in Ockham's times for arguments having to do with what was later called analyticity, it is formulated in terms of what God could do without contradiction; in the Quodlibeta's formulation, OTh IX, p. 473: '... non est contradictio quod Deus faciat cognitionem universalem sine tali ficto... ' The argument in this case is directly inspired by Chatton (see Gal 1967, p. 199, and Weyand Etzkorn 2002, p. 234). 30. This is argument 2 in Quodl. IV, q. 35 and in Quaest. in Phys., q. 1. In the former's formulation: '... tale fictum impediet cognitionem rei... quia '" est... quoddam tertium medium inter cognitionem et rem'. A similar argument had been standardly used since Peter John Olivi in the polemics against Aquinas's theory of the mental word (see Olivi's Tractatus de Verbo 6.2.3, ed. in Pasnau 1993, p. 144: '[tale verbum] potius esset ad impedimentum'; on the use of the argument by other authors between Olivi and Ockham, especially William of Ware and Walter Burley, see Panaccio 1999a, pp ). Chatton puts forward a similar argument againstpeter Aureoli in the Prologue ofhis Lectura (Wey 1989, p. 87), but only in connection with intuitive knowledge, which is irrelevant for Ockham's theory of the jictum which had to do only with general concepts. 31. Adams 1987, p This is argument 7 in Quaest. in Phys., q. 1, OPh VI, p. 398: '... tale idolum plus differret a re quacumque quam quaecumque res differtab alia... et ideo tale idolum minus assimiletur rei... ' It had first been put forward by Ockham in the Exp. in Perih., Prooemium, 7,OPh II, pp , but is not found in. the Quodlibeta version. 33. This is argument 3 in Quodl. IV, q. 35 and in Quaest. in Phys., q. 1. In the Quodlibeta's formulation (OTh IX, p. 473): '... ab aeterno erat coordinatio tot entium fictorum quot possunt esse diversae res intelligibiles, quae fuerunt ita necesse est, quod Deus non potuiteadestructere...' It is also found in Chatton's Prologue to the Lectura (Wey 1989, p. 87), but who devised it first is unknown. 34. See Exp. in Perih., Prooemium, 10, OPh II, pp On Chatton's connection to Scotus, see Gal 1967, p. 191, and Way and Etzkorn 2002, p. IX. Some of Chatton's dicta on concepts are telling in this regard, for example in Rep. I, dist. 3, q. 2 (in Gal 1967, p. 2(9): '... conceptus iste speciti<:us [that is. the concept of man] hahet humanitatem pro ohiecto primo.' This is something Ockham could not very well accept. 36. This is argument 1 in Quodl. IV, q. 35, OTh IX, p. 472, and Quaest. in Phys., q. 1, OPh VI, p In SL I, 12, OPh I, p. 43, Ockham uses a more familiar version of the Razor in support of the actus-theory: 'frustra fit per plura quod potest fieri per pauciora'. On Ockham's general use of this principle, see Maurer This is the gist of argument 4 in Quodl. IV, q. 35, OTh IX, p. 473 and in Quaest. in Phys., q. 1, OPh VI, p. 397; in the latter's formulation: '... sine tali ficto potest sufficienter haberi subiectum et praedicatum'. 39. See Gal 1967, p. 198 and Weyand Etzkorn 2002, pp Note that a term can have a supposition, for Ockham, without actually suppositing for something; it might simply purport to refer. See SL 1,72, OPh I, pp Exp. in Perih., Prooemium, 7, OPh II, p See Quodl. IV, q. 35, OTh IX, p. 474: '... per actum potest salvari quidquid salvatur per fictum, eo quod actus est similitudo obiecti, potest significare et supponere pro rebus extra, potest esse subiectum et praedicatum in propositione, potest esse genus, species, etc., sicut fictum'. The corresponding passage in Summa Logicae I, 12, OPh I, p. 43, mentions only signification and supposition, these being, in the last analysis, the really basic functions of the concept: '... supponere pro alio et significare aliud ita potest competere actui intelligendi sicut alii signo. Igitur praeter actum intelligendi non oportet aliquid aliud ponere.' Biard 1989, pp , has.rightly insisted on the centrality of the idea of the concept as a sign in Ockham's final adoption of the actus theory. 43. Exp. in Perih., Prooemium, 7, OPh II, p. 361: '... praedicta [that is, representing, suppositing, being part of a proposition... ] propter alias rationes non possunt competere intellectioni, quia difficile est salvare quid intelligam tali intellectione...'; what Ockham calls 'intellectio' in these lines is the act of intellection. See also Ord. I, dist. 2, q. 8, OTh II, p. 268, where the two arguments adduced against the actus-theory basically amount to the same point. 44. Ibid.: 'tamen omnia ilia [that is, representing, etc.] possunt verius competere alicui qualitati exsistenti in anima, quae terminat actum intelligendi' (the italics are mine). 45. See Ord. I, dist. 2, q. 4, OTh II, pp See Lindbergh 1976, Tachau 1988, chap Ockham's main discussion ofthe intelligible species theory is in Rep. II, q , OTh V, pp For his discussions of the species in medio and of the sensible species, see Rep. III, q. 2 and 3, OTh VI, pp For instance in Tachau 1988, Tweedale 1990, Spruit (Vol. 1, 1994), Perler 1996a, Pasnau 1997, or Stump See Spruit (VoLl, 1994), p. 295, Pasnau 1997, pp , and Stump 1999, passim. Perler 1996b, p. 280, approves ofthis diagnosis: 'Spruit', he writes, 'correctly points out that Ockham's critique [ofthe inteliigible species] was based on a misunderstanding.' 50. Rep. II, q , OTh V, p Spruit (VoLl, 1994), p In Rep II, q , Ockham t1rst presents Aquinas's version ofthe theory (probably, as the editors point out, through the exposition of it by some disciple - OTh V, pp , n. 2), and then briefly remarks that Scotus too subscribed to it, although for different reasons, which Ockham does not even care to mention or discuss in this context (OTh V, p.256). 5~. See in particular Aquinas's Summa Theologiae I, q. 84-6, and his Expositio in Aristotelis De anima 54. Rep. II, q , OT!I V, p. 253: 'Hic primo suppono quod species sit illud quod est praevium actui inlelligendi et potest manere ante intellectiones et post, etiam re ahsente. Et per consequens dislinguitur ah hahitu. quia hahitus intellectus sequitur actum intelligendi. sed species prael'edit tam aclum quam hahitum.'

26 40 INTELLECTUAL ACTS Thomas Aquinas, Quaest. disp. de potentia, q. 8, art. 1: 'Differt autem [conceptio] a specie intelligibili: nam species intelligibilis, qua fit intellectus in actu, consideratur ut principium actionis intellectus...;' the same point is made, for example, in Summa contra Gentiles I, 53, and in Quodl. V, 5, Pasnau 1997, p: Ibid., p Cf. Thomas Aquinas, Summa contra Gentiles I, 46: 'Species intelligibilis in intellectu praeter essentiam eius existens esse accidentale habet.' 59. Rep. II, q , OTh V, pp Ibid., p. 273: 'Sed isto modo assimilatur intellectus suffieienter per intellectionem causatam ab obiecto et receptam in intellectu, igitur non requiritur species.' 61. Tachau 1988, p See Rep. II, q , OTh V, p. 273: 'Item, tanta assimilatio requiritur in notitia intuitivaquanta in abstractiva; sed in intuitiva non requiritur 'aliquid, praevium cognitioni, assimilans; igitur nee in abstractiva.' On Ockham's acceptance of cognitive assimilation for both intuitive and abstractive knowledge, see also OTh V, pp Rep. II,q; 12-13, OTh V, p. 274; the point about intuitive cognition had been made a few pages before (p. 268). 64. Ibid., p. 274: '... sicut obiectum sufficienter repraesentat se in una cognitione, ita in alia quae immediate sequitur intuitivam.' 65. Ibid., p. 275: '... sicut tu ponis quod corporale potest esse causa partialis ad causandum speciem in spirituali, ita ego pono quod corporale est causa partialis ad causandum intellectionem in spirituali.' 66. Ibid., pp Aquinas, for one, frequently insists that the intelligible species are not normally the objects of intellectual cognition, but the means of it, not the id quod but the id quo of intellection. See, for example, In De Anima III, 2: 'Manifestum est etiam quod species intelligibiles quibus intellectus possibilis fit in actu, non sunt obiectum intellectus. Non enim se habent ad intellectum sieut quod intelligitur, sed sieut quo intellectus intelligit... ' This is a point Ockham could hardly have missed. He uses it actually in Ord. I, dist. 2, q. 8, OTh II, p. 269, to show that intelligible species, if they were admitted, could not be identified with universals, since universals are generally supposed to be the objects of general intellection while the species are not. 68. Exp. in Perih., Prooemium, 5, OPh II, pp. 35D-51. It should be noted that the thesis Ockham discusses in this passage - namely that concepts are intelligible species - is not Aquinas's- who draws, on the contrary, a clear distinction between concepts and intelligible species (see above n. 55), but nothing suggests that Ockham thought it was. Ockham's editors mention Roger Marston as a possible candidate here (OPh II, pp , n. 1). 69. Rep. II, q , OTh V, p. 268: '... sed nullus videt speciem intuitive; igitur experientia non inducit ad hoc'; the sentence is quoted in translation and discussed in Stump 1999, p Ibid.: 'Item, nihil est ponendum necessario requiri naturaliter ad aliquem effectum nisi ad iliud inducat ratio certa procedens ex per se notis vel experientia certa; sed neutrum istorum inducit ad ponendum speciem, igitur etc.' This passage almost immediately precedes the sentence quoted in the previous footnote. 71. Perler 1996a, especially pp Ord., Prol., q. 1, OTh I, p. 16: '... inter actus intellectus... unus est apprehensivus et est respectu cuiuslibet quod potest terminare actum potentiae intellectivae, sil'c sit complcxum sil'c incomplexum' (italics are mine). 71. For example. in.'ii, I. 3. OPh I. p. II; or Quodl. II..l. OJ''' IX. p Exp. in Praedic., 4, 1, OPh II, p. 148: 'Striete dicitur incomplexum simplex dictio, hoc est una dictio sine additione alterius dictionis, sieut hie: homo, currit, leo, capra. Et per oppositum dicitur complexum omne compositum ex diversis dietionibus.' 75. In the context of a discussion with Ockham, Walter Chatton gets explicit about this, by assuming that a simple act, for both of them, is one 'that does not include several acts' (actus simplex non ineludens plures actus); this is in the Prologue of his Lectura, q. 1, art, 1, ed. Wey 1989, p Ockham makes a similar point about simple sentences in his Exp. in Perih. I, 4, OPh II, p. 395: 'Intelligendum quod enuntiatio categorica dicitur simplex non per carentiam cuiuscumque compositionis et quarumcumque partium...; sed dieitur simplex quia non componitur ex pluribus enuntiationibus.' 77. See, for example, Rep. II,q , OTh V, pp. 28D-81, where Ockham admits that one can have a simple intuitive grasping ofa whiteness and a blackness simultaneously, and then form a simple abstractive act pertaining to the same objects. 78. See, for exemple, Ord. I, dist. 3, q. 2, OTh II, p. 405: 'Tertio, dico quod Deus potest cognosci a nobis in conceptu simplici connotativo et negativo sibi proprio..., et iste conceptus est simplex, quamvis distincta significet, sive principaliter sive secundario, hoc est, vel in recto vel in obliquo... ' 79. See SL II, 1, OPh I, p. 241: 'Propositio categorica est illa quae habet subiectum et praedieatum et copulam, et non includit plures tales propositiones'; see also Exp. in Perih. I, 4, OPh II, p That the copula, and not only the subject and predicate, can sometimes be complex, according to Ockham, can be gathered, for example, from his treatment of such phrases as 'begins to be' (incipit esse) or 'ceases to be' (desinit esse) (SL I, 75, OPh I, pp ), and from his discussion of truth-conditions for modal propositions without dictum (SL II, 10, OPh I, pp ). On Ockham's admission of such special copulas as 'is believed by you to be', see Panaccio, forthcoming, section Ockham calls all such complex propositions 'hypothetical' (see SL II, 1, OPh I, p. 241: 'Propositio hypothetica est illa quae ex pluribus categoricis est composita'); they include conjunctive, disjunctive, conditional, causal and temporal propositions. 82. See Gregory of Rimini, In I Sent. Prol., q. 1, art. 3, ed, Trapp et al. 1978, pp See Exp. in Perih. I, Prooemium, 6, OPh II, pp ; and Quaest. in Phys., q. 6, OPh VI, pp See Exp. in Perih., Prooemium, 6, OPh II, p. 356: 'Ad istud potest dici quod propositio potest esse actus intelligendi aequiva1ens toti uni propositioni compositae ex realiter distinctis, si talem ordinem haberent qualem habent in voce.' 85. Ibid.: 'Aliter posset dici quod... in propositione in mente correspondente isti propositioni prolatae "omne animal est album" correspondet unus actus tamquam pars propositionis isti toti "omne animal" et alius actus isti "album", sed in propositione in mente correspondente isti propositioni "omne album est animal" correspondet unus actus isti toti "omne album" et alius isti termino "animal".' X6. Quaest. in Phys., q. 6, OPh VI, p. 410: 'Propositio autem mentalis aliqua componitur ex subiecto et praedicato et copula, et aliqua erit aequivalens tali composito' (the italics are mine). The same is said, albeit a little more tentatively, in Exp. in Perih., VPh II, p. 358: 'Sed de propositione concepta, quae tantum est in mente, potest dici quod aliq!1a...' X7. See in particular SL I, 1-4, where the real composition of mental propositions is explicitly stated and analysed. xx. This approach can, of course, be further exploited. Spade 1996, pp , quotes an anonymous laic fourteenth-century text mentioning a doctrine according to which for something 10 be the suhjecl of a mental proposition, it does not have to OCt'ur al a special position in a given order. hut it must. instead. he accompanied hy a special mental,n'/i( /III'X0/'('lI/a. whidl identities it as Ihe suhject. and so 100 for the predicate, I'a/'l'

27 42 INTELLECTUAL ACTS 43 Spade, this notion appears to be but a rather natural elaboration on Ockham's ideas in the texts we have just been discussing. 89. See, for instance, Ord. I, Prol., q. 1, OTh I, pp , where Ockham argues that Plato can all at once love Socrates and know that he does, which requires the simultaneity of at least three mental acts: loving Socrates, knowing Socrates, and knowing about one's love for Socrates, two of which are cognitive acts. Also: Quodl. IV, 17: '... falsum [est] quod partes conceptus compositi succedunt sibi in mente sicut in voce; quia in mente conceptus illi simul existunt in subiecto indivisibili, in voce autem non.' 90. See Exp. in Perih., Prooemium, 6, OPh II, p. 358: 'Nec umquam invenitur ab Aristotele quod negaret duos actus intellectus posse simul esse in intellectu...' 91. In Rep. II, q , OPh V, p. 280, Ockham holds that the intellectual act corresponding to the copula is always complex: 'Sed actus terminatus ad copulam est complexus quatenus terminatur immediate ad totum complexum.' But this was written when Ockham favoured the fictum-theory: the mental act 'terminating' in the copula is not here identified with the copula itself (see also his early discussion of the copula in Rep. II, q. 1, OPh V, pp ; Karger 1996, pp , offers a good study of these passages). His later formulations, in the Summa Logicae in particular, generally presuppose the simplicity of (at least some) copulative acts. See, for instance, SL I, 3, OPh I, p. 11, where he admits that there are verbs (and hence copulas) among incomplex mental acts (but see above n. 80). 92. This condition, however, is not necessary since Ockham, as we saw, accepts the possibility of simple propositional acts that are merely 'equivalent' to complex combinations of terms (see above n. 84 and 86). 93. For a further discussion of this point, see below chap. 8, sect See Ord. I, Prol., q. 1, OTh I, p. 16: 'Alius actus potest dici iudicativus, quo intellectus non tantum apprehendit obiectum, sed etiam illi assentit vel dissentit.' 95. As Karger 1996, p. 226, n. 105, remarks, this identification is riot explicitly asserted in Ockham's earlier works, but it is presupposed. See, for example, Rep. III, q. 2, OTh VI, p. 85, where he refers to 'the apprehension or formation of a complex unit' (apprehensionem sive formationem complexi). 96. See Ord. I, Prol., q. 1, OTh I, p. 17: '... actus iudicativus respectu alicuius complexi praesupponit actum apprehensivum respectu eiusdem.' 97. Quodl. V, q. 6, OTh IX, p. 501: '... duplex estapprehensio: una quae est compositio et divisio propositionis sive formatio; alia est quae estcognitio ipsius complexi iam formati...' 98. See Quodl. II, q. 12, OTh IX, p. 165, where Ockham distinguishes reflexive acts from direct ones: '... vocatur actus rectus quo intelligimus obiectum extra animam, et actus reflexus quo intelligitur ille actus rectus.' 99. See Quodl. III, q. 8, OTh IX, pp : 'Loquendo de primo assensu, dico quod ille actus non habet complexum pro obiecto; tum quia ille actus potest esse per solam formationem complexi et sine omni apprehensione complexi...; tum quia laicus sciens quod lapis non est asinus, nihil cogitat de propositione... Licet assentiat et sciat quod sic est in re vel non est, mediante propositione formata in intellectu, tamen hoc non percipit...' 100. Ibid., p. 234: '... dico quod proprie loquendo, non debet dici quod aliquid scitur isto actu, sed quod isto actu scitur quod lapis non est asinus...' See also Quodl. IV, q. 16, OTh IX, pp , and Quodl. V, q. 6, OTh IX, p. 500, where the same distinction is drawn Quodl. III, q. 8, OTh IX, p. 234: '... iste actus [namely the non-reflexive judgement that a stone is not a rabbit] habet res extra pro obiectis, pula lapidem et asinum... ' 102. Ibid., p. 234: '... et iste actus aequivalet quantum ad multa ali<.:ui complexo quo aliquid scitur.' 103. Ibid., p. 234: 'Sed loquendo de secundo actu sciendi vel assentiendi, dico quod ille proprie est actus complexus habens pro obiecto complexum'; and Quodl. IV, q. 16, OTh IX, p. 377: 'Sed secundus assensus, naturaliter loquendo, necessario praesupponit apprehensionem complexi... ' 104. See Exp. in Phys., Prol., 2, OPh IV, p. 6: 'Quarto modo dicitur scientia notitia evidens veri necessarii nata causari ex notitia evidenti praemissarum necessariarum applicatarum per discursum syllogisticum.' The term 'science', of course, can also be taken in less stringent ways, as Ockham explains in the same text, to designate, for example, the certain knowledge of any old truth, or the evident knowledge of a truth, or again the evident knowledge of a necessary truth (ibid., pp. 5-6) Ibid., p. 9: 'Nam obiectum scientiae est tota propositio nota ~.. ' Please note that this was most probably written when Ockham had already turned to the actus-theory ofconcepts, and drawn the distinction we commented above between two sorts of judicative acts in Quodl. III, See above nn See Exp. in Phys., Prol., 2, OPh IV, p. 6: '... scientia... aliquando acclpltur pro collectione multorum habituum ordinem determinatum et certum habentium.... Et scientia isto modo comprehendit tamquam partes aliquo modo integrales habitus principiorum et condusionum, notitias terminorum, reprobationes falsorum argumentorum et errorum, et solutiones eorum. Et sic dicitur metaphysica esse scientia et naturalis philosophia esse scientia, et ita de aliis.'

28 Chapter 3 Concepts as Signs Generality for Ockham is but a matter of signification, and concepts are the prime bearers of it. That concepts should be signs, therefore, is a crucial requirement of Ockham's nominalism. What it means, however, is not self-evident. In his 1994 book, Nominalisme, the French scholar Cyrille Michon has forcefully argued that Ockham's position on this point is ultimately unintelligible. My aim in the present chapter will be to take up Michon's challenge. I will first explain what the problem is with concepts being labelled as signs in the Ockhamistic framework (section 1). Ockham's definition of a sign in the Sum oflogic will then be scrutinized (section 2), and how it applies to concepts will be spelled out (section 3). I will explore, finally, the consistency of taking concepts to be signs in this precise sense with two central Ockhamistic tenets: conceptual atomism (section 4) and nominalism itself (section 5). 1. The problem: How can a concept ever be a sign? The standard conception of a sign, in Ockham'stimes, was directly dependent upon Augustine's famous definition: 'a sign is a thing which, in addition to the impression it produces upon the senses, brings by itself something else in our thought'.i Taken literally, this venerable definition squarely excluded intellectual entities such as concepts from the realm ofpossible signs, since it required all signs to produce some 'impressions' of their own upon the senses. Already in the thirteenth century, however, the habit had been taken to sometimes classify intellectual representations as signs too and to drop, consequently, the Augustinian clause about the impression upon the senses. Roger Bacon, for one, was explicit about this in the 1260s: 'not all signs', he wrote, 'are given to the senses as a popular description of sign supposes, but some are given to the intellect only'.2 The rest ofthe Augustinian definition came to be taken as what really mattered: saying that a concept is a sign was, prima facie, to say that it brought to the mind something different from itself. One intuitive way of seeing how this raises a special difficulty for Ockham's later view that concepts are intellectual acts is to remark that the very notion of a sign as bringing about the cognition of something else seems to involve aternary structure where Ockham would admit of only two items. For we should have, apparently, not only the sign itself (that which brings about something) and the external thing (that about which a cognition is activated by the sign - its 'significate' in Ockham's vocabulary), but also a third item: the cognition - or recognition - that the sign is supposed to bring about within the agent who understands it. Describing the concept as a sign seems to require the postulation of some intellectual interpretation of it within the mind, which interpretation, of course, could not in tum be identified with 4~

29 46 CONCEPTS AS SIGNS 47 a sign without the threat of an infinite regress. Ockham, on the other hand, simply equates the conceptual sign with the intellectual act, leaving no room for a third item which could serve as interpretation for the concept. Without interpretation - or understanding - one could ask, how can we still speak of signs? How, in other words, could Ockham literally apply to concepts the key-clause of the Augustinian definition of a sign, namely that a sign should 'bring something in our thought'? There is indeed a passage of the Ordinatio where the Venerabilis Inceptor discusses the meaning of a closely related phrase: ducere in notitiam (to lead to the cognition [of something]). It does nothing much, unfortunately, to alleviate our difficulty. Ockham there distinguished two senses of ducere in notitiam, none of which can be applied to the conceptual acts of his later theory. Here is the text: That something leads to the cognition of something else can be understood in two ways. Either it causes the cognition of this other thing by means of being cognized itself, in such a way that the cognition of it should be the cause of the cognition of the other thing. Or immediately, without being cognized, just as the intellect leads as a cause to the cognition of any intelligible object. 3 Conceptual acts - as Michon rightly points out 4 - cannot 'lead to the cognition' of their significates in the first of these two senses, since they do not have to be themselves the objects of a cognition to function as signs, in Ockham's view. But the second sense will not do either. A concept could 'lead to the cognition' of something in this sense, only if it caused some other cognition within the mind. Which brings us back to our original riddle about interpretation and understanding: a third item something like an interpretation - seems to be needed for the concept to be a sign, in addition to the concept itself and its external significate, something that Ockham, surely, cannot accept. s I do agree with Michon, then, that this passage from the Ordinatio does not shed much light on the general idea of a sign which is required by Ockham's mature theory. This is not too surprising, though. Not only did Ockham still subscribe to the fictum-theory at the time, but he had not yet systematically developed the semiotical approach to concepts which came to be central in his later logical works and in the Quodlibetal Questions. The context of the passage, moreover, is a discussion of the ideas of image (imago) and trace (vestigium); with no explicit reference being made in the surrounding pages to the term signum. The explicitation of ducere in notitiam found in the Ordinatio simply does not meet the needs of Ockham's iater theory of signs and concepts; but, let me insist, it was not supposed to. In order to understand in what sense exactly a concept can be said to be a sign in Ockham's mature theory, we must tum, instead, to the first chapter of the Sum of Logic, Where a full-fledged definition of signum is provided: Nevertheless, to silence hairsplitters, it should be pointed out that the word 'sign' has two different senses. In one sense a sign is anything which when apprehended brings something else to mind, even though it does not bring about a first-time cognition of the thing, as has been shown elsewhere, but merely an actual cognition following,upon a habitual cognition of the same thing. In this sense of 'sign' a spoken word is indeed a natural sign, just like any effect signifies at leasl its cause; and in this sense too is lhe barrel-hoop a sign of wine in the tavern. However. I am not here using the tenn 'sign' in such a general sense. In another sense a sign is that which brings something to cognition and can supposit for that thing, or be added to a sign of this sort in a proposition (for example, syncategorematic expressions, verbs, and other parts of speech lacking a determinate signification), or again that which can be composed of signs of either sort (for example, propositions). And taking the term 'sign' in this sense the spoken word is not the natural sign of anything. 6 In the first sense of signum, according to this characterization, a sign, just like an image or a trace, is typically reminiscent of something which has previously been known in some other way. But, Ockham says, this is not the meaning he wants in the context of his logical treatise; it is the second one. Since conceptual signs are central to logic in his view, we can safely conclude that a concept, in his theory, is a sign in the second sense. This, no doubt, is Ockham's considered answer to the question we are discussing in the present chapter. Michon, however, takes this answer to be unsatisfactory. All signs other than concepts, he remarks, are signs in the first sense. The spoken word, the barrel-hoop in front of the tavern, even smoke with respect to fire, all of these are, he says, 'recordative' signs: they serve as signs in so far as they recall for us something that we have an antecedent knowledge of; and they do so by becoming themselves the objects of a new cognition.? The second sense, consequently, is purely ad hoc. In extending the ordinary scope of the term ~ignum to make it include intellectual acts, Ockham, according to Michon, renders the notion intolerably 'equivocal's and introduces a new meaning for it that has nothing left to do with the customary use of signum: 'The primary notion of a sign has been forgotten.'9 The new definition still boasts the Augustinian authority, since its first clause is that a sign, even in the logical sense, should 'bring something to cognition' but this traditional phrase has now to be taken in a completely new sense: a conceptual sign, in Ockham's mature theory, cannot bring something to cognition in the same sense as a recordative sign does. A wholly new understanding of the cognitive scheme is required. But Ockham, Michon says, always explains conceptual knowledge on the basis of signification. So we face an apparently impenetrable circle: '... until now', Michon writes, 'the sign was defined on the basis of cognition, but if cognition is itself a signification, the resulting circle renders both notions unintelligible'.l0 Such is the 'paradox' of the concept as sign withinockham's actus-theory of concepts, according to Michon. In order to assess this critique, we have to take a close look again at the passage of the Summa Logicae I, 1 to which it pertains. 2. The two meanings of signum The first thing to note about this passage is that it is a reply to an objection. Ockham in these lines is answering some 'hairsplitters' (protervos). To whatthey would have objected, he does not explicitly say, but we can find it out ourselves from the very structure of the passage and its place in the chapter to which it belongs. As it happens, this aspect is quite relevant for the ongoing discussion. Let me show how. Ockham had just distinguished in the preceding paragraphs between three sorts of terms: spoken, written and conceptual. And he had identified two differences

30 48 CONCEPTS AS SIGNS 49 between the first two on the one side, and the third one on the other, two differences which are closely related to one another: first, the concept naturally signifies whatever it signifies, 'whereas the spoken or written term signifies only conventionally';11 and second, the signification of spoken and written words can be changed at will by the speakers, while that of concepts can't. This is where the (real or imagined) hairsplitter breaks in. What he must have asked, I take it, is this: aren't the spoken words natural signs too? Looking at the text again with this in mind, its organization and progression become clear. Ockham introduces the first sense of signum and then concedes that in this sense, yes, 'a spoken word is indeed a natural sign'. But, he adds, this is not the sense he is interested in. Having given the second ~ and more relevant - meaning, he finally concludes that 'taking the term "sign" in this sense the spoken word is not the natural sign of anything'. The distinction between the two senses of signum is introduced to settle a difficulty concerning spoken words, and not especially a problem about concepts. It is true, of course, that it does clarify, in the process, the precise meaning in which we can say that a concept is a sign in Ockham's epistemology. But it does so by spin-off, the strategical point of the distinction being to neutralize the hairsplitter's objection that spoken words, after all, are natural, rather than conventional, signs. It so happens that both questions are simultaneously settled by the clarification of the second sense of signum. But, as the concluding sentences indicate at the end of each definition, the point of the move, first and foremost, was to elucidate the status of spoken words, not that of concepts. This in itself tells against Michon's claim that the second sense of signum was an ad hoc invention of Ockham in view, solely, of including concepts among signs. Sticking to the first - and admittedly more traditional - meaning of 'sign', nothing would have prevented him from saying, with the hairsplitter, that spoken sounds are natural signs. 12 Doesn't a laughter, or a groan, or even a simple intonation of the voice, reveal, by some natural connection, the mental state of whoever produces it? Ockham agrees with this, and goes even further: spoken speech, just like anything else, naturally evokes its cause, exactly like smoke calls fire to the mind. So there is a case for saying that a spoken word is a natural sign of something: it can, like anything else, recall to mind whatever it has a natural connection with (a causal connection, for instance). But the sense in which this is true is of no special interest for logic as Ockham conceives of it. '3 The second sense - the one he favours in the context - corresponds to a completely different notion; Michon is certainly right that there is an equivocity here for signum. This equivocity, however, does not depend - as Michon thinks it does - on Ockham's eagerness to include concepts among signs. 14 It follows, rather, from his attempt to delineate, besides the usual sense of signum, another one - a technical one - which could be of service for logic as a science of arguments, propositions and their components, whether spoken, written or mental's A precise understanding ofthis second definition is needed to catch its motivation and relevance, and to see how exactly the two meanings of si/?num are connected with each other. Ockham's formulation is rather complex and punctuated with 'and's and 'or's. Even the logical form of the definition is not immediately perspicuous. The surface structure of the sentence apparently suggests a conjunction of two clauses, the second one being a disjunction. A sign in the second sense, according to this reading, should: (A) and either (B 1) or (B2) or (B3) either (A) or (B2) or (B3) bring something else to cognition, be able to supposit for that thing in a proposition, be able to act as a syncategorematic term in determining the logical and semantical roles of other terms within the proposition, be composed of such signs, like a complex expression or a sentence. It would follow that all signs in the second sense, like all signs in the first sense, would 'bring to cognition' something different from themselves. This, however, can hardly be what Ockham had in mind, since syncategorematic terms at least - which are explicitly counted as signs according to sense number two - do not satisfy clause (A) in his view: there are no particular objects that syncategoremata such as 'if', 'and', 'all', or 'not', specifically bring to mind. They lack, as Ockham stresses in this very passage, 'a determinate signification'. 16 In all consistency with the rest of the semantical theory, the definition, instead, should be read as a disjunction of three clauses, the first one of which is a conjunction. A sign in this sense must: bring something else to cognition, and (B 1) be able to supposit for that thing (in a proposition) be able to be added to the signs of the precedent category within well formed sentences in some language, be composed in an appropriate way of signs of the previous two categories. It thus becomes clear that the second sense of signum does not merely correspond to a particularcase ofthe first, which would then have been artificially enlarged simply to include concepts, as Michon suggests. Signs which specifically belong to category (B2) in particular (syncategoremata, that is), do not (in so far as they are of interest to logicians) 'bring something else to mind'. And this is true not only of conceptual, but also of conventional syncategoremata. Ockham simply takes it for granted that the term signum as it is used in logic, must apply to syncategoremata. This has nothing to do with his postulation of an oratio mentalis. The medieval logicians' interest for syncategoremata and the habit of calling them signa were well entrenched long before himp His definition here simply takes due account of the practice and terminology of his fellow logicians: a technical sense of signum was needed to cover - at least - the case of syncategoremata, whether spoken, written or mental. The point, admittedly, is not as obvious for signs of category (B3), that is, sentences and other complex expressions, but in this case too, the advantage of the disjunctive reading of the definition, as I just reconstructed it, is that it does not force us to countenance propcr signiticatcs for such special kinds of signs. Sentences as a whole, in this perspective, do not need. any more than.i'yflcllfegoremala, to bring somcthing distinctive to mind in order to be correctly described as signs in the

31 50 CONCEPTS AS SIGNS 51 logical sense. This is in harmony with Ockham's general policy of leaving it to the sole categorematic terms to have proper significates. Even limiting ourselves to categorematic terms, the clause of definition number two which allows us to count them as signs - the conjunction of conditions (A) and (B I) - does not correspond to a particular case of the first sense of signum. The first definition of signum - the general and non-logical one - had three clauses, only one ofwhich matches clause (A) ofthe seconddefinition. A sign in Ockham's first sense must not only 'bring something else to cognition' - the Augustinian condition - but it must do so, in addition, by being itself cognitively apprehended (which could be called the apprehension condition) and by reviving a previously possessed knowledge of the thing (the reminiscence condition). Now, these last two conditions are utterly absent from Ockham's second definition of signum. His use of the phrase 'to bring something to cognition' does not activate by itself the apprehension condition nor the reminiscence condition: some things can 'bring something else to mind' without functioning as 'recordative' signs in Michon's sense. Is this the case only for concepts, as Michon takes it to be? This is doubtful. The correct use of a demonstrative in ordinary language, for example, does not presuppose on the speaker's or the hearer's part any previous knowledge of its referent. The reminiscence condition, at least, is irrelevant here. And learning new names, whether proper or common, cannot presuppose either, in Ockham's view, a previous acquaintance with all of their significates: 'horse' signifies all horses for me, even those I have never seen; 'baobab' in my spoken language signifies all baobabs even though I have never seen any. Nothing implies - or even suggests - in the context of Ockhamism that all linguistic signs other than concepts should satisfy all three conditions for being a sign in sense number one; quite to the contrary. In so far as Ockham introduces a new sense for signum, it is to delineate the sort of objects that interest logic as a scientific field. These objects are characterized by certain functions - such as bringing something to the mind, suppositing for something within a proposition, or being able to act as syncategoremata - which are not specific to concepts (even though, of course, concepts are their prime bearers in Ockham's epistemology). The old Augustinian expression 'to bring something else to mind', which is still used by Ockham to introduce the first sub-category of signs in the second sense the categorematic terms - cannot receive, as we have seen in section 1, any of the meanings Ockham had identified a few years before in the Ordinatio for a related phrase, ducere in notitiam. 18 This, however, in no way entails that his use of it at the beginning of the Summa is out of place, arbitrary, or especially ad hoc. It does leave aside, admittedly, the causal connection that might have been suggested prima facie by 'facit venire' (to bring to): the cognition brought about by a sign does not, in Ockham's new view, have to be something different from the sign itself. But it retains what Ockham takes to be really essential in the Augustinian phrase for characterizing categorematic terms in general (whether spoken, written or mental): the referential drive, the fact that they relate the mind to something different from themselves. The role of clause (A) in Ockham's technical definition of sirnum is to introduce the semantical connection proper: the reference of (some) signs to objects in the world. The other three clauses in the definition - to he able to supposit in a proposition (BI), to be able to play syncategorematic roles (B2), or to be composed of such signs (B3) ~ correspond to derivative functions which all have to do with the combination ofsigns with each other within sentences. Ockham's technical definition restricts the domain of signs which are of interest for logic, to those that can occur, in some capacity or other, in the well-formed sentences of a language. The primary advantage of this move is to provide at the outset, for students of logic, a precise characterization of signs in terms of the semantical and syntactical functions which will be studied in detail in the subsequent chapters of the Sum oflogic. In so far as this strategy produces something new, it is not only to legitimize the classification of concepts as signs; it is, mainly, to identify a well-delineated network of semantical and syntactical features which are basic for the study of logic. These features do not especially depend upon Ockham's particular theory of concepts. The point, originally, is to explain in what sense of signum spoken words can be said to be conventional, rather than natural, signs. And this leads in the process to the formulation of a new definition, based on what we call today semantical and syntactical relations. This definition is suitably tailored to the needs of logic as a science (terminist logic, in particular). It also suits Ockham's epistemology, of course; but this was not its sole - or even primary - intent. There is no objectionably ad hoc manreuvre here, after all. 3. Conceptual roles To say that internal thought is made up of signs comes down, in this perspective, to crediting it with a propositional structure, the very subject predicate structure indeed that the terminist logicians had been studying in spoken languages for decades before Ockham. The basic idea for this language-minded approach to thought was already in Aristotle, for whom concepts (noemata), although prior to linguistic words, could combine, just like words, into propositions, true or false. 19 The whole Aristotelian tradition had acknowledged this requirement, even when it had not made use of the vocabulary of signification to talk about concepts. Thomas Aquinas, for one, tended to reserve the term signum for recordative signs,20 but he accepted without qualms the idea of true or false mental propositions, structurally composed of 'concepts' representing external things. 21 This amounted in effect to an acceptance in principle ofthe oratio mentalis model. Ockham's originality was to systematically draw the consequences of the approach, and to mobilize the whole apparatus of terminist logic - significatio, suppositio, and so on - for the fine-grained description of thought. Identifying the concept with the cognitive act (as in Ockham's mature theory) rather than with its object (as in his former theory and in that of Aquinas) does not render this terminological transposition less intelligible; on the contrary. Once the sign is defined by its semantical and syntactical functions, nothing in principle prevents these functions from being attributed to mental states (or 'acts') rather than 10 mysterious ideal objects such as the ficta of the former theory or the verha n1cntalia of Thomism: 'for an act of understanding can signify something and can supposit for something just as well as any sign', Ockham remarks. 22 What is required to legitimize a systematical extension of the terminist vocabulary to the study of

32 52 CONCEPTS AS SIGNS 53 thought is the general validity of a certain functional description of human mind, according to which some mental units with a referential import can be combined with each other into true or false propositions. This picture in effect was quite generally accepted in Ockham's times. This is not to say that it is a trivial picture. The precise functional model Ockham finally adopted supposes a rather complex conception of the working of the mind. Let us consider again, to bring it out, the functional roles that are mentioned in one way or another in the new technical definition of signum. There are four of them: (RI) (R2) (R3) (R4) bringing something else to cognition (this is clause (A) in the presentation I gave above of Ockham's definition number two), suppositing for this in the context of a proposition (this is clause (Bl», being added within a proposition to signs satisfying both (Rl) and (R2) (this is clause (B2», being a well-formed combination (in one language or other) of signs satisfying either (RI) and (R2), or (R3), a complete sentence for example (this is clause (B3». (Rl) corresponds within Ockham's semantics to signification in the strong sense the significatio jinita, (R2) to supposition, (R3) to the syncategorematic functions (that of the copula, for instance, or of quantifiers, or connectors), and (R4) to the functions of the proposition itself, that, in particular, of being the bearer of a truthvalue. Ockham's definition exploits these four types of roles to differentiate three kinds of signs: those that fulfil (Rl) and (R2), they are the categorematic terms; those that fulfil (R3) only, they are the syncategoremata; and those that fulfil (R4), such as propositions (or sentences) and other complex expressions. Three out of these four roles - (R2), (R3) and (R4) - directly have to do with the formation of propositions. Taken together, they constitute the distinctive element in the technical sense of signum Ockham wants for logic. Human beings are depicted as capable of coherently entertaining true or false propositionally structured beliefs about things other than themselves. The basic role of logic, in this perspective, is to explain what it is for a proposition to be true or false, and what it is for a proposition to be incompatible with another one, or to follow from it. The. technical notion of sign Ockham thinks appropriate for the fiel\i is built up with these tasks in view. Admittedly, this model of the mind also calls for an epistemology; Ockham does provide one, which is discussed in some detail in other parts of this book. But his distinctive epistemological doctrines about concepts, intellectual acts, intentional similarity and so on, are not presupposed by the new definition he proposes for signum. The identification of the concept with a sign on the one hand, and with an intellectual act on the other hand does allow, within Ockham's mature theory, for a wealth of connections between ontology, logic, epistemology and philosophy of mind. But this, on the whole, is quite legitimate. Terminist logic had been interested in semantical and syntactical functions such as supposition and syncategorematic roles long before Ockham, and the general intelligibility of such notions docs not depend upon how Ockham develops his epistemology in the Ordinatio or in the Quoe/liht'tal Questions. The second definition of sirnum, which is built up on the basis of such functions,did not have as its main objective to cover the case of concepts, but to delineate with some precision - in a purely functional way - the set of units which are required by logic as.a science. 4. Atomism or propositionalism? A number of interesting problems still remain about Ockham's treatment of concepts as signs in the second sense. One of them is whether Ockham's semantical theory is best labelled as 'atomistic' or 'propositionalistic', given his technical definition of signum, a debate that went on not too long ago in.the literature. 23 It comes out, at this point, that propositional roles loom large in the logical definition of a sign. Yet it must be stressed that this definitional insistence on propositional combinations is by no means incompatible with the theory of conceptual meaning as being of an atomistic sort basically. A correct understanding of how concepts can function as signs is at stake here. Simple concepts, for Ockham, receive their signification one by one, on the basis of the intuitive acts that originally triggered their formation. As we have seen in the previous chapters, getting intuitively acquainted, say, with a particular chickadee is both sufficient and necessary, under normalconditions, for causing the formation of a simple natural-kind concept for chickadees. More general concepts.- genus concepts - presuppose the antecedent formation of several species concepts; their meanings, nevertheless, are not given to them by a process of syntactical - let alone propositional - combination, but by a causal chain. 24 Structured propositional thoughts become possible only after their categorematic components - the concepts - have, each of them, independently received a signification. A number of commentators have suggested, in the 1960s and 1970s, a 'propositionalist' reading of Ockham's semantics, according to which the proposition rather than the term is 'the primary complete linguistic unit' and the 'primary reality' of cognition in Ockham's system. 25 Their main argument, precisely, rested on the technical definition of 'signum' we have just been considering. Look, they would say, Ockham defines the very idea of a sign on the basis of propositional roles such as supposition; doesn't this show that signification as a property of single terms must presuppose the formation ofsome propositional combinations?26 We are now in a position to see with some precision why thisis not so in Ockham's theory. Among the four defining functions for signs (Rt) - (R4), three can befulfilled only within the context of propositional complexes: supposition (R2), syncategorematic roles (R3) and properly propositional roles (R4). But this is not so for (Ri). Which is no detail. Categorematic terms are the fundamental stuff out ofwhich thought and language are made in Ockham's semantics, and (Rl) identifies their primary function: that of calling something else to mind. This is signification proper. It must helong to single categorematic terms before any proposition involving them can be.limned. And this is especially true for conceptual signs: 'All judicative acts presuppose... the incomplex cognition of the terms, since it presupposes an apprehensive act, and all apprehensive acts with respect to complexes presuppose the incomplex cognition ofthe terms.'27 In other words, a mental judgement presupposes the formation of a menlal proposition, and mental propositions, in so rar as they are

33 54 CONCEPTS AS SIGNS 55 complex arrangements of terms, presuppose that their categorematic components have already been endowed with cognitive signification: they should already, independently of any proposition, refer the mind to some objects in the world. Functions (R2) and (R3) can be fulfilled in the language of thought only if some terms - the categorematical concepts, namely - antecedently fulfil (RI). Such is Ockham's brand of conceptual atomism. That the idea of supposition occurs in the definition of 'signum' in chapter I of the Summa Logicae, and in the definition of 'significare' in chapter 33, in no way reflects a priority of supposition over signification. The definitional order, here, does not reveal the order of priority among the functions themselves. Why should it? The ideas of signification (as a prepropositional function) and supposition (as a propositional function) are intimately connected in Ockham, to the point of interdefining each other. 28 The corresponding functions, however, are clearly hierarchized: signification is prior and more fundamental, supposition is derivative, and so.are the other propositional roles in their own way. The notion of a propositional act that would be primary - or even simultaneous - either chronologically or logically with respect to the signification of its categorematic components ultimately finds no place in Ockham's semantics. And the (chronological or logical) primacy of a mental proposition over the signification of its conceptual components is incompatible with Ockham's most distinctive tenets in epistemology.29 On the other hand, there is an important kernel of truth in the propositionalist interpretation that must not be minimized. All three sorts of signs in the logical sense - categorematic terms, syncategoremata, and complex phrases - must, by virtue of the definition of signum favoured by Ockham, be able to play a part within a proposition. This has to be true of concepts, in particular, in so far as they are signs in the technical sense. Nothing is a concept unless it can be combined with other concepts and mental syncategoremata into mental propositions. The propositionalist commentators went too far, no doubt, when they suggested that a sign 'has no meaning or validity, but in the process ofa proposition'.30 But what they were rightly pointing at is that any unit of thought or language, for Ockham, must be either a proposition, a combination of propositions, or an element capable of occurring within a proposition. Note that in the latter case, it needs not ever actually appear in any propositional context in order to be a sign: the definition of signum does not require the actual occurrence of the sign in a proposition. But under normal conditions all conceptual signs soon find their way into propositions. It is their normal fate. The original formation of a general concept F, in so far as it is based on intuitive cognition in Ockham's picture, would normally be immediately followed by the formation of at least one mental proposition of the form 'this is an F' or 'this G is an F'. Such propositions only come after their component general categorematic terms have been endowed with significations, and as the effects of such endowments (effects which, let me insist, special circumstances could impede);3l yet in the vast majority of cases a newly formed concept will immediately enter some elementary mental proposition, and will most probably reoccur later on in the context of new - and possibly more complicated - propositional combinations. At any rate, it must he able to. This requirement directly follows from Ockham's technical definition of 'sign', plus his central contention that concepts are signs in this very sense. All concepts, in virtue of their being signs, must be able to contribute something to the truthconditions of some mental propositional complexes, and to do so as constitutive parts of these complexes. Conceptual atomism, in Ockham's theory, is the thesis that concepts acquire their signification one by one before entering any proposition whatsoever. This is a central tenet. Butit goes hand in hand with a truly compositional view ofthe working ofthe intellectual mind. The combination of these two elements - conceptual atomism and a compositional theory of thought - is indeed the gist of Ockham's doctrine of the oratio mentalis, just as it is today of Jerry Fodor's doctrine of the language of thought. 32 S. Types and tokens Yet this raises another important - and delicate - question about the integration of concepts among signs within the framework of Ockham's nominalism. It has to do with the inevitable type/token distinction. One might ask, shouldn't a signin the logical sense be repeatable? Ockham's intellectual acts are not. They are singular and transient qualitative episodes in the life of the mind. The repeatability requirement, on the other hand, seems implied by the very wording of Ockham's technical definition: in orderto be a sign in this sense, as we insisted, a concept must be able to enter some propositions. It must be the same concept, consequently, that first receives a signification atomistically, and that enters, afterwards, different propositional acts. How is this consistent with Ockham's nominalism? Repeatability, prima facie, is a feature for universals. But Ockham, ofcourse, would tolerate no universal in his ontology. Concepts, in the mature theory, are identified with singular qualities of individual agents. How can such entities ever serve as repeatable signs? In semiotics, most of the time, it is the types that seem to matter, and not the singular episodes of their instanciations. Standard statements are of the form 'x is a name' or 'x is combinable with y in a certain way', or 'p follows from q', etc., which make more sense about types prima facie than about tokens. Ockham himself, indeed, often speaks as if it was types he was talking about. Think of a principle like the following, which is central in his semantics: a given term can have different suppositions in different contexts. 33 It looks more readily applicable to types than to (okens, no doubt. The Venerabilis Inceptor never tackled the issue directly. But what he is committed to on the topic can be reconstructed safely enough from what he explicitly says. For one thing, Ockham was well aware that sucha distinction as the type/token distinction can be drawn about mental acts. He thought of it as a special case of Aristotle's familiar distinction between numerical identity on the one hand, and specific or generic identity on the other hand. 34 Speaking of mental acts in a quite different context - that of a discussion on virtues - Ockham states that an act which is numerically one 'cannot first occur without circumstances, and then with circumstances'.1~ Strictly speaking, a singular act of cognition is a transient episode indissoluhly linked to its temporal and spatial context of occurrence. Yet this should not keep us, as Ockham makes clear in the same passage, from speaking sometimes

34 56 CONCEPTS AS SIGNS 57 of 'the same act generically, rather than numerically'.36 Having in view generic (or specific) sameness, it is acceptable to say that the same act - the act of loving somebody, for example - reoccurs at different times and places in a given process. The remark, of course, is directly relevant for the case of conceptual signs, in so far, precisely, as they are supposed to occur first outside any propositional combination, and then within various such contexts successively. What the passage shows at any rate is, that Ockham's frequent speaking of types in psychological and semiotical matters was not inconsiderate. He felt ccmfident that talk which is apparently about types in such matters can be rendered legitimate within his nominalistic system, as a special case of resorting to generic or specific identities. In other words, the status of semiotical types such as the word 'horse', the concept 'animal', or the proposition'a horse is an animal', is but a special case with respect to the general question of universals. Ockham's approach to it, then, should be expected to follow his preferred strategy for dissolving universals and abstract entities: that of reinterpreting an apparent reference to a single universal into a distributive reference to several singular entities at once. 'Horse' signifies singular horses, and when taken in propositional contexts normally suppasits for singular horses; no common horseness is required at any stage. This, surely, is the approach Ockham would want to extend to the treatment of apparent references to semiotical types. Speaking of the word 'horse' or the concept 'horse' should be understood, in this perspective, as a way of distributively referring to singular speech or thought episodes. How this is supposed to work obviously calls for some spelling out. But we can build progressively on Ockham's own texts. Let us start with type identity within a single agent. The case is easier to figure out. How, let us ask, can different mental episodes in the life of a given agent be grouped into legitimate types? How is it that several numerically different conceptual acts' of mine can be correctly counted as instances of the same concept? The obvious place in Ockham to look for an answer to such a question is the theory of causal chains within the mind. As we have seen above in chapter 2, a singular conceptual act would normally belong, according to Ockham's epistemology, to a definite causal chain of acts and habitus. Such chains normally start up with intuitive acts, and are made, afterwards, of abstractive acts (or concepts) and their corresponding habitus in alternance: 'ahabitus is the cause ofan act, but not of the same act numerically as the one it was caused by, but it is the cause of a further act. Similarly, an act is the cause ofa habitus, but not of this habitus by which it was caused, but of another habitus or of another degree in the habitus.'3? A typical chain has the following form: intuitive act ~ first abstractive conceptual act (prima abstractiva) ~ habitus] ~ conceptual act z ~ habitusz ~ conceptual act 3, ~ habitus 3... There is, in this picture, a natural causal connection between conceptual acts one, two, three and so on. And this causal connection is taken to implement semantical equivalence among them. Bow exactly it does that is something we will have to discuss in some detail later on. 38 Let us only assume, for now, as a crucial part of Ockham's epistemology, that if a conceptual act is caused in the right way by a certain. habitus, then it inherits the semantical features of the act that had caused (in the right way) this very habitus. Causal chains of the relevant kind must be semantically conducive. This appropriately provides us with the possihility ofa clear criterion for \Pl'c!!;c identity among the conceptual acts of a given agent. Let me propose the following as a plausible candidate: two singular conceptual acts within a given agent belong to the same conceptual type if and only if (1) they both belong to a single causal chain of the kind indicated above, and (2) they are semantically equivalent to each other. 39 Of course, the satisfaction of condition (2) should follow on, normally, from that of condition (1). But since we have not yet examined how this link is secured, we better at this point mention both conditions, simply keeping in mind, for the moment, that condition (2) might turn out to be superfluous. 40 However that may be, once we have clear and stringent identity conditions such as those - with or without condition (2) - there is no harm, presumably, in making use of a type-idiom, with such typical phrases as 'the term a' or 'the proposition p'. Such phrases canlegitimatelybe taken as distributively referring to certain singular acts of the same kind, just as 'horse' distributively refers to singular substances of the same kind, and 'whiteness' to singular qualities of the same kind. Type-phrases, of course, do not usually occur in the plural. We do not say 'the names "equus"', 'the concepts "horse'" or so on, in the way we say 'horses' or 'animals'. But this can be seen as a surface feature of our spoken and written languages,probably reflecting the fact that what is said of a given mental act when we treat it as part of a semiotical system should equally be said of a number of other tokens connected with it in some determinate way. The same approach can be extended to the interpersonal situation. In order to legitimize the grouping of conceptual acts into types across agents, we need to devise a similarly inspired set of precise conditions. We can follow for this the very same policy we have been following in the intrapersonal case: that of combining a causal condition with a semantical one (the latter, maybe, being ultimately dispensable), What this approach yields is this: two singular conceptual acts a 1 and a z of two distinct agents can be said to be of the same conceptual type if and only if: (1) the intrapersonal causal chain to which a 1 belongs has been set in motion (in the right way) by the intuitive apprehension of an individual i 1 which is of the same natural kind as the individual i z which triggered (in the right way) the intrapersonal causal chain to which a z belongs; and (2) a l anda z are semantically equivalent. Condition (1) presupposes, of course, that two different individuals in theworld can indeed be correctly said to be 'of the same natural kind'. But Ockham has no problem with this. He willingly admits that some things are, as a matter offact, of the same natural kind as certain others, independently of the human mind: 'that something is similar to another thing is not brought about by the intellect,he writes, any more than that Socrates is white, or that Plato is white'.41 Ockham's rejection of universals as distinct entities was not intended to compromise all natural groupings. Specific and generic identities are OK, for him, as long as we do not imagine that special extra entities are needed to serve as kinds. The legitimacy of general terms, in Ockham's view, docs require individual things to he naturally ordered somehow independently of the human mind. 42 What his nominalism resolutely wants to avoid is Ihe reification of such orderings into special entilics of their own. The semantics

35 58 and the epistemology are devised to achieve just this. The criteria for semiotical types, then, can legitimately make use of specific or generic groupings, especially if they can be independently ascertained. In the end, the characteristic type-idiom of semiotics can be legitimized, within Ockham's system, by changing, as he often invites us to do, an apparent reference to an abstract entity into a distributive reference to several well-determined singular entities. The salient particularity of semiotical typing, as we have reconstructed it here, is that it connects somehow semantical features with causal history and natural kinds. Ockham's conviction, obviously, is that certain singular conceptual acts are so naturally connected with each other that semantical equivalence between them is (normally) secured. On the whole, then, it is quite true that treating concepts as signs raises a number of delicate questions for Ockham's mature theory of concepts. Some, as Cyrille Michon's, have to do with the exact sense of 'sign' that is used in such a context. Others pertain to the consistency of this semiotical approach with Ockham's conceptual atomism or with his strictly nominalistic ontology for conceptual acts. What I have tried to show in this chapter is that the theory has fruitful resources at its disposal for dealing with these puzzles. The technical definition of 'sign' Ockham provides at the very beginning of the Summa Logicae turns out to be an unusually elaborate one. It harmonizes, in particular, the Augustinian semantical requirement of 'bringing something to mind' with the special interest of logic for syntactical combination. Its motivation, mainly, comes from the needs of terminist logic as it had developedin the previous century. Butthe definition does inform us, at the same time, as to the exact sense in which a concept is said to be a sign in Ockham: in virtue of its causal history, it represents within the mind several individuals of the world in such a way that it can stand for them in various mental propositional combinations, the truth-conditions of which it will then determine in definite ways. Ockham's distinctive nominalistic claim is that a semiotical system of this sort is indeed realized within the mind through causal chains of singular mental acts and habitus. Notes 1. Augustine, De doctrina christiana, II, 1.1, ed. Martin 1962, p. 32: 'Signum est enim res praeter speciem, quam ingerit sensibus, aliquid aliud ex se faciens in cogitationem venire.' 2. Roger Bacon, De signis 2, ed. Fredborg et ai. 1978, p. 82: '...non omne signum offertur sensui ut vulgata descriptio signi supponit, sed aliquod soli intellectui offertur... ' On the progressive acceptance of purely intellectual signs in the thirteenth and early fourteenth century, see Meier-Oeser 1997, chap. 2, and Panaccio 1999a, chap Ord. I, dist. 3, q. 9, OTh II, p. 544: 'Sed tamen aliquid ducere in notitiam alicuius potest intelligi dupliciter: vel tamquall causativum notitiae alterius mediante sua notitia, ita quod notitia ipsius sit causa notitiae alterius. Vel immediate sine notitia, sicut intellectus ducit tamquam causa in notitiam cuiuslibet intelligibilis.' 4. Michon 1994, p This problem, actually, also arises in connection with the tirst sense of ducerl' in notitiam: both meanings, as characterized in the passage quoted above. require the causation of some cognition. CONCEPTS AS SIGNS 6. SL I, 1, OPh I, pp. 8-9: 'Propter tamen protervos est sciendum quod signum dupliciter accipitur. Uno modo pro omni ilia quod apprehensum aliquid aliud facit in cognitionem venire, quamvis non faciat mentem venire in primam cognitionem eius, sicut alibi est ostensum, sed in actualem post habitualem eiusdem. Et sic vox naturaliter significat, sicut quilibet effectus significat saltern suam causam; sicut [= sic?] etiam circulus significat vinum in taberna. Sed tam generaliter non loquor hic de signa. Aliter accipitur signum pro ilia quod aliquit facit in cognitionem venire et natum est pro ilio supponere vel tali addi in propositione, cuiusmodi sunt syncategoremata et verba et iliae partes orationis quae finitam significationem non habent, vel quod natum est componi ex talibus. cuiusmodi est oratio. Et sic accipiendo hoc vocabulum "signum" vox nullius est signum naturale.' The English translation given here is very much inspired by that of Loux 1974, p. 51, but it differs from it in many details. Among other things, I favour the reading of sic instead of sicutat the beginning of the sentence about the wine in the tavern (thus following a number ofgood manuscripts; see the critical apparatus for SL I, line 58 in OPh I, p. 9). 7. Michon 1994, pp Ibid., p Ibid., p Ibid., pp SL I, 1, OPh I, p. 8: '...conceptus seu passio animae naturaliter significat quidquid significat, terminus autem prolatus vel scriptus nihil significat nisi secundum voluntariam institutionem' (transl. Loux 1974, p. 50). 12. In what sense spoken words could be said to be natural signs had been discussed, for example, by Roger Bacon in paragraphs 8 to 14 of his De signis (ed. Fredborg et al 1978, pp. 83-6). 13. It should be noted that signs in the first sense of signum are not all of them 'natural' in Ockham's view. The barrel-hoop in front of the tavern, which he also mentions as an example of signs in the first sense, was a paradigmatic example of a conventional sign in medieval discussions (see, for instance, Rosier 1994, p. 101). The barrel-hoop evokes wine for whoever has already drank wine (or even only seen some I), but this is due to a conventional connection. This is why the choice ofsicut by the editors in line 58 ofthe above quoted passage (from OPh I, p. 9) seems unhappy. I would favour sic, along with many good manuscripts. The barrel-hoop, according to this reading, is given as an example, not of a natural sign, but of a recordative sign in general. The point, anyway, is minor. 14. See Michon 1994,pp See SL I, 1, OPh I, p. 7: 'Omnes logicae tractatores intendunt astruere quod argumenta ex propositionibus et propositiones ex terminis componuntur.' This is the very first sentence of the treatise proper. 16. William uses this very same phrase (non habentfinitam signijicationem) to characterize syncategoremata in general in SL I, 4, OPh I, p On the special importance of syncategorematic terms in medieval logic, see Kretzmann See the text quoted above in n See Perihermeneias, 1, 16a See Thomas Aquinas, Quaest. Disp. de Veritate IX, 4, ad. 4; and S. Theol. III, 60, 4, ad This is the Thomistic theory of the 'mental word' (verbum mentis) as it is presented. for example, in Aquinas's commentary on John's Gospel (Super Evang. S. Ioannis 1) and in his commentary on the Pcrihermeneias. I have examined this doctrine in some detail elsewhere (Panaccio 1l)lJ2b, Il)l)l)a, chap. 6, 200 I). 22. SL I. 12. OPh I. p. 4]:... supponere pro alio et signiticare aliud ita potesl competere aclui intclligcndi Sleul alii sigllo' (lrans!. Lnux p. 74). 59

36 23. 'Propositionalist' readings of Ockham have been - independently - proposed by De 24. See supra, chap. 1, sect See pe Andres 1969, pp , 233-6, and passim. 26. They would usually mention, in the same spirit, Ockham's discussion of the verb the verb'significare' in the definition of personal supposition. 34. See Aristotle, Metaphysics 6, 1016b32-3. sine circumstanciis et postea cum circumtanciis'. JH. See i"ji /I. c:hap. 7. ('eadem per aequivalentiam'); see Ord. I, dist. 2, q. 8, OTh II, p Semantical equivalence is explicitly acknowledged by Ockham as relevant for the grouping together of distinct tokens: these are then said to be 'identical by equivalence' 40. If it did, Ockham's system could be literally labeled as 'naturalistic' with respect to semantical features. 41. Ord. I, dist. 30, q. 5, OTh IV, p. 385: 'et ad hoc quod unum sit simile alteri non plus facit intellectus quam facit ad hoc quod Sortes sit albus vel quod Plato sit albus'. See also Ord. I, dist. 30, q. 1, OTh IV, p. 316; and Quodl. VI, q. 25, OTh IX, p Ockham's position on natural cospecificity is discussed in some detail in Panaccio 1992a, pp More on this below in chap. 5, sect OCKRAM ON CONCEPTS CONCEPTS AS SIGNS 61 Andres 1969, Rochart 1972, Loux 1979, and Biard I have discussed these in detail in Panaccio 1984 and 1992a. A moderate propositionalist approach.,.-. which is very close to what I am defending here - is developed in Biard 1989, chap. 3, pp , and summarized in Biard 1997, pp 'significare' in SL 1,33, OPh I, pp. 95-6, where the first two - and more basic - senses of it are characterized in terms of the propositional Junction of supposition: 'Nam uno modo dicitur signum aliquid significare quando supponit vel natum est supponere pro illo... Aliter accipitur "significare" quando illidsignum in aliqua,propositione de praeterito vel de futuro vel de praesenti vel in aliqua propositione vera de modo potest pro illo supponere' (po 95). 27. Ord. I, Prol., q. 1, OTh I, p. 21: '... omnis actus iudicativus praesupponit... notitiam incomp1exam terminorum, quia praesupponit actum apprehensivum. Et actus apprehensivus respectu alicuius complexi praesupponit notitiam incomplexam terminorum...' And a few lines further down the same page: '... patet quod intellectus nullam propositionem potest formare, nec per consequens apprehendere, nisi primo intelligat singularia, id est incomplexa.' 28. In SL 1,63, Ockham uses the adverb 'significative' (significatively) in his definition of 'suppositio' in general; in the next chapter, he uses both the adverb 'significative' and 29. Even in the basic case of the intuitive cognition and the simple propositional acts that it causes in the mind (such as 'Socrates exists', 'Socrates is sitting', or even 'this exists'), Ockham is eager to maintain that they art; distinct from each other (see Ord. I, Prol., q. 1, OTh I, p. 69: '... dico quod notitia illa intuitiva et illud iudicium distinguuntur realiter... '), and that the latter is caused by the former. 30. Rochart 1972, p Ockham admits that an intuitive cognition could happen, in special circumstances, not to be followed by any judgement as is normally the case(see Ord I, Prol., q. 1, OTh I, p. 70: '... dico quod potest fieri illa notitia intuitiva sine iudicioconsequente~). It must be so, a fortiori, for the abstractive cognition with respect to any proposition whatsoever. 32. Conceptual atomism and the compositional-propositional picture of human thought are both at the heart of Fodor's recent theory of concepts (Fodor 1998). 33. This is presupposed, for example, by such statements as the following: '... any term which can in any way be a part of a proposition can exhibit material supposition' ('suppositio materialis cuilibet quod quocumque modo potest esse pars propositionis competere potest', SL 1,67, OPh I, pp ; transl. Loux 1974, p. 197), which strongly suggests that the very same term can have personal supposition on some occasions, and material supposition on other occasions. The same, mutatis mutandis, is said of simple supposition in SL I, 68, OPh I, p Quaest. Var., q. 7, art. 3, OTh VIII, p. 351: '... non potest esse idem actus numero primo 36. Ibid.:'... idem actus saltern genere, - non dico numero... ' The distinction between specific and numerical identity is also applied to acts in Quodl. III, q. 21, OTh IX, p Rep. III, q. 8, OTh VI, p. 205: 'Dnde habitus est causa actus, sed non eiusdem numero a quo causabatur, sed est causa alterius actus. Similiter, actus cst causa habitus, sed non illius a quo causabatur, sed vel alterius habitus vel altcrius gradu in habitu.' The same point is made in Quodl. lll, q. 21, OTh IX, p. 287.

37 Chapter 4 Connotative Terms in Mental Language Ockham understands it, include simple connotative terms? 6.1 Being signs in the logical sense, concepts, for Ockham, are the primary objects of logic. The newly introduced tools of terminist logic, in particular, are thus systematically developed and used by him as a framework for the fine-grained analysis of human thought. Concepts are seen as the general categorematic terms of mental language, and the technical ideas of signification (significatio) and supposition (suppositio) serve in the Summa Logicae as the basis for a detailed nominalistic theory of truth-conditions for mental propositions. As I have tried to show elsewhere, it is one of Ockham's original contributions to Western philosophy that he inventively turned the technical apparatus ofthe language sciences ofhis time into a theory of how the human mind ideally works.! This approach, as Ockham was well-aware, raises a number of intriguing and fruitful questions as to which semiotical categories exactly can legitimately be transferred to the logical study of the language of thought. One such question, in particular, has recently come to the forefront in the Ockham literature, and it has. somewhat surprisingly, turned out that the understanding of Ockham's whole programme about conceptual thought is at stake in this apparently technical point. This chapter and the following two will be devoted to this discussion. 2 The problem, in its simplest form, is the following: does mental language, a.\' The distinction between absolute and connotative terms out1il1ed by Ockham in Part one, chapter 10, of the Summa Logicae, plays a major role in his general nominalistic approach. Certain signs - the connotative ones such as 'white', 'father'. 'movement', 'time', and a lot of others - are endowed with a hierarchized internal semantical structure: in addition to their primary signification, they are said to have a connotation (or secondary signification) and this semantical duality allows in crucial cases for major ontological simplifications. The non-connotative categorematic terms, on the other hand, are called 'absolute'; they have only a primary signification and no connotation. In an influential paper published in 1975, Paul Vincent Spade has ably defended the idea that, although he does not explicitly say so, Ockham's best theory should be that: 'There are no simple connotative terms in mental language; all simple or noncomplex concepts are absolute mental terms.'3 This has been accepted by many of the best Ockham scholars. 4 But it has long seemed to me that such an admission would be fatal to Ockham's nominalism for the following reason: all relational terms, according to him, are connotative (this is precisely what allows him to do away with relations in the ontology). If, then, mental language had no simple connotative terms, it would follow that there are no simple relational concepti; buc. as is well-known since at least Bertrand Russell, it is logically impoulbll to construct all relational concepts exclusively from non-relational simple 0lIl.

38 64 CONNOTATIVE TERMS IN MENTAL LANGUAGE 65 Ockham's nominalistic doctrine of thought and relations would therefore be doomed to failure. s This worry prompted me to re-examine carefully a large number of relevant passages in Ockham, and this led me to the surprising conclusion, which I Will defend here, that Spade and all these other commentators whom I deeply respect are just wrong on this: Ockham consciously thought that there are simple connotative concepts in mental language (section 2 below) and nothing he says commits him to the contrary (sections 3 and 4). 1. Connotative terms Let us first recall some of Ockham's main theses about the distinction between absolute and connotative terms. It is presented by him as a division among categorematic terms, which are those signs - such as 'man', 'white', 'mother', 'concept', etc. - which are normally used to refer (or purport to refer) to individual things in the world or in the mind. 6 In a nominalistic vein, any such term is said to primarily signify each individual of which it is true. A general term like 'man', for instance, primarily signifies not a general entity such as human nature, nor an abstract Platonistic idea or a Fregean sense, but only men themselves, those individuals, that is, of which it would be true to say 'this is a man'. Now, some of the categoremata - but not all - have, in.addition to this primary signification, a secondary sign(fication: they direct the mind not merely toward their primary significates, but also toward some other things. Take, for example, a relational term like 'father'. Its primary significates in the strictest sense are at any time those individuals of which it is at that time true to say 'this is a father'. But the term 'father' in virtue of its very meaning also turns the mind toward those individuals that have a father, namely the children. These are its secondary significates, or, as Ockham also says, its cannatata. Whenever a categorematic sign has secondary significates, it is a connotative term. Otherwise, it is an absolute one.? An absolute term, then, treats all of its signijicata on a par. 'Man' - which is, for Ockham, a paradigmatic case of absolute term - signifies men and nothing but men, and it does not signify any man differently than any other. Such terms roughly correspond to what is nowadays called 'natural-kind terms'. They are relatively rare. Since only singular substances and singular qualities (singular whitenesses for instance) are admitted in Ockham's ontology, the only absolute terms there are are the natural-kind terms from the categories of substance (for example, 'man', 'horse', 'animal', 'tree') and quality (for example, 'whiteness', 'colour'). All other categorematic terms are connotative. This includes all relational terms such as 'father', 'owner', and so on, all concrete accidental terms such as 'white' (which, according to Ockham, primarily signifies the individual substances that are white, and secondarily signifies their singular whitenesses), all categorematic terms from the Aristotelian categories of quantity, action, passion, time, place, position and habitus, as well as all negative or privative expressions like 'immaterial' or 'blindness', and also such philosophically significant terms as 'true', 'good', 'intellect', 'will', and so on. One can easily see the extreme importance of connotation theory for Ockham's nominalism. Every sign, be it linguistic or mental, must, if it is to be truly significant, refer the mind in some way to singular substances or singular qualities, and to nothing else (since nothing else exists). Butit cando so through connotation as well as through primary signification, and this duality accounts for much of the richness of language without jeopardizing ontological economy. Ockham, for example, avoids being committed to the ontological acceptance of relations by analysing relational terms as signifying nothing but individuals, some of which they primarily signify and some ofwhich they connote. And quantities are disposed of in a similar fashion. Frege's puzzle about identity, which led him to introduce intensional entities such as senses besides denotations, can also be solved in an Ockhamistic framework through connotation theory. How is it that expressions such as 'the Morning Star' and 'the Evening Star' are not synonymous even though they have the same denotation? Ockham's answer would be that while they have the same primary significate, they are connotative terms with different cannotata. And ofcourse, we will have a similar solution to Quine's puzzle about 'renate' and 'cordate' which are supposed to have the same extension while not being synonymous. Very simple, Ockham would say! No need here for ontological eccentricities: 'renate' and 'cordate' primarily signify the same animals, but the former connotes kidneys while the latter connotes hearts and that is precisely why they are not synonymous. So it is not surprising to find out that Ockham himself calls upon his notion of connotation to solve a large array of theoretical problems, not only in logic, but in theology and in natural philosophy as well. The particularities of theological discourse, for example, are in large part explained by connotation. How is it, Ockham asks, that we have in theology different non-synonymous concepts all applicable to God (such as 'good', 'wise', 'creator', etc.) if He really is a perfectly simple being? His answer, of course, is that connotation allows in such a case for a multiplicity of concepts primarily signifying the same thing, but connoting different other beings;8 And in many occasions does William, with the help of connotation, subject the terminology of natural philosophy and of mathematics to reductive analysis in view of showing that neither numbers, nor.geometrical points or lines, nor movement, void, place or time have any reality distinct from that of singular substances and qualities. 9 Ockham's nominalism, in short, relies heavily on connotation theory. Now, why would anybody think that such a useful and crucial device only belongs to the superficial level of conventional languages and that it is utterly absent (at least as a property of simple terms) from the deep structures of mental language? I will, in section 3 and 4, scrutinize in some detail the arguments used by Spade and others to sustain suchan exclusion. But details apart, itmust at this pointbe saidthat their interpretation rests. for a large part on a particular thesis which is clearly and frequently asserted by William himself: a connotative term always has a nominal definition (a definition quid nominis), while strictly speaking no absolute term does. lo Since absolute terms uniformly signify whatever they signify, the only sort of definition they are liable to receive is a quid rei definition, which is a description of the essential features of their significates but in no wayan explicitation of their I,('/"hal meaning. Connotative terms, on the contrary, both havc a primary and a seclmdary signification. and this internal hierarchizcd structure, Ockham thinks, can

39 66 CONNOTATIVE TERMS IN MENTAL LANGUAGE 67 always be rendered explicit with the help of a complex expression, which will precisely be the term's nominal definition. Spade's idea is that ifthis is so, then for every connotative term, there exists what he calls a fully expanded nominal definition which is strictly equivalent with that term and in which the only occurring.non-complex categorematic terms are absolute ones. II The point is that if a connotative name occurs in the nominal definition of another connotative name, it can be replaced in this definition by its own nominal definition; the process can be repeated until only absolute terms are left, along with syncategoremata. In other words, connotative terms, since they are definable, are all logically dispensable. And since, according to Ockham himself, there is no synonymy in mental language, all connotative terms, Spade concludes, should be represented in mentalese by their fully expanded nominal definitions. The problem here, I am afraid, is that this conclusion runs head on against Ockham's own considered beliefs about connotative terms. This is what I will now show in section 2 by going through some relevant texts of Ockham. And I will return in section 3 to Spade's synonymy argument in order to find out what exactly is wrong with it. 2. Mental connotation Let us first tum to Quodlibeta V, question 25, where the problem under discussion is precisely the following: is there in the mind a real distinction between absolute, connotative, and relative concepts? Ockham's answer is an unequivocal yes: I reply that, according to philosophers, the [affirmative answer to this question] is certain. For the concept human being is absolute, the concept white is connotative, and the concept father is relative. And [the three types of concepts] do not overlap except as superior and inferior, since every relative concept is connotative, but not vice versa. 12 So, there are connotative terms in qlentallanguage. Of course, Spade's thesis does not directly contradict that, since in its most careful formulation, it is only that there are no simple connotative terms in mentalese. But if this is whatockham thought, the least that can be said is that he missed a good chance of saying so! His examples here are all simple expressions such as 'man', 'white', and 'father'. Ifhe thought that such written expressions really corresponded to mental complexes containing only absolute terms and syncategoremata, his way of speaking here has to be seen as very misleading. Moreover, if we should both accept Ockham's explicit thesis in this passage (that there really are.connotative terms in mentalese) and Spade's thesis (that no connotative mental concept is simple), there would follow, as Spade clearly saw himself, that mental connotative terms are identical with their nominal definitions. 13 And this directly contradicts something Ockham frequently repeats: 'a definition is not identical with what is defined'.14 His explanation of the point is precisely that the definition is a longer expression than the defined term. That this is not an accidental feature follows from something Ockham considers as a commonplace about definition, namely that it is more explicit than what is defined. I> Other passages are even more direct. Speaking about the knowability of God in Book One of his commentary on the S/'nt/'1/('/'s. Ockham writes that 'God can he known by us through a simple connotative and negative concept which is proper to Him.' Such a concept, he adds, 'is simple,even though it signifies distinct things, principally or secundarily, directly, that is, or obliquely'.16 Concepts, as we already know, are the basic terms of mental discourse. So what is unambiguously presupposed here is that there are simple mental connotative terms and that their internal hierarchized semantical structure does not preclude them from being simple. The notion of a simple term which is involved in this passage is quite clearly, it seems to me, the same as in Spade, namely: the old Aristotelian idea of a sign no part of which is itself a sign. It is sufficient to prove it to remark that simple concepts are, all along the surrounding pages, contrasted with composite concepts, which are themselves characterized as having different parts, each of which is independently 'abstractible' from things, which comes pretty much to the same as saying that each of these parts has an independent signification. I? In the following question of the same book, we find another very clear assertion: 'I say', Ockham writes, 'that of the same thing there can be a plurality of different simple denominative concepts, and this is due to the diversity of their connotata.'18 Denominative terms are paronyms, terms such as 'white', 'just', 'courageous' which are used to attribute to individual things certain properties for which there exist corresponding abstract names such as 'whiteness', 'justice' or 'courage.' 19 As is clear from the above quotation(as well as from many other passages), denominative terms are all seen by Ockham as connotative. So once more the conclusion is unavoidable: there are simple connotative terms in the mind. One can still resist, nevertheless, and remind us that these passages from the Ordinatio were written early in Ockham's career and that he could have changed his mind on this as he did on other subjects. This is true, of course: he could have. But as far as I can see, he didn't. Let us look at Summa Logical' I. Here the evidence is admittedly a little more indirect, but still, it seems to me, overwhelming. On the one hand, Ockham announces at the beginning of chapter 3 that he is now about to introduce a number of distinctions among 'incomplex terms';20 this is certainly meant to include the absolute/connotative distinction drawn in chapter 10. On the other hand, we find, at the beginning of chapter 11, the following statement: 'All divisions we have considered so far apply both to terms which naturally signify and to terms which are merely conventional signs.'21 This impliesthat the distinction of the immediately preceding chapter is among those which pertain to mental as well as to conventional language. That some mental terms are both simple and connotative neatly follows from these two considerations taken jointly. A similar argument can also be derived from the way Ockham understands the Aristotelian division of the ten categories. This, he says, is a classification of incomplex terms, mental as well as verbap2 And it is well-known, on the other hand, that all terms from the categories of relation and quantity, as well as many terms from the other categories, are, according to him, connotative. 23 A doubt might be raised, at this point, about what exactly Ockham means in these passages from the Summa by 'incomplexum'. Couldn't it be, after all, something different than what was meant by 'simplex' in the previously quoted passages from the Ordinatio'! In chapter 4 of his commentary on Aristotle's Categories, Ockham had distinguished two senses of 'incomplexum': 'In the strict sense, an incomplex is a simple word. a word. that is. to which no other word is added. such as: man. runs.

40 68 CONNOTATIVE TERMS IN MENTAL LANGUAGE 69 lion, goat. And by contrast, a complex is any expression composed of different words.'24 In this sense, then, an incomplex sign is precisely a sign of which no part is itself a sign. If Ockham is using this strict sense in the passages we have looked at, then the conclusion we have drawn from them does contradict Spade's thesis. But there is for Ockhamanother meaning of 'incomplexum': In a second sense, an incomplex is the term of a proposition, whether it is a single word or composed of different words; 'white man', for example, is an incomplex in this sense, and so is 'white musician' and all similar expressions. And by contrast a complex is an expression composed of a name and a verb. 25 In other words, an incomplex term in this sense is whatever can be the subject, the predicate, or the copula of a proposition. In this sense a nominal definition can itself be said to bean incomplex term and holding,. therefore, that there are incomplex connotative terms in mental language in this large sense would not preclude them to be identified with their fully expanded definitions, just as Spade wants it. Assuming that no other sense of 'incomplexum' is relevant here, the question we now have to face is this: which one of these two senses is Ockham using in the passages I have quoted? An important clue is provided by Ockham's distinction in Summa Logical' I, chapter 2, ofdifferent senses of the word 'terminus'. In the first of these, a terminus is any expression that can be the copula, the subject, or the predicate of a proposition. Please note that this corresponds exactly to the second sense of 'incomplexum' in the commentary on the Categories. But 'terminus' also has another meaning according to which, Ockham says, 'all incomplexes are called terms'.26 My point here is that the word 'incorrtplexum' in this last sentence can only be used in its first sense (that is,a sign no part of which is a sign), since otherwise there would be no difference between the first and the second meanings of 'terminus'.27 If this is so, then we have a strong presumption that Ockham keeps using 'incomplexum' in the same sense in the other passages we have referred to from Summa Logical' T, especially in the one from chapter 3, which occurs only a page down. To sustain the rival interpretation, one would have to suppose not only that Ockham changed his mind between the Ordinatio and the Summa about connotative concepts - and without ever saying so...,- but also that he changes the meaning of the word 'incomplexum' from chapter 2 to chapter 3 of Summa Logical'. Such a daring interpretative hypothesis, I would say, should only be admitted in the face of strong evidence, which, as we shall see in the next sections, we do not have in this case. But before we get to that, I would like to show that not only did Ockham consciously think that some simple concepts were connotative (as I hope to have established by now), but also that this is not, in his system, an isolated thesis which he could simply have dropped without harm: the thesis, I maintain, is directly implied by at least one other important philosophical doctrine which Ockham explicitly holds. This doctrine has to do with the nominal definitions of correlative terms such as 'father' and 'child', or 'double' and 'half'. None of these, Ockham says, can be completely defined without the help of its correlative. Thus a complete nominal definition of 'father' would have to be something like 'male animal having 1I child', and the complete definition of 'double' has to include the correlative term 'half'.28 Whether these examples, which are from Ockham himself, are well-chosen or not, the important point here is that some relational terms (if not all) cannot be defined without recourse to other relational terms; therefore, since all relational terms are connotative, Ockham is directly committed to denying the thesis Spade attributes to him that all connotative terms are liable to fully expanded nominal definitions, 'each of whose constituent non-complex categorematic terms is abso~ute'.29 The resulting circularity in nominal definitions (since, for example, 'child' must occur in the definition of 'father', and 'father' in the definition of 'child') is not a problem for Ockham: if two concepts are really correlative to each other, they have to be acquired simultaneously and there is no difficulty in admitting that they mutuallydefine each other - nominal definitions do not necessarily proceed from the better known to the lesser known. 30 Ockham, then, just like Bertrand Russell some six centuries later, explicitly denied that relational terms are all logically constructible from absolute ones. They are not dispensable in mental language as he understands i\)' So, what we have now is that Spade's interpretation contradicts at least three of Ockham's explicit and repeated statements: Definitions are always distinct from the terms they define. Some connotative concepts are simple. Relative concepts cannot all be completely defined with the sole help of absolute terms and syncategoremata. But this is still not enough to settle the matter. Spade's point is that Ockham is committed to the denial ofeach one of these three theses. And of course, he could be without realizing it. His theory of connotation, after all, could be inconsistent! What is left for us to scrutinize, then, is the arguments which led such renowned commentators as Spade, Adams and Normore to believe that William is so committed. 3. Synonymy.and nominal definitions The crucial argument we have to discuss can be summarized in the following way: (1) each connotative term has a nominal definition; (2) a term which has a nominal definition is synonymous with it; (3) there is no synonymy in mental language. Therefore: (4) connotative terms and their definitions cannot exist as distinct units in mental language. Premiss (I), as we have seen in the first part of this paper, is very clear in Ockham, it is often repeated and nowhere denied. Premiss (3) is also quite explicit: 'For there is no multitude of concepts corresponding to the multitude of synonymous [spokenl names.' I! Mental language, according to Ockham, is maximally economic: if a distinction between words has no semantical relevance, then there normally is no

41 70 CONNOTATIVE TERMS IN MENTAL LANGUAGE 71 corresponding distinction in mentalese. What is wrong with the argument is premiss (2): as we shall see, it is not the case in Ockham that a connotative term is exactly synonymous with its definition. It is true that Ockham says in different passages that a connotative term and its nominal definition 'signify the same'.33 But, as Spade readily admits, it is not sufficient, in Ockham's vocabulary, that two terms signify the same individuals in order for them to be synonymous. 'Parent' and 'child' for example both signify exactly the same individuals: 'parent' primarily signifies the parents and secondarily the children, while 'child' primarily signifies the children and secondarily the parents. But, of course, they are not synonymous for the obvious reason that they do not signify the same individuals under the same modes. 34 In Ockham's sense of 'synonyma', two terms are synonymous if and only if whatever is signified by one of them under a certain mode is also signified by the other one under the same mode. 35 Spade clearly realizes this, but he thinks that for Ockham, a connotative term and its definition always signify exactly the same things under the same modes, although, as he acknowledges, Ockham never explicitly says so. Spade reasons as follows. A term like 'blind' is supposed, according to Summa Logical' I, chapter 33, to signify sight negatively. 'And this', Spade remarks, 'seemsto be exactly the way it is signified by its nominal definition, which would presumably be something like "animal not possessing sight".' 36 If this is so, 'there seems', he concludes, 'to be no reason not to generalize this to all connotative terms'.37 But, on the contrary, we now know that there is a very good reason to avoid such a generalization: Ockham's doctrine of connotation would be rendered inconsistent by it! Moreover, there are a number of passages which strongly suggest that Ockham consciously refused the general principle of a total synonymy between the definition and the defined terms. 38 I have quoted above a passage where he says that definitions are more explicit than what they define,39 and this probably expresses his most basic intuitions about the matter. But there is much more. In Quodlibeta V, question 19, which is wholly about the distinction between quid rei and quid nominis definitions, he holds that verbs, adverbs and conjunctions are all liable to nominal definitions and his examples are the following: '... someone who wants to define "where" will say that it is an interrogative adverb of place; likewise, he will saythat "when" is an interrogative adverb of time'.40 The definitions in these cases are metalinguistic while the defined terms are not, and consequently they cannot be synonymous with them. Even if they are not about connotative terms, these examples show at least that Ockham did not subscribe to a general principle.of total synonymy between definitions and defined terms. 41 The most explicit passage I know of is in Book One, chapter 20, of Ockham's commentary on Aristotle's Sophistical Refutations. Here he insists that a nominal definition cannot always be substituted for its definiendum without harm. 42 The definition of 'pug', for example, is 'concave nose', but if we substitute this definition for 'pug' in 'pug nose', we get the unacceptable result 'concave nose nose'.43 More generally, Ockham warns the reader against uncritically using in inferences the idea... that a name and its definition signily exactly the same thing and that the one can he suhstituted to the other and that whatever can he correctly added to lhe one can also he correctly added to the other... as if it was totally certain that such a principle was true, since in fact it is not true although it seems to be...44 Where exactly do the semantical differences lie between the definiendum and the definition? Ockham does not say. But maybe we can find it by ourselves. To do that, we first have to take a look at which modes of signification there are according to Ockham. Many passages suggest a basic' distinction between grammatical and logical modes. 45 The grammatical modes of signification of an expression are wholly determined by its grammatical features: number, case, gender, and so on. 46 The logical modes - which are certainly more relevant for synonymy - must be those which are listed in Summa Logical' 1II-4, chapter 10. According to this passage, a term can signify: in recto or in obliquo (that is: primarily or secondarily); affirmatively or negatively; categorematically or syncategorematically; as a proper name or as a common name. 47 It is to be noted that these logical distinctions are independent of the grammatical features of the term under consideration: 'father', for example, signifies the children obliquely even when it is not taken in one of the oblique cases (genitive, dative, ablative and accusative), and 'blind' signifies sight negatively even if its external grammatical form does not display any negation. This being clarified, we can readily notice that there are in fact some obvious differences between the logical modes of signification of a connotative term taken alone and those which are found in its nominal definition. First, there are cases where the definition itself includes a connotative term. We have seen that relational terms are not ultimately dispensable. A term like 'father' is defined as 'male animal having a child'. In such a definition, the term 'child' is itself connotative. What does it secondarily signify? The parents, of course. So the fathers are connoted (or secondarily signified) in the definition of 'father', but they are certainly not connoted by the term 'father' itself, of which they are the primary significates! Attention has been drawn to another sort of case by John Boler in a very interesting paper about 'Connotative terms in Ockham'.48 Let us consider the definition of 'father' again. It includes the expression 'animal' which occurs there in the nominative case and which, of course, signifies in recto all animals. Should we say, then, that 'father' itself signifies in recto all animals? Certainly not, because that would amount to saying that it primarily signifies all animals, which it does not, since it is not true of each animal that it isa father. We have found so far two major differences between the modes under which certain things are signified in a nominal definition and the modes under which the defined term signifies them (if it does signify them at all): some things are connoted by certain parts of the definition which are not connoted by the defined term; and some things are signified in recto by certain parts of the definition which are not signified in recto by the defined term. Haven't we gone astray, however, in taking into consideration the modes under which certain things are signified by the paris of the definition rather than the modes

42 72 CONNOTATIVE TERMS IN MENTAL LANGUAGE 73 under which they are signified by the definition itself taken as a whole? After all, Spade's thesis is that a connotative term is synonymous with its nominal definition as a whole. But the trouble with the idea of considering the global signification of the definition is that we do not find in Ockham any clear indication about how the signification of such complexes as definitions should be constructed out of the significations oftheir constituent parts. Spade proposes with regard to this a principle which he calls the Additive Principle: 'A complex expression signifies just exactly the sum total of what is signified by its constituent non-complex categorematic terms.'49 Theleast that can be said is that such a principle - which, as Spade remarks, can be found in Buridan 50 - is far from being explicit in Ockham himself, who rarely speaks ofthe global signification of such complex wholes as nominal definitions. But even if it could plausibly be attributed to Ockham,it still would not teach us anything about the modes of signification of the definition as a whole, since it has to do only with the total list of significates, and not with the modes under which each of these significatesis signified. Are the modes of signification of a complex expression different from those of its constituent parts (as the objection we are now considering requires)? And if so, what are the rules which lead from the modes of the parts to the modes of the whole? Ockham is totally silent about these questions, and we should. avoid attributing to him some theses on this that would render his theory inconsistent! The best attitude here is to limit ourselves to considering the modes ofsignification which are involved in the definition withouttrying to attribute some other modes to the definition as a whole. 51 On the other hand, we do have some clearhints in Ockham about certain rules which could lead from what is going on in the nominal definitions to conclusions about the modes of signification of the defined connotative terms. The following two rules, for example, are plausible candidates: (Rl) (R2) A connotative term T negatively signifies an object 0 if and only if 0 is one of the primary significates of a term T' which occurs within the scope of a negation in the nominal definition of T.52 A connotative term T signifies in obliquo an object 0 if 0 is one of the primary significates ofa term T' which occurs in one ofthe oblique cases in the nominal definition of T.53 Such rules account for the facts that a term like 'blind' is said to signify sight negatively (its definition being something like 'animal not having sight') or that a term like 'father' is said to signify the children obliquely (its definition being 'male animal having a child', where 'child' is in the accusative case). And other such rules could be constructed (in order, especially, to yield the primary signification of the defined connotative terms). But even taken jointly, they do not (and should not) imply the complete synonymy between the definitions and the dejinienda. We have seen, on the contrary, that certain semantical features of (parts of) the definition have no equivalent in the defined term. It might even be admitted that in some cases, certain things are signified by some parts of a definition which are not, under any mode, signitied by the dctinicndum. Nowhere, for example, does Ockham say that 'father' signif1es in any way all the animals there are, although the general term 'animal' occurs in its def1nition. Of course, he sometimes asserts, as we saw, that a definition and its dejiniendum 'signify the same'. But a close look at the texts reveals that he has a tendency to make a distinction between 'signifying the same' and 'signifying exactly the same'. A good example of this is found in the Quaestiones Variae, where, at one point, he writes that "'man" and "animal" do signify the same, although not exactly, since "animal" signifies whatever "man" signifies, and more..;'54 And remember that in his commentary on the Sophistical Refutations, he clearly warns the reader against taking as true the principle that a name and its definition signify exactly ('omnino' in this case) the same thing. 55 So it might very well be that what he means when he says that the definition and the definiendum signify the same (without the qualification: exactly) is only that whatever is signified by the definiendum is also signified under some mode in the definition (while the converse might not necessarily hold). Be that as it may, we can at least conclude from this discussion that, as far as we know, nothing in Ockham's writings requires that there be a total synonymy between a connotative term and its definition. Many considerations, on the contrary, point in the other direction. Since, in particular, the admission of the principle leads to inconsistency, it should, it seems to me, definitely be discarded. As to what role exactly the doctrine of nominal definitions plays in Ockham's scheme, we will come back to this in detail in chapter 5 below. 4. Connotative terms and exponible propositions What we have said so far also disposes of a somewhat different argument which might be tentatively designed in order to exclude connotative terms from mental language. AlthoughI have never seenthis argumentin print, itis nevertheless clearly suggested here and there. It is based on Ockham's thesis that any proposition in which there occurs a connotative term is exponible. 56 To say that a proposition is exponible - or that it has exponentes - is to say that it is equivalent in truthconditions to a conjunction. In order for a proposition such as 'something white is running' to be true, for example, it is necessary that the following be true: something is running and in this thing there inheres a whiteness. Each part of such a conjunction is called an exponens of the original proposition. The argument, then, would run as follows: (5) any simple proposition in which there occurs a connotative term has at least two exponentes; (6) in the final analysis, the exponentes should not themselves include any connotative term; (7) the conjunction of the exponentes is synonymous with the exponible proposition; (X) there is no synonymy in mental language; therefore: (9) an exponible proposition and the conjunction of its cxponcntc.l' cannot both exist as distinct units in mental language.

43 74 CONNOTATIVE TERMS IN MENTAL LANGUAGE 75 What has to be rejected here is premiss (6) and premiss (7). Premiss (6) has sometimes been asserted by very good commentators,57 but our previous discussion now shows it to be unacceptable to Ockham. At least one of the exponentes in his theory is built out from the nominal definition of the original connotative term: 'white', for example, is defined as 'something in which there inheres a whiteness', and that is why the proposition 'in (this) thing there inheres a whiteness' has to be counted among the exponentes of 'something white is running'. But we now know that in certain cases, connotative terms cannot be eliminated from complete nominal definitions. There is no reason, therefore, to think that they can be eliminated from the exponentes. If that is true, then premiss (7) also has to fall. As we have seen, some things are sometimes signified in a definition which are not signified under the same modes by the definiendum. If such a definition (or something very close to it) occurs in the exponentes, then some things will be signified in the exponentes which are not signified under the same modes in the original exponible proposition and hence they will not be synonymous. It is interesting to notice that this point has been explicitly discussed by some of Ockham's successors from the end of the fourteenth century (Peter of Ailly, for example) up to the sixteenth. The question was raised as to whether the exponible propositions were really distinct in mental language from the conjunctions of their exponentes. And the usual answer, as we learn from Jennifer Ashworth's study of this discussion, was an unequivocal 'yes'.58 Many arguments were put forward in favour of this conclusion, for example, that in some cases people clearly understand an exponible proposition without knowing what its exponentes are. 59 In particular, an illuminating distinction was drawn, which, I think, was already implicit in Ockham, 'between two kinds of equivalence, equivalence in significando and equivalence in inferendo, or in truth and falsity'.60 It is certainly true that Ockham views the conjunction of the exponentes as equivalent to the original exponible proposition. But he says so only in relation with truth-conditions. Since, as we have seen, the synonymy in such cases cannot always be total, a distinction is strongly suggested here between synonymy (equivalence in significando, according to the later terminology) and equivalence in truth-conditions. In short, the argument from exponibles and exponentes fails for the very same reason as the Spade-Adams-Normore argument: connotative terms cannot always be seen as strictly synonymous with their nominal definitions. The principle according to which there is no synonymy in mentalese simply does not apply in such cases. The conclusion of section 2, then, has to be confidently maintained: there are simple connotative terms in Ockham's mental language. His nominalistic theory of connotation, so crucial for his whole philosophical programme, in no way requires the (impossible) elimination of relational terms. Their indispensability, on the contrary, is explicitly recognized and consistently taken into account. Appendix: A reply to Richard Gaskin Richard Gaskin has recently proposed a detailed critical discussion of the case for simple connotative concepts as I have just presented it. ' >! He hreaks it down into six pieces ofevidence and purports to show that none of them overwhelmingly forces us to accept simple connotative terms in the mental language of Ockham's later theory. Gaskin concedes that the point has been well-established 'as far as the early Ockham is concerned',62 but he thinks William changed his mind on this between the writing of the Ordinatio and that of the Summa Logicae and the Quodlibetal Questions. Gaskin concedes, moreover, that 'Ockham nowhere signals a change of mind' on the matter, but he thinks that 'if challenged on the issue, [Ockham] would agree that his mature position was that there are, in fact, no simple connotative terms at the mental level'.63 The question, of course, is: why would that be? Not only is it the case that Ockham never says anything of the sort, but as I mentioned above, there are a number of passages from the Quodlibetal Questions and the Summa Lqgicae which strongly testify to the contrary. They correspond to arguments (1) to (4) in Gaskin's enumeration: (l) Quodlibetal Questions V, question 25, where Ockham argues that there is a real difference in the mind between absolute and connotative concepts;64 (2) Quodlibetal Questions V, question 15, which explicitly posits that a definition differs from the defined term (contrary to the identification required by Spade between Ockham's connotative concepts and their mental definitions);65 (3) Summa Logicae I, chapters 3 and 11, from which it follows, when taken jointly, that the distinction between absolute and connotative terms drawn in Summa Logicae I, 10 obtains among simple concepts;66 (4) Summa Logicae I, chapters 10 and 41, where the Aristotelian ten categories are presented as a division among simple terms, whether mental or spoken: since most of the categories contain only connotative terms, according to Ockham, it follows, once more, that he is straightforwardly committed to the acceptance of simple connotative concepts. 67 To all of these, Gaskin replies in basically the same way. He agrees that Ockham expresses himself in these passages 'as ifthere were simple connotative terms '.68 But it can all be seen, he contends, as mere far,/ons de parler. In none of these cases, according to Gaskin, is the question of whether there are or not simple connotative concepts really central to Ockham's ongoing purpose in the context. Ockham, therefore, might be using convenient 'shorthands' there, having no need, in these precise contexts, for nuances and complications about the simplicity. or complexity of concepts. This strategy calls for several remarks. For one thing, the implication that there are indeed simple connotative concepts in the mind is not that immaterial in all the passages invoked. When Ockham neatly asserts in Summa Logicae I, 41 that 'the terms subsumed under the categories are simple terms' and that 'this holds for simple terms of both the mental and the vocal sort',69 he is well aware that he is thus taking a stand on one of the most controversial issues in the history of philosophy, and one, moreover, that crucially matters for his own thought: what are the ten Aristotelian categories, categories of? The quoted sentences from Summa LOf.:icae I, 41 provide nothing less than Ockham's considered answer to this age-old riddle; it is difficult to think that he doesn't express himself with great care in this context. The very way in

44 76 CONNOTATIVE TERMS IN MENTAL LANGUAGE 77 which he chose to formulate his general position about the status of the ten categories strongly suggests that there are simple mental concepts from all ten categories within the human mind. This suggestion, therefore, should be taken prima facie with utmost seriousness. If Ockham had thought that all simple terms at. the mental level belong to the categories of substance and quality - as Gaskin wants him to - 70 he would normally have seized the occasion of this general discussion 'on the distinction of the Categories' to say so. Secondly, in some cases, Gaskin's eagerness to make the quoted passages compatible with Ockham's supposed rejection of simple connotative concepts leads to devices which are quite implausible, interpretatively speaking, such as the (distinctively non-ockhamistic) postulation of an 'infinitely expanded nominal definition' in the discussion of argument 2,71 and a convoluted reading of the first sentence of SummaLogicae I, 11 in the discussion of argument Thirdly (and the most important aspect), Gaskin's strategy on the issue makes sense only if there is strong counter-evidence available to neutralize the prima facie literal reading of the quoted passages. This, indeed, is the very basis of Gaskin's hermeneutical method as he precisely explains it itself: The general principle here... is thatif, in a given passage, Ockham writes as though there were simple connotative terms at the mental level, but if that simplicity plays no role in the argumentation of the passage - if his purpose would have been equally well served had Ockham substituted the nominal definitions ofthe simple terms for those terms themselves - then, given that there is pressure on Ockham from elsewhere in his philosophy to deny the existence of simple connotative terms at the mental level, we cannot be sure that in writing as though there were simple mental connotative terms Ockham does not intend that as mere shorthand which could and would be eliminated were one to insist on strict accuracy.73 Gaskin's reinterpretations of the passages referred to in arguments (1)-(4) overtly depend for their plausibility on the assumption 'that there is pressure on Ockham from elsewhere in his philosophy to deny the existence of simple connotative terms at the mental level '. But what pressure? That's the question. Gaskin's answer is that the rejection of simple connotative concepts neatly follows from Ockham's rejection of synonymy within mental language along with the assumption that a nominal definition is always synonymous with the defined term. Which, as the reader will recall, was exactly Spade's original argument. My point against it was that the synonymy of the definition with the defined term is simply not to be admitted in Ockham's theory, and that he did in fact explicitly reject it.7 4 Gaskin, obviously, is not convinced by this refutation, which he discusses independently in a longfootnote of the same paper.75 A lot is now seen to depend on which of us is right about this: whether or not nominal definitions and the defined terms were thought of as synonymous by Ockham turns outto be central to the whole discussion about connotative concepts. For my part, I must say, I still take the textual dossier I provided above on the matter to be decisive. Gaskin's dismissal of its impact is based partly on points made by Paul Vincent Spade and CyrilleMichon, to which I have, I think, sufficiently replied;7o and partly on an original suggestion of his own for constructing complex signitications out of simple ones, which is not found in Ockham. 77 (A) (B) (C) The situation at this point is the following. We all agree that: Ockham accepted simple connotative concepts in his earlier theory. he never signalled afterwards a change of mind on the question. he kept expressing himself 'in many passages' of his later works as iihe still admitted simple connotative concepts. Given this basis, a strong reason is needed if we are to accept Gaskin's supposition that Ockham did in fact change his mind on simple connotative concepts and finally excluded them. 78 But the only reason we are given for this crucially rests on the attribution to Ockham of another thesis (the synonymy of the nominal definition with the term it defines) that he not only never asserted, but that he, in fact, explicitly rejected. Not a very good prospect, it seems to me.7 9 The only way out for Gaskin's interpretation would now be to hold that even though Ockham inadvertently rejected in so many words the synonymy-thesis, he is nevertheless strongly committed to it for yet other reasons, that have to do with what nominal definitions are supposed to accomplish in the framework of his theory of thought and language. This will be the object of the next chapter. Notes 1. Panaccio 1999a. 2. The con;; ofthe present chapter follows Panaccio 1990a, with a few minor amendments. Some ofthe questions and objections this paper raised in the literature are addressed in the footnotes and the Appendix, while others will be discussed in chapters 5 and Spade 1975, p. 67. See also, along the same lines, Spade Two salient examples are Adams 1987, pp , and Normore On the other hand, several scholars, in recent years, have expressed their general agreement with the interpretation I am arguing for here; see in particular Tweedale 1992, Goddu 1993, and Beretta 1999, pp Spade did in fact reach the conclusion that Ockham's reductionist programme is doomed to failure, and for closely related reasons. See on this Spade 1990 and A similar point is made by Mertz 1997, who also thinks that the ontological rejection of relations as distinct entities goes hand in hand in Ockham's thought with the logical elimination ofall polyadic predicates (for example, pp. 64-7). 6. Syncategoremata, by contrast, do not by themselves denote any object neither in the world nor in the mind. They can only be used within complex expressions to modify the syntactic and semantic functions of the categoremata. See SL I, 4, OPh I, pp The locus classicus for the explication of the distinction is SL I, 10, OPh I, pp Ord. I, dist. 3, q. 3, OTh II, p See for example, Exp. in Praedic. 10, OPh II, pp , Tract. de Corp. Chr. 32, OTh X, pp , and SL I, 44--c6,OPh I, pp Goddu 1984 is particularly explicit about the central role of connotation in Ockham's analysis of the language of physics. On its use in mathematics, see also Goddu See SL I, 10, OPh I, p. 35: 'Immo, proprie loquendotalia nomina [= absoluta] non habent definitionem exprimentem quid nominis...;' and p. 36: 'Et tale nomen [= connotalivum I proprie habet definitionem exprimentem quid nominis... ' II. See Spade pp

45 78 CONNOTATIVE TERMS IN MENTAL LANGUAGE Quodl. V, q. 25, OTh IX, p. 583: 'Respondeo: conclusio est certa secundum philosophos, nam conceptus hominis est absolutus, conceptus albi est connotativus, et conceptus patris est relativus. Et non coincidunt nisi sicut superius et inferius, quia omnis conceptus relativus est connotativus, et non econverso' (trans!. Freddoso 1991, p. 486). 13. See Spade 1975, p. 68: '... then all connotative mental terms, if indeed there are any, are complex terms and are their own nominal definitions' (italics are Spade's); see also his note 41 on the same page: '[connotative mental terms] would be "synonymous" with their nominal definitions only in the degenerate sense of being identical with them'. 14. Quodl. V, q. 15, OTh IX, p. 541: '... definitio non esteadem cum definito, quia definitio est sermo longus mentalis, vocalis vel scriptus, etper consequens non esteadem realiter cum re extra nec cum uno termino definito' (italics are mine; trans!. Freddoso 1991, p. 453). As Cyrille Michon has rightly remarked (1994, pp ), this particular sentence occurs in the context of a discussion about real - rather than nominal- definitions. The point it makes, however, is certainly meant to hold for both. Ockham repeats it almost verbally in SL III-3, 22 (see OPh I, p. 680: 'definitio non est idem realiter cum definito'), a passage which, contrary to what Michon claims (1994, p. 375, n. 1), explicitly pertains to 'all definitions', the distinction between real and nominal definitions being introduced in the following chapter only. In SL 1,26, Ockham insists that a nominal definition - just like a real one - is always a complex phrase ('oratio') explicating a single term (see OPh I, p. 88: 'Definitioautem exprimens quid nominis est oratio explicite declarans quid per unam dictionem importatur'). 15. See Summ. Phil. Nat. 1,3, OPh VI, p. 162: 'communiter dicitur quod definitio significat illud idem explicite quod definitum significat implicite'. The sentence that follows makes it clear that Ockham is speaking here of nominal definitions. In Panaccio 1990a, I had also quoted a passage from SL III-2, 14, OPh I, p. 530: 'definitio explicat plures res quam definitum'; but, as was remarked by Spade 1996, p. 232, this passage has to do with real rather than nominal definitions, while our present discussion is about nominal definitions only. 16. Ord. I, dist. 3, q. 2, OTh II, p. 405: '... dico quod Deus potest cognosci a nobis in conceptu simplici connotativo et negativo sibi proprio... et iste conceptus est simplex, quamvis distincta significet, sive principaliter sive secundario, hoc est, vel in recto vel in obliquo... ' Ockham seems to have considered, at this time, these simple connotative concepts as conventional; see ibid., p. 403: '... dico quod [Deus] est cognoscibilis a nobis in conceptu aliquo modo simplici ad placitum instituto ad significandum, et hoc in conceptu connotativo et negativo sibi proprio'. More on this in chapter See, for instance, Ord. I, dist. 3, q. 2, OTh II, pp : '...dieo quod essentia divina vel quidditas divina potest cognosci a nobis in aliquo conceptu sibi proprio, composito tamen, et hoc in conceptu cuius partes sunt abstrahibiles natwaliter a rebus' (italics are mine). See also p. 404, where Ockham is even more explicit about his notion of a composite concept. 18. Ord. Ii dist. 3, q. 3, OTh II, p. 425: '... dico quod eiusdem rei possunt esse plures conceptus simplices denominativi,et hoc propter diversitatem connotatorum... ' 19. See Aristotle, Categories 1, lai3-15; and SL I, 13, p SL I, 3, OPh I, p. 11:. '... prosequendum est de divisionibus termini incomplexi'. 21. SL I,ll, OPh I, p. 38: '... Positis divisionibus quae possunt competere tam terminis naturaliter significantibus quam etiam terminis ad placitum institutis... ' (trans!. Loux 1974, p. 72). 22. See SL I, 41, OPh I, p. 116: '... illa quae sunt in praedicamentis sunt incomplexa ex quibus sunt propositiones natae componi. Et hoc est verum tam de incomplexis mentalibus quam vocalibus... ' 23. See SL 1, 10, OPh I, pp. 37-X: 'Immo, qui ponunt quod quaclibet res est substantia vel qualitas, habent ponere quod omn;u ('(}lit('fltu ;n uli;s ''''urt!;l'uml'flt;s II.whsllllll;a ('I qualitate sunt nomina connotativa' (italics are mine). It should be noted that Ockham is using 'nomina' here in a wide sense since he acknowledges elsewhere that the categories of action and passion contain only verbs, and no nouns in the strict sense (SL 1,56-7, OPh 1, pp and 187), and that the categories of time and place contain only adverbs (SL I, 59-60, OPh I, pp ). Martin Tweedale has surmised that the Ockhamistic distinction between absol.ute and connotative terms. might not pertain to verbs, adverbs and prepositional phrases at all (Tweedale 1992, pp ). This, however, neglects the wide use of 'nomina' that Ockham frequently allows himself: after all, verbs, according to Aristotle, 'are really names and have a determinate meaning' (perihermeneias 3, 16bI9-20), and this must apply to adverbs as well, and to prepositional phrases such as 'on a horse', 'rapidly', etc. Ockham is entirely explicit that connotation is at work even in the categories such as those of action, passion, time and space, which contain no nouns in the strict sense; see the passage from SL I, 10 quoted above in the present note; see also Exp. in Praedic. 12, OPh II, pp : '... relatio non importat aliquam rem quae non sit de genere substantiae vel qualitatis, et tamen de nulla re praedicatur in quid... et hoc quia connotat aliquid praeter ipsum de quo praedicatur. Et sicut est de quantitate etrelatione, ita est de omnibus aliis sex generibus' (italics are mine). His analysis of the verb 'to heat' (calefacere), for example - a verb which is assigned to the category of action - clearly comes down to treating it as a connotative (SL 1,57, OPh I, p..184). 24. Exp. in Praedic. 4, OPh II, p. 148: 'Stricte dicitur incomplexum simplex dictio, hoc est una dictio sine additione alterius dictionis, sicut hie: homo, currit, leo, capra. Et per oppositum dicitur complexum omne compositum ex diversis dictionibus.' 25. Ibid.: 'Secundo modo dicitur incomplexum terminus propositionis sive sit una dictio sive plures dictiones, sicut "homo albus" dicitur incomplexum et similiter "album musicum" et huiusmodi. Et per oppositum dicitur complexum compositum ex nomine et verbo...' 26. SL I, 2, OPh I, p. 9: '... et sic omne incomplexum vocatur terminus'. 27. Richard Gaskin claims that my reading here rests on a misinterpretation of Ockham's second sense for 'terminus' in SL I, 2 (Gaskin 2001, p. 258). His point is that 'terminus' in this sense is explicitly contrasted by Ockham with 'oratio' (which Gaskin interprets as meaning 'sentence' in this context) and that, consequently, a term in the second sense of 'terminus' could very well be complex (like 'white man' or 'the man I saw yesterday') as long as it is not itself a complete sentence. This is wrong-headed, however. 'Oratio', here, is not to be taken as meaning 'sentence' as Gaskin presupposes, but 'complex phrase'. Ockham is explicit about this distinction between these two usual senses of 'oratio'; see Exp. in Perih. 3, OPh II, p. 390; '... oratio dupliciter accipitur: Uno modo large, et sic omnis congeries dictionum est oratio... Isto modo hoc quod dico "homo albus," et similiter "album animal" et huiusmodi sunt orationes. Aliter accipitur "oratio" stricte, et sic oratioest congrua dictionum ordinatio, ubi verbum congruit, et nomen vel aliquid loco nominis...' As I read him, Ockham in SL I, 2 is contrasting 'terminus' in the second sense with 'oratio' in the wide sense: no complex phrase, he wants to say, is a term in this sense. Gaskin's alternative reading of this passage, by contrast, would leave us with a contrived and bizarre way of distinguishing the different senses of 'terminus'. 28. See SL III-3, 26, OPh I, p. 690: 'Et est sciendum quod relativum potest definiri vel complete vel incomplete. Incomplete potest definiri sine suo correlativo... Sed complete non potest definiri sine suo correlativo; ut definiatur sic "pater est substantia sensibilis habens filium".' The same thesis is found in the Exp. in Porph. 2, OPh II, pp. 31-2: 'Est autem notandum quod quando aliqua sunt correlativa, ilia tantum habent quid nominis et non quid rei, et ideo non est inconveniens quod utrumque per alterum detiniatur detinitione exprimentc quid nominis: immo hoi' ('st I/('l'('s.wr;um, cum

46 80 CONNOTATIVE TERMS IN MENTAL LANGUAGE 81 correlativa sint simul in intellectu...' (italics are mine). The matter is discussed in some details in Quodl. VI, q. 24, pp See also Exp. sup. Elench. II, 16, OPh III, p. 296, and Ord. I, 30, q. 3, OTh IV, pp ; 29. Spade 1975, p See SL III-3, 26, OPh I, p. 690: 'Et tali definitione relativa mutuo se definiunt. Nee est hoc inconveniens, quia sicut relativa sunt simul in intellectu, ita simul imponuntur, et propter hoc non est inconveniens si mutuo definiant se. Unde sciendum est quod ista definitio non est per notiora, sed sufficit quod sit per aeque nota.' 31. Tweedale 1992, pp ,has rightly remarked that not only does Ockham admit, for each of two correlative terms, of the possibility of an incomplete nominal definition making no use of the other correlate, but that he considers, in addition, this kind of definition, even if it is sometimes incomplete, as being more properly a definition than the one that contains the correlate (ibid.: '... est magis proprie definitio, quamvis non sit semper ita completa'). Tweedale wrongly concludes, however, tliat the best definition of a relational term is the one that does not contain its correlate, and that, consequently, all relational terms are supposed to be eliminated in the end from the definitions of other relational terms. It is important, here, to precisely understand Ockham's own vocabulary. Ockham distinguishes betweendefinitions 'properly speaking' and complete definitions. The former is the one that goes from the better known to the lesser known; it can be used to teach a new term to somebody who does not know it yet and this, no doubt, must be the reason why it is said to be a definition 'properly speaking'. But the second one is said to be 'more complete'. Why would that be? I can see only one possible explanation: this definition alone unfolds in a complete way the meaning of the definiendum. Ockham's thesis is that the meaning of a relational term cannot be completely given without the help of its correlate. Contrary to what Tweedale thinks, then, the elimination of all relational terms is ultimately. impossible without semantical impoverishment. 32. Quodl. V, q. 9, OTh IX, p. 518: '... quia multitudini nominum synonymorum non correspondet multitudo conceptuum' (trans!. Freddoso 1991, p. 432). See also Quodl. V, q. 8, OTh IX, pp , and SL 1,3, OPh I, pp More on this below in chap. 7, sect See for example SL III-3, 22, OPh I, p. 680: 'Tamen non obstante quod definitio et definitum non sint idem realiter, tamen significant idem realiter' (italics are mine). 34. See Quaest. in Phys., q. 25, OPh VI, p. 461: '... illa nomina "pater" et "filius" significant omnino idem, et tamen haec estfalsa "filius estpater".' See also Exp. in Phys. III, 6, OPh VI, p. 480: '... quidquid importatur per unum correlativum, importatur per reliquum, tamen non eodem modo, quia iliud quod importatur per unum in recto, importatur per reliquum in obliquo et e converso... sicut quidquid importatur per hoc nomen "pater" importatur per hoc nomen "filius" et e converso, quamvis non eodem modo.' The example is much better, of course, if we use 'parent' and 'child' rather than 'father' and 'son', 35. See SL I, 6, OPh I, p. 19: '... dicuntur ilia synonyma quae simpliciter idem significant omnibus modis, ita quod nihil aliquo modo significatur per unum quin eodem modo significatur per reliquum...' 36. Spade 1975, p Ibid., p Martin Tweedale thinks I am needlessly complicating things in trying to demonstrate this point, since the relation of synonymy, according to Tweedale, never connects but simple terms, in Ockham's vocabulary (Tweedale 1992, pp ). To which it must be replied, first, that there is at least one case where Ockham does consider the possibility ofa perfect semantical equivalence between a simple term (any verb, in this case) and a complex phrase (a participle with the verb 'to be'). Such an equivalence, he says, would suffice - in conformity with premiss (3) above - to exclude this redundancy from mental language: 'Nam verbum et participium verbi sumptum cum hoc verbo "est" semper videntur in significando aequivalere. Propter quod... videtur quod non oportet participiis vocalibus distinctos conceptus in mente correspondere' (SL 1,3, OPh I, p. 11). See also Quodl. V, q. 8, OTh IX, pp : '... verbum semper et participium verbi sumptum cum hoc verbo "est" in significando aequipollent et sunt synonyma' (italics are mine). It is true, however, that a difference in structure between two expressions often comes together, for Ockham, with some differences in meaning. But we have to understand why. That is precisely what I am trying to do in the next paragraphs. 39. See above, n Quodl. V, q. 20, OTh IX, p. 556: '... sed definitio exprimens quidnominis estverborum, adverbiorum, coniunctionum, quia qui vult definire "ubi" dicet quod est adverbium interrogativum loci; similiter dicet quod "quando" est adverbium interrogativum temporis, et sic de aliis, ubi praedicatur definitio de definito supponente materialiter' (trans!. Freddoso 1991,.p. 465). 41. Spade 1996, p. 233, rejects this argument of mine by remarking thatif the definition is metalinguistic in such cases, so are the defined terms themselves since they are taken in material supposition, which, he contends, re-establishes semantical equivalence between the two. But this is off the mark. The question is whether the definition and the defined term are synonymous, and synonymy is a matter of signification, not of supposition. Surely the metalinguistic phrase 'an interrogative adverb of place' signifies (in Ockham's sense) certain things that the adverb 'when' does not signify even when taken in material supposition (for taking a term in material supposition does not change its signification), namely adverbs. Hence they are not synonymous. 42. Exp. sup. Elench. I, 20, OPh III, pp See also on the same theme Exp. sup. Elench. II, 16, OPh III, pp Ibid., OPh III, p. 131: '... exemplum est de simo, nam definitio exprimens quid nominis ipsius est "naris cava"; si igitur quidquid potest addi nomini potest addi definitioni, sequitur quod sicut convenienter dicitur "naris sima", ita convenienter dicetur: naris naris cava.' The example is that of Aristotle himself (Soph. Ref 173b9-1O). 44. Ibid., OPh III, p. 133: '... si opponentes non petant in principio sibi dari iliam propositionem, scilicet quod nomen et sua de/initio idem omnino significentet quod loco unius possit alterum poni et quod quidquid convenienteradditur uni potest convenienter addi alted... quasi omninosit certum quod praedicta propositio sit vera, cum tamen non sit vera quamvis videatur esse vera... ' (italics are mine). Michon 1994, p. 369, claims that this remark occurs 'in the context of a study of real definitions', but the text is explicitly intended to cover nominal as well as real definitions (see Exp. sup. Elench. I, 20, p. 129). The redundancy ~ or nugatio - problem (as in 'concave nose nose') typically arises, Ockham says, with nominal definitions (ibid., p. 130: 'sed etiam [causa deceptionis] habet locum in aliis nominibus in quorum definitionibus exprimentibus quid nominis necessario ponitur subiectum'). Since the rejection of the total synonymy between the definition and the defined term is meant to solve the nugatio problem precisely, it better applies to nominal definitions! See also Exp. sup. Elench. I, 20, pp See, among many other examples, Quodl. II, q. 7, OTh IX, pp ' about the synonymy between 'esse' taken as a name and 'essentia': 'Tamen "esse" aliquando est nomen, et tunc significat omni modo grammaticali et logicali idem cum "essentia'" (italics are mine). 46. See, for example, Ord. I, dist. 2, q. 8, OTh II, pp : 'Verbi gratia, isti voci "homo" competit talis modus grammaticalis quod est singularis numeri, nominativi casus, masculini generis, et sic de aliis' (italics are mine). Also Quodl. II, q. 7, OTh IX, p. 144: 'Aliquando l"esse"l est verbum, et tunc signiticat idem verbal iter quod "essentia" signiticat nominaliter.'

47 82 CONNOTATIVE TERMS IN MENTAL LANGUAGE SL III--4, 10, OPh I, p. 817: '... necesse est cognoscere quas res tennini significant et qualiter significant eas, scilicet an in recto vel in obliquo,et an affinnative vel negative, et antamquam categorematici vel syncategorematici, et an tamquam nomina propria vel communia... ' 48. See Boler 1985, esp. pp Spade 1975, p Ibid.., n. 11. On Buridan's Additive Principle, see also King 1985, p Michon 1994, pp and Gaskin 2001, p. 239n. Both suggest that the signification of a complex expression - such as a definition - should not be seen within Ockham's framework as the sum of the significations of its components (as required by the 'Additive Principle'), but that some sort of special semantical restriction should be supposed to operate in such cases to exclude from the significates of the complex expression as a whole some of the significates of its components. I can see, however, no textual support in Ockham for such a strategy. Since it is inconsistent with some of the things he explicitly holds (as in the passage quoted above in n. 44), it should not, as a matter of principle, be attributed to him. Moreover, it is difficult to see, in the case of mental expressions, how the natural signification of their constituent simple concepts could be cancelled by the surrounding tenns. If, for example, the concept 'animal' naturally calls to mind every animal there is, why should it cease to do so when it is followed by the concept 'having a child'? This, actually, is the very reason why Spade, contrary to Gaskin, thinks that Ockham is committed to the Additive Principle. He does not acknowledge, however, that this principle is incompatible, in many cases, with the synonymy of the definition with the defined tenn. 52. This fonnulation is directly inspired by one of Spade 1975, p. 67. But his is meant to yield the mode of signification of the definition as a whole rather than that of the defined tenn, which he only afterwards infers from that of the definition. I don't see any reason for this detour, unless we presuppose that the defined term and its definition are wholly synonymous, which, as we saw, we should not do if we care to avoid attributing a fatal inconsistency to Ockham. 53. I say 'if' in this case, rather than 'if and only if', because I want to leave open, at this stage, the question of the necessary conditions for oblique signification. R1 and R2 are merely given here as examples of rules, and not, of course, as a complete theory. 54. Quaest. Var. q. 6, art. 5, OThVIII, p. 231: '... licethomo et animal significent idem, quamvis non adaequate quia quidquid significat homo et animal et plus... ' 55. See above n This doctrine of the exponibilia is expounded in SL II, 11, OPh I, pp Attempts to eliminate connotative terms from the mental language on the basis of it are to be found in Michon 1994, pp , andyrjonsuuri 1997, pp But, as we shall see, nothing in Ockham indicates that a complete 'exposition' of any given 'exponible' sentence can always be entirely devoid of connotative tenns. Quite to the contrary. 57. See for example Pinborg 1976, p See Ashworth 1973, pp Ibid., p Ibid. 61. Gaskin 2001, pp Ibid., p Ibid., p This is Gaskin's reply to what he counts as my sixth argument: Ockham never explicitly rejected simple connotative concepts. 64. The text is quoted in n. 12 above. 65. See above n See above nn See above nn Gaskin 2001, p. 257, ad 1 (italics are mine). See also ibid., p. 260, ad 4: 'In many passages Ockham writes as if there were simple mental tenns...' (with Gaskin's own italics this time). 69. SL. I, 41, OPh I, p. 116: '... illa quae sunt in praedicamentis sunt incomplexa... Et hoc est verum tam de incomplexis mentalibus quam vocalibus... ' (transi. Loux 1974, pp ; italics are mine). 70. Gaskin 2001, p. 260, ad Ibid., p. 257, ad 2. This is how Gaskin ultimately accommodates Ockham's idea that a nominal definition should differ from the term it defines (which was argument 2 above). 72. Ibid., p. 259, ad 3. According to Gaskin, Ockham's assertion that the distinctions apply both to mental and to conventional signs (as quoted above in n. 21) should not be taken as pertaining to each individual one of the distinctions drawn in the Summa Logicae before that point. This, however, would be quite strange. The whole intent of this particular sentence is to serve as a transition marker within Book One of the Summa Logicae in order to separate those distinctions among simple tenns that do apply to both concepts and words (chapters 4 to 10, including the absolute/connotative distinction treated in chapter 10) from those that don't (chapters 11 to 13). 73. Gaskin 2001, p. 260, ad 4; italics are mine. 74. See in particular the passage from Exp. sup. Elench I, 20 quoted above in n Gaskin 2001, pp , n See above nn. 15,41, See my discussion ofthe point in n. 51 above. 78. Gaskin even describes this as what Ockham came to be 'officially maintaining' (2001, p.240). 79. In addition to (1)-(4) above, Gaskin discusses two more arguments of mine in favour of simple connotative concepts, which he numbers as (5) and (6) (2001, pp ). Argument (5) is that correlative tenns, such as 'parent' and 'child' or 'half' and 'double' are said by Ockham to be ultimately ineliminable from each other's complete nominal definitions (see the texts referred to in nn. 28 and 30 above). Gaskin's reaction to this is to point out, as Tweedale (1992) had done, that they might nevertheless be eliminable from the incomplete definitions. This is true, but of no help, I am afraid, for Gaskin's general interpretation. See my discussion of the point in n. 31 above. As to argument (6), which is that Ockham never explicitly rejected simple connotative concepts, Gaskin grants it, but thinks there is indirect evidence available. Which brings us back to his discussion of arguments (1)-(4)!

48 Chapter 5 The Role of Nominal Definitions Many, as we have seen, have counted Ockham as a resolute supporter of what we call today the 'Definitional View' of concepts, also labelled sometimes as' the 'Classical Theory'. 1 According to this doctrine, most of our mental concepts are actually definitions. They are, in other words, complex mental representations encoding, on the basis of some primitive terms, necessary and sufficient conditions for their own application. In Paul Spade's original interpretation of Ockham, the primitive basis was supposed to be provided by a limited range of simple absolute concepts, acquired by the agent as a consequence of direct encounters with individual exemplifications; all the connotative ones, by contrast - the bulk of our intellectual equipment, certainly - were identified with their nominal definitions. The discussion of the previous chapter now allows us to renounce this picture of Ockham's theory of thought, but its demise, admittedly, leaves us with a pressing puzzle: given that all - and only - connotative terms have a nominal definition, according to Ockham, what is the connection exactly, in his epistemology, between simple connotative concepts and their nominal definitions? Paul Spade himself has forcefully raised the point. If connotative names are not synonymous with their nominal definitions, he asks, then what would the criterion be of a 'correct' nominal definition for a given concept? In other words, 'if it's not synonymy, then what is it?'2 In a 1998 paper in Franciscan Studies, Spade is even more explicit, insisting that his main reason for attributing to Ockham a reductionist programme with respect to connotative terms, had been that such a programme fits so well with Ockham's theory of nominal definition: 'if he did not accept the moderate reductionist programme I have described', Spade asks again, 'then what is the point a/the theory a/nominal definition? '3, This is an entirely relevant question and I would like to address it in this chapter. It has to do, let me insist, not with definitions in general, but only with nominal definitions (de/initiones exprimeyj;tes quid nominis), which Ockham sharply distinguishes from real ~ or essential - definitions (definitiones exprimentes quid rei).4 Real definitions, in the strict sense, are descriptions of the intrinsic essences of things. A man, for example,can be 'really' - or essentially - described as a rational animal. Nominal definitions, on the other hand, primarily say something about names (whether spoken, written or mental). What it is exactly that nominal definitions say about names, and what role they play in Ockham's general nominalistic programme is what we will now be concerned with. The strategy will be the following. I will first recall four characteristic - and nontrivial - theses of Ockham about nominal definitions (section 1). I will then explain what I take the function of these definitions to be and show how the four theses fit with this conception (section 2). A numher of interesting consequences will finally he drawn about Ockham's epistemology and metaphysics (section 1). H~

49 86 THE ROLE OF NOMINAL DEFINITIONS Four theses about nominal definitions The theses r have in mind form at first sight a disparate collection. Yet they are very precise and they are clearly and repeatedly stressed by Ockham. Any understanding of what his theory of nominal definition is about should certainly take them into account. Thesis I: Some nominal definitions irreducibly contain connotative terms. Correlative terms, Ockham often repeats, interdefine each other. 5 These correspond, in his vocabulary, to such pairs as 'father' and 'child', 'double' and 'half', 'larger' and 'smaller', and so on. They are, in modemwords, pairs ofnonsynonymous converse relational terms such that if one of them is true of two individuals in a given order, the other is true of the same individuals.in the reverse order (for example, if x is larger than y, y is smaller than x, and vice versa). Since correlative terms are relational, they are ipso facto connotative. Saying that these terms interdefine each other, therefore, amounts to hold that the nominal definitions of certain connotative terms irreducibly contain at least one other relational connotative term. Nominal definitions in such cases cannot be expected to eliminate all connotative terms. Thesis 2: A nominal definition is not always synonymous with the defined term. That Ockham subscribes to this thesis is also something I take to have been established in the previous chapter. 6 Since, however, the point has been in dispute in the recent literature,? it will be useful to quickly review what the main arguments are for this ascription: (i) Ockham says himself that the principle of a total synonymy between a nominal definition and the defined term is not true. 8 (ii) He nowhere asserts that such synonymies generally hold;9 (iii) Crediting him with the total synonymy principle between nominal definitions and defined terms introduces a prima facie tension in his account of mental language;, since he is committed to accepting on the one hand, the possible coexistence in the mind of simple connotative concepts and their complex nominal definitions,lo while he rejects, on the other hand, total synonymy from the mental language. 11 (iv) In the case of verbs, adverbs and conjunctions, Ockham clearly admits of nominal definitions which he could not have taken as synonymous with the defined terms; 12 this shows at least that he does not accept synonymy as a general constraint on nominal definitions. (v) Ockham holds that nominal definitions signify more explicitly than the defined terms; 13 this suggests that they signify differently, and hence that they are not synonymous with them in the strong sense. (vi) Some things are usually signified in certain ways by parts of the nominal definition, which are not signified in the same way by the defined term)4 Taken jointly, these considerations are decisive, it seems to me, for attributing thesis 2 to Ockham. The possibility is not entirely excluded that some nominal definitions might, in particular cases, be taken as synonymous with what they are used to define, but such situations cannot be but exceptional. Thesis 3: Each connotative term has only one nominal definition. Ockham explicitly says so: 'properly speaking, a name that has a nominal definition has only one such definition'.15 There can be variants, to be more precise, but Ockham's pointis that they cannot differ from each other more than by the substitution of a synonym for a synonym. 16 Nominal definitiol}s for Ockham are always complex phrases. If this is so, it can routinely happen in some spoken or written languages that at least one of the terms occurring within the correct nominal definition of a certain term T has a synonym in the same language. The phrase which differs at most from the nominal definition of T by the substitution of this synonym, then, will clearly be a correct definition of T as well. The discrepancy, however, is not allowed to go further. Any two correctnominal definitions for a given term mustbe isomorphical to each other and such that their corresponding simple components are all strictly synonymous with each other. Since there is supposedly no such synonymy in mental language, there cannotbe for any connotative term more than one appropriate nominal definition of that term within the mental language of a given agent. Ockham, on the other hand, also holds that all connotative terms independently of whether they are spoken, written or mental - do have a nominal definition. 17 It follows from both considerations taken together that any connotative term has exactly one correct definition in the language of thought. Thesis 4: The nominal definition ofa connotative term is a complex phrase one component of which normally is a term 'in recto' and at least one other component a term 'in obliquo'. See, among others, this passage from Summa Logicae III-3: 'Connotative terms are defined by their subjects taken in recto and by the names of their connotata taken in obliquo, or by verbs.' 18 Disregarding the clause about verbs - to which I will soon come back - what the passage says is that a typical nominal definition for a connotative term contains at least two component terms, one taken in recto and one in obliquo, and that the former is expected to be the'subject' of the connotative term, while the latter is supposed to be the 'name of its connotata'. This is crucial for any understanding of what a correct nominal definition is, according to Ockham, and we must, accordingly, pause a bit to see what it means exactly. The terminology 'in recto'i'in obliquo' also standardly occurs in Ockham to characterize the semantical duality of connotative terms in general. Connotative names, he would say, 'signify something in recto and something else in obliquo'.19 Yet his use of the same phrases in the theory of definitions, although not unrelated, is markedly different. It is relevant for our present purpose to graspboth how these two uses are related and how they differ. When Ockham says that a connotative term signifies something in recto and something in obliquo, he istalking about the modes of signification of that term. To signify in recto is the same, in this vocabulary, as to signify primarily, and to signify in obliquo is the same as to signify secondarily. 'White', for example, is said to signify in recto - or primarily - all the white things, and in obliquo - or secondarily - all their whitenesses; 'father' is said to signify in recto - or primarily - all the fathers, and in ohliquo - or secondarily - all their children. The distinction, here, is purely semantical; it has to do with two permanent semantical features of every

50 88 THE ROLE OF NOMINAL DEFINITIONS 89 connotative term. By contrast, when Ockham uses the couple 'in recto'l'in obliquo' in the description of nominal definitions, he is speaking of grammatical features. A term is said to be taken (sumptus) in recto, according to this use, when it is in the nominative case, and in obliquo when it is in one of the so-called 'oblique' cases: the genitive, the dative, the ablative, or the accusative. This, indeed, was the standard terminology of grammar at the time. It primarily pertains to languages with declensions like Latin or Greek. Yet, what Ockham uses it to say about nominal definitions exceeds the limits of such languages. A term taken in obliquo in a certain context simply is a term which plays in this context the grammatical role of a complement, while a term taken in the nominative or in.recto is a term which is not, in the relevant context, playing the role of a complement. In the end, it all amounts to this: a nominal definition, according to Ockham, is a grammatically structured phrase that normally contains one categorematic term which is not, in this very phrase, playing the role of a grammatical complement, and at least one term which is. In 'animal having a child', which is supposed to be the nominal definition of 'father',20 'animal' is not a grammatical complement (it is in recto), while 'child', being the complementof the verb 'to have',is in obliquo. Now, the term taken in recto,as we have seen, should be the 'subject' of the connotative term. 21 What this means, I take it, is that it should circumscribe the domain of individuals among which the defined term is predicable. It should indicate, in other words, a natural group - presumably the smallest one - to which all the primary significates of the defined term belong: the group of animals, for example, is the smallest natural one in which the term 'father' has its primary significates. It is easy, then, to see the connection between Ockham's two uses of 'in recto': the term which is grammatically taken in recto within the nominal definition of a connotative term circumscribes the domain where what the defined term signifies in recto is to be found. As to the terms taken in obliquo within the definition, they are said by Ockham to be the names of the connotata of the defined terms (nomina connotatorum).22 Not the proper name, of course. What Ockham means by 'the name of the connotata' is a term which applies exactly to the connotata of the defined term. 'White', for example, is defined as 'a body having a whiteness'.23 'Whiteness', here, is the term in obliquo; it is true exactly of all and only the connotata of the defined term, namely the whitenesses. The same, to give only one more example, holds for 'father': 'child' occurs as a complement in its definition and applies exactly to its connotata, namely all the children. This isto be generalized, thus revealing what the relationis between Ockham's two uses of the phrase 'in obliquo': the terms taken in obliquo in the nominal definitions of connotative names should have as their extensions exactly what the defined terms signify in obliquo. We may add, I suppose, that if the defined term has more than one group of connotata - if, that is, it connotes some things under a certain mode and some under another mode -, then the correct nominal definition for such a term will normally include more than one term taken in obliquo, eachone occurring with,ln the definition as a distinct grammatical complement. 24 It is still possible in some definitions, as Ockham admits, that this role of naming the connotata of the defined term be assigned to a term in the nominative case. But this does not change the general picture since it happens only when the term in question occurs as a grammatical attribute introduced within a relative clause by a verb such as 'to be', sometimes accompanied by a negation.2~ Ockham'scxample is the negative term 'non-man', the definition of which, he says, is 'something which is not a man'.26 In this particular case, 'something' (aliquid) is the main term in recto - the subject of the connotative- and 'm.an' (homo) names the connotata of the defined term - namely all men- because, although it is also in the nominative, it occurs nevertheless in a subordinate grammatical position and within the scope of a negation. As to the verbs themselves, when Ockham says - without explanation- that they too can occur in nominaldefinitions,27 he is probably thinking of such verbs as 'to be' or 'to have', ~n so far as they can be treated as syncategorematic copulas. 28 The definitions of 'father' and 'white', for example, both contain an occurrence of 'to have'. Other verbs, such as 'to run' are taken to be analysable into a categorematic component plus a syncategorematic copula: 29 the connotative term 'runner', for example, is presumably best de.fined in mentalese, according to this view, not as i a running creature' (which contains no term taken in obliquo),but as something like 'a creature having (or making) a run', where the verb 'having' (or 'making:) is a syncategorematic copula, and where the grammatical complement';l run' designates, exactly like in the previous cases, the connotata of the defined term, namely the acts of running, considered here as real qualities. 2. What defining amounts to: a reconstruction To this cluster oftheses must be added, of course, the two more familiar ones which formed the basis for Spade's reductionist reading of Ockham: (5) Every connotative term has a nominal definition. 3D (6) No absolute termdoes. 31 These six tenets are the pieces of the puzzle that any sound reconstruction of Ockham's theory of nominal definition should assemble. The redllctionist interpretation, by contrast, is incompatible with (I), (2) and (3) taken separately,32 and it provides ho account for Thesis (4). The reason for this inadequacy is that the interpretation in question simply projects on Ockham's idea of a nominal definition the Frege~Russell conception of what an 'explicit definition' should be. An explicit definition for a certain term '1', according to this modem tradition, is a phrase which is wholly synonymous with 'I' and substitutable for it in all contexts (except those, of course, where T is mentioned rather than used). Thus, when a given expression is 'explicitly' defined with the help of some other terms, it appears as a mere abbreviation for its own definition and becomes eliminable from the language in which it is so defined,. without any semantical loss. This, however, was not how Ockham conceived of nominal definitions in general. The key to what he expected from such definitions is given in a cameo formulation by the following sentence from the chapter 'On definition' of the Summa Logicae: 'A nominal definition is a complex phrase (oratio) explicitly indicating what is importl'd (importatur) by a certain single term (per unam dictionl'ni).'11 This is what we now must explicate. A signiticant feature of this characterization is that it contrasts the nominal

51 90 THE ROLE OF NOMINAL DEFINITIONS 91 definition as a complex phrase with the defined term, which is normally taken to be a simple expression. 34 The nominal definition of a connotative name, in Ockham's system, is expected to unfold the meaning of the defined term. Now, the complete meaning of a connotative term lies in its signifying some things under a certain mode, and some under another mode. What the nominal definition should do, then, in order to unfold such a meaning, is to explicitly - and separately - identify the primary significates of the defined term and its secondary significates, while disclosing in each case the mode under which they are signified by the defined term (in recto or in obliquo, affirmatively or negatively, and so on). The point of the operation is revealed by Ockham's use ofthe verb 'importare' in the above quotation from Summa Logicae I, 26: 'Importare' is the general verb Ockham standardly uses to indicate the ontological import of an expression: a term is said to 'import' whatever objects of the world it signifies in whatever way.35 The primary goal of a nominal definition, for Ockham, is to make explicit the ontological commitment which is to be associated with the normal use of a given term. That is the whole secret: it is all a matter of ontology. Which is why nominal definitions are so important within the framework of Ockham's nominalism. Their function is to make it perspicuous that the things that are referred to in one way or another by a meaningful connotative term are but singular things, and more precisely, singular substances and singular qualities. The role of nominal definitions is not to eliminate connotative terms, but to show, on the contrary, that they are ontologieally innocuous, even when they are ineliminable. Nothing in Ockham's nominalistic ontology prevents us from having simple connotative concepts in our basic mental apparatus. What the nominal definitions must show, precisely, is that such simple connotative concepts require no undesirable complexification of the ontology. How is that to be done? We can, I submit, extract from Ockham's own practice and explanations with respect to the nominal definitions of connotative names, a relatively precise set of conditions he wanted such definitions standardly to satisfy. The following three can be proposed as a fair approximation. (Cl) The nominal definition should contain - usually as its first component - a term which is not in a grammatically subordinate position within this definition. This is the term in recto ofthesis (4) above. It should have among its primary significates all the primary significates of the defined term. Its extension, in other words, should include the extension of the defined term. Consider our familiar examples of nominal definitions again. In 'animal having a child', the term 'animal' occurs in recto: it is a categorematical name not occurring here as a complement. Its primary significates - all the animals there are - include all the primary significates of the definedterm 'father'. In 'abody having a whiteness', itis 'body' thatplays this role: its primary significates are all the bodies there are, including, of course, the primary significates of the defined. term 'white', namely the white bodies. The general idea is that the term in recto should delineate, at the outset of the nominal definition, a usually wider - range of things among which the primary significates of the defined term are to be looked for: that of animals in the case of 'father', and that of all the bodies there are in the case of 'white'. A passage of the Ordinatio strongly suggests that this term in recto should in general be an absolute name. 1fi (C2) For each group of ('onnot(j((j of the defined term, there should occur within the definition one and only one grammatically subordinate categorematical term normally a term in obliquo in the medieval vocabulary - having as its own primary significates exactly these connotata ofthe defined term. Take 'white'. It has a single group of connotata, according to Ockham, namely all the whitenesses there are, and its definition is said to be 'a body (or something) having a whiteness'. It can be seen that the term 'whiteness', which occurs as the sole grammatical complement in this definition, has as its own primary significates exactly the individuals which happen to be, according to Ockham, the connotata of the defined term, namely the whitenesses. Or take 'father'. Its nominal definition is supposed to be 'animal having a child', where it is observed again that the term occurring in obliquo, 'child' namely (which, in this case, is itself a connotative term), has as its own primary significates exactly those individuals that are taken to be the secondary significates of the defined term, namely the children. More complicated cases could be devised if we turned to terms with several series of connotata,37 but the general principle would be the same: each grammatically subordinate name within a good Ockhamistic nominal definition should primarily signify exactly those individuals which are connoted under a determinate mode by the defined term. (C3) The nominal definition should connect its categorematic term in recto with the complements it contains, so as to reveal the modes under which the secondary significates of the defined term are connoted, and to make it perspicuous in this way how the primary significates of the term in recto have to be related in reality with the primary significates of the term(s) in obliquo, in order to be the primary significates of the defined term as well. Whether in mentalese or in external language, this is secured by grammatical structure and by the use of appropriate syncategorematic phrases (including copulas such as 'being' or 'having'). Let us return to the definition of 'white' once more: 'a body having a whiteness'. As suggested by Paul Spade, 'having', here, might be seen as a syncategorematic copula connecting the term in recto 'body' with the complement 'whiteness'}8 What is required from it in this circumstance is that its logic be adequate to express how singular bodies should stand with respect to whitenesses if they are to be among the primary significates of the defined term 'white'. It makes little doubt that Ockham's conception of such syncategorematic copulas remains largely undeveloped,39 but it would certainly be relevant for the correctness of a definition with 'having' as a copula that 'having' is asymmetrical in a way that 'being' is not. Since syncategorematic terms, according to Oc~ham, do not signify special determinate things of the world, their contribution to the definition can only rest upon such logical and grammatical properties. What we end up with, in short, is this: the correct nominal definition of a connotative name should precisely delineate (through the term(s) in obliquo) the connotata of the defined term, while revealing the modes under which they are connoted; it should provide, moreover, through the syncategoremata and the grammatical structure, a kind of algorithm for picking out the primary significates of the defined term among the primary significates of the term in recto. 40 Whatever does that job for a given term T will be a good nominal definition of T. This reconstruction fits well with the passages where Ockham is explicit about the role of the nominal definitions of connotative names,41 and with most of his actual examplesy It accounts, in particular, for the otherwise surprising Theses (I )-(4) above. which we can now return to (in the reverse order).

52 92 THE ROLE OF NOMINAL DEFINITIONS 93 First, it is easily seen in the light of CI-C3 why a nominal definition should normally contain a tefill in recto and at least one grammatically subordinate categorematical term,as stated in Thesis.(4): the latter identifies the connotata of the defined term (by C2), while the former delineates the group of things among which its primary significates will be found (by CI). Taken together, they indicate what must exist in the world for the defined term to be significant. Why now should there be only one correct nominal definition for any given connotative name, in the sense indicated by Thesis (3)? Well, this is precisely because all the terms occurring within a nominal definition have, by CI-C3, a very precise job to do: the term in recto normally corresponds to the smallest natural group ofthings among which the defined term is predicable, and the terms in obliquo must have as their extensions exactly the connotata of the defined term. The role of revealing in the indicated way the ontological commitment associated with the use of a connotative name leaves virtually no room.for manoeuvre, at least at the.level of mental discourse. One might object that since, according to CI and C2, what matter for the correctness of a nominal definition are only the primary significates (that is, the extension) of its categorematic components (plus of course,the syncategoremata), nothing in our set of rules prevents any particular categorematic term occurring in obliquo within a definition. to be replaced in this definition by some other coextensive, but non-synonymous, term (contrary to what Ockham says himself).43 But note that mere actual coextensiveness will not do: a nominal definition should hold for all possible worlds. And given Ockham's ontological atomism,.there. can hardly be, in his' epistemology, necessary coextensiveness between any two non-synonymous simple concepts" whether absolute or connotative. 44 Let us tum now to Thesis (2): a definition is not always synonymous with the defined term. At least two reasons can be given why a nominal definition satisfying CI-C3 will not, in general, be synonymous with the defined name. First, the definition is supposed to contain, in virtue ofci, a general term in recto, the primary significates of which will include - and usually exceed ~ those of the defineq teffi1. Thus more individuals are normally signified by the definition than by the defined term, and hence the two of them cannot be synonymous in Ockham's sense. A second reason is that,as I have repeatedly stressed, the terms in obliquo which are called for by condition C2 need not be absolute tenlls. What is required from them is that their primary significates precisely correspond to the secondary significates of the defined terms. Whether or not they have secondary significates themselves, and whatthese secondary significatesare, is not directly relevant for the correctness ofthe definition. It is bound to happen, then, that some parts of the definition will connote certain things that are not Gonnoted by the defined term (as in the definition of 'father' as 'animal having a child'), and this brings about another excess in the meaning of the definition with respect to that of the defined term. Why, finally,is it impossible to completely eliminate all connotative terms from such nominal definitions, as stated by Thesis (I)? This is also something we can now clearly understand by reflecting upon the role of the terms in ohliquo within the nominal definition according to C2. What the complements should do, we have said, is to delineate exactly the ('onnotata of the defined term. But there is no reason why the ('of/f/otala of any connotative term should coincide with a natural kind. or, if you prefer, with the significata of an absolute term (or combination of absolute terms). Only a connotative, in many cases, willbe able to play the role expected from a term taken in obliquo within a correct nominal definition. Take a relational term like 'smaller', for example. Its connotata, presumably, are all the things that are such that something else is smaller than they are. This, however, does not correspond to any natural kind, and there is no absolute concept signifying precisely these individuals. If the complete nominal definitionof sucha,term is to satisfy C2, it will have to use as a complement its own correlative, namely 'larger', which is a connotative too. Theses (5) and (6), finally, also naturally follow from conditions CI-C3: every connotative name will have a nominal definition of the required form, since every such name, as Ockham frequently insists, signifies something in recto and something in ohliquo. No absolute term, by contrast, can be given such a definition: having no connotata, it could nevef receive a definition complying with C2 and C3. The whole network of these apparently disparate theses, then, is seen to be highly cohesive when we realize that the primary function of a nominal definition for Ockham is to explicate the ontological import of the defined term. 3. Some consequences Three consequences can be drawn, which are of special interest for the ongoing discussions on Ockham's theory of concepts. 3.1 Definitions and abbreviations Since a connotative name is notin general synonymous with its nominal definition, itcannot be seen as a mere abbreviation for it, whether in mental, spoken or written languages. Ockham, on the other hand, does admit of the possibility of spoken or written conventional abbreviations for any complex sequence of terms whatsoever: 'For the speakers of a language can,ifthey wish, use one locution in place of several. Thus, in place of the complex expression "every man", I could use "A"; and in place of the complex expression "man alone", I could use "B", and so on with other expressions.'45 Such conventional abbreviations, admittedly,.will be entirely equivalent in meaning with the abbreviated phrases and they will be substitutable for them in every context (short ofmaterial supposition), since that is what they are instituted for. But this should not be conflated with the having of nominal definitions by normal connotative names. For one thing, abbreviations can.be used inplace of any sequence, and not only of those that contain, in conformity with Thesis (4), a term in recto and at least one gmmmatically subordinate term. Take Ockham's second example in'the passage from Summa Logicae 1,8 just quoted: 'b' in this case is introduced as an abbreviation for 'tantum homo', which does not have the required form for a regular nominal definition. The abbreviating procedure, indeed, can be used even ifthe abbreviated phrase has no logical or grammatical unity. Another of Ockham's examples, just a few lines further, is that of an abstract term like 'humanitas'. In some of its uses, Ockham says, it is a mere abbreviation for'homo necessario' ('man necessarily'),4tl Yet the phrase 'homo f/c(,c.i',\'(/rio' has in itself no logical or grammatical unity:

53 94 THE ROLE OF NOMINAL DEFINITIONS 95 except when taken in material supposition, it cannot be the subject or the predicate of a well-formed sentence and cannot, properly speaking, supposit for anything. The two components 'homo' and 'necessario' simply happen to follow each other in some sentences (as in 'homo necessario est animal': a man necessarily is an animal), and only in such sentences can they be replaced by the abbreviation. In cases such as the latter, where the abbreviated phrases are not in themselves significant units, the resulting abbreviations are not connotative names. Strictly speaking, they are not names at all, as comes out from Ockham's lengthy analysis of the abstract terms 'punctus', 'linea', 'instans', 'mutatio', 'generatio', and so on in his Tractatus de Quantitate. 47 Such words, he insists, are introduced 'in place of an expression or of something composed of several words', but 'they do not have precisely the value of a name';48 properly speaking, they cannot be the subject or the predicate of well-formed sentences and they cannot supposit for anything. 49 These limitations, of course, do not hold in all case~ of abbreviations: if a single letter, let's say, was introduced as an abbreviation for an absolute or a connotative name, it could properly be used as a subject or predicate and have a personal supposition just as the original name could. The general point, however, is that the relation of an abbreviation with the abbreviated sequence is entirely different, in Ockham's view, from the relation of a connotative name with its nominal definition. The abbreviation process merely aims at a gain in brevity, in elegance or in poetic rhythm. 50 A nominal definition, on the other hand, even ifit can be used too as a procedure for introducing new terms in conventional languages, normally has a completely different function in Ockham: that of revealing the ontological import of a given term, whether mental or conventional. 3.2 Possession ofconcepts and knowledge ofdefinitions It has often been assumed that a nominal definition, for Ockham, is what a person should know if she is to be attributed the corresponding concept. 51 We can now see that this cannot be generalized. Although Ockham is not explicit on the point, it follows from the theory of nominal definition presented above that, in the case of simple connotative concepts at least, to have the concept is not the same as to know what its nominal definition is. You could have naturally acquired a simple relational concept such as 'similar' or 'smaller', without being able to tell at the first blush and with certainty what its correct nominal definition is, even if such a definition exists, and even if you do possess all the terms that are required to build it. This follows from the very notion of what a nominal definition is for, as we have reconstructed it. The objective of such a definition, we have said, is to make explicit the ontological import of the defined term. Now, you could be mistaken about the exact ontological import of one of your own simple concepts. This happens to realist philosophers all the time, according to Ockham! They typically think that a concept like 'white', which is presumably a good example of a simple connotative concept directly acquired on the basis of intuition,52 refers in some way to some single universal property of whiteness. By Ockham standards, they are mistaken about the significates of the simple concept 'white', and consequently, about what its correct nominal definition is. Yet they can hardly be denied the possession of this concept. The trouble with realist philosophers is not that they lack certain ordimlryconcepts like 'white'; it is, rather, that they misunderstand the ontological import of these concepts. Which amounts to say that they are wrong about their nominal definitions. One could object that there are a number of passages where Ockham insists that the knowledge of the correct nominal definition of a given connotative name is presupposed by any sound use of this name in demonstrations or in disputations.53 This, however, does not entail that the possession of the corresponding concept is impossible without the express knowledge of the definition. Disputations and demonstrations are scientific activities conducted, in general, in some public conventional language, and their scientific and interpersonal character do require, no doubt, that all the participants have a common and clear understanding of the words they use. But the point, here, is that we might possess certain simple concepts which we are not ready to make use of in such scientific arguments, precisely because our grasping of them has not yet been rendered explicit enough. When Ockham writes, for example, that 'the significate of a word cannot be known without the nominal definition, if it has one',54 he presumably uses a strong sense of the verb 'to know' (scire), and he uses it in connection with what the significate of the connotative term is. In other words, what he is saying is that we cannot have a clear knowledge of what things exactly are signified by a given connotative term without knowing the nominal definition of that term. This should not surprise us: it is precisely what a nominal definition is for, according to the theory reconstructed above. What should be concluded is not that we cannot have a concept without knowing its definition, but rather that if we do have a concept without the definition, our speculative mastering of this concept is imperfect. We might, for example, applyit correctly to its most typical instances in everyday life, without being able to explicate it, and without, consequently, being able to make a rigorous use of it (or of the associated spoken or written words) in scientific demonstrations or academic disputations. It thus turns out - even if he was not quite clear about it himself - that Ockham's theory of connotative concepts and nominal definitions challenges the idea that the conceptual content of the mind is always immediately perspicuous to the thinking agent. 3.3 Real orderings According to condition C3 above, a good nominal definition for a given connotative name should show how the primary significates of this name are connected in reality with its secondary significates. This implies that real singular things are ordered in the world in certain precise ways independently of the human mind, and that the human mind can sometimes acknowledge such connections. Ockhamindeed is quite explicit that ordering does not result, in general, from a merely creative work of the intellect: 'it is not the intellect', he writes, 'which brings it about that one thing is similar to another, any more than itbrings it about that Socrates is white orthat Plato is white'.55 In the Summa Logicae, he enumerates a number of ways in which something can be prior to something else in reality: 'it can temporally precede the other, or it can exist without the other but not conversely, or it can be more perfect than the other'.56 This is an incomplete list, no doubt. Things can be really ordered for Ockham in many ways, without the human mind having anything to do with it: temporally, spatially, causally, axiologically, rnereologically (as parts to wholes),

54 96 THE ROLE OF NOMINAL DEFINITIONS 97 hylemorphically (as fonns to chunks of matter), or inherently (as accidents to substances). What is required for thought and knowledge to take place is that the mind be organized in such a way as to be able to recognize in certain cases that.certain things are ordered in certain ways with respect to certain others, that the one, for example, is larger than the other, or that the one inheres in the other, and so on. When such a reccignitionoccurs on the basis of an intuitive grasping, it involves, as we shall develop in the next chapter, the fonnation of a' simple connotative concept associating in some precise way its primary significates with its secondary significates. This is what the nominal definition is supposed to express, in virtue of C3, by its grammatical structure and its syncategorematic components. Ockham is committed to admit, consequently, that the human mind has innate capacities for grasping, in at least some respects, how things are ordered outthere in the world, and to map these orderings into the very fonnation of its connotative concepts. This is no breach to Ockham's distinctive brand of nominalism,. since, as he sees it, it does not require the positionof orders or relations as distinct things in addition to singular substances and qualities. A relation, he.often says, is not an additional 'small thing' (parva res) in between the relatay The spatial order, for example 'involves only absolute things' such that one of them is more distant than another from the same thing, or closerthan another, without any spatial connection inhering in any of them as a distinct entity. 58 An order is not like a bond between two bodies, 'as if these bodies would not be ordered.,. without such an intermediary, as Simplicius has imagined in his commentary on the Categories'.59 In general, then, Ockham's theory of connotative concepts and nominaldefinitions.implies that singular things are ordered in certain ways, but this does not add to the ontological commitments he associates with such concepts since the orders themselves 'are not countenanced as extra entities. What should be said, strictly speaking, is not that there are orders in the. world, but that things are ordered. The difference, here, being between the correct use of a complete true sentence (for example, 'things are spatially ordered') andthe - strictly speaking, incorrect - use of a pseudo-name such as 'order'. That both theses - things are ordered, yet no order exists - can be held together rests on how Ockham understands what the truth-conditions are for relational sentences: these do not require any absolute things in the world other than singular substances and singular qualities, precisely because connotative names (including relational ones) signify nothing else, whether primarily or secondarily.60 Of course, condition C3, as I have already remarked, is not explicit enough and it remains mysterious to some degree how the grammatical and logical structure of a definitional phrase is supposed to'" mirror reality. Should mental language.be endowed, for example, with a special syncategorematic copula for each type of ordering thatthe mind is inflately able to recognize?6j Or should we be content with definitions' being incorrigibly. imprecise in this.regard (with, for instance, only one copula being available for all asymmetrical orderings)? Those are questions that are raised by Ockham's approach, but that he never directly addressed. That he didn't do so, however, is not surprising. They would have required an in-depth inquiry into the structure of the human mind, which he simply did not have the tools to achieve. And however important they may he in themselves, they were not, on the whole, that central to his own purpose. His main concern in the lheory of nominal definition, as in many other parts of his system, was with nominalism, let me repeat, and he was more interested, consequently, in the role of categorematic tenns within definitions (conditions Cl and C2) than with the role of syncategoremata or grammatical structure (condition C3), for it is the categorematic terms in the definition that ultimately reveal, in his view, the neat ontological import of the defined tenns, by identifying both their primary and their secondary significates. As to the remaining problems about the ways in which nominal definitions are supposed to reveal how the primary and secondary significates should be connected with each other, what Ockham is committed to is that they can be solved in principle without further ontological commitments. This, it seems to me, is not obviously doomed to failure, contrary to the reductionist programme Spade - anachronistically - read into the Venerabilis Inceptor's theory of nominal definitions. Notes For a good account of the so-called 'Classical Theory','see Laurence and Margolis 1999, pp 'It would be difficult', th,e authors write, 'to overstate the.historical predominance of the Classical Theory' (p. 10). Fodor 1998 devotes two chapters (pp ) to the criticism of such approaches. Spac!e 1996, p Spade 1998, p. 355 (the italics are Spade's). What he calls here the 'moderate reductionist program' holds that 'any statement containing connotative terms can be paraphrased by (and so in that sense "reduced to") a strongly equivalent (perhaps even strictly synonymous) statement that does not contain connotative terms, but only absolute categorematic terms plus syncategoremata' (ibid., p. 350). See SL I, 26: 'De definitione: quot modis dicitur definitio?', OPh I, pp See the texts quoted above chap. 4, nn. 28, 30 and 31. The interdefinability ofcertain terms was accepted, long before Ockham, by Porphyry; see Porphyrii lsagoge. Translatio Boethii, ed. Minio-Paluello 1966, p. 9: 'Nosse autemoportet, quoniam et genus alicuius est genus et species alicuius est species, idcirco necesse est et in utrorumque rationibus utrisque uti..' Porphyry's point is that 'genus' and 'species' are correlative terms and therefore interdefine each other. See above chap. 4, sect, 3. See Michon 1994, pp ; Spade 1996, pp ; Gaskin 2001, p.239, n. 60. See Exp. sup. Elench. I, 20, OPh III, p The text is quoted above in chap. 4, n. 44 and Michon's objection to its use in this context is rebutted in the same footnote. It is true that Ockham sometimes writes that the definition and the defined term 'signify the same' (see, for example, the passage from SL I1I-3, 22,OPh I, p. 680 quoted above in chap. 4, n. 33), but, as we have seen, this should not be taken to mean that they are synonymous in. the strong sense. He also says, sometimes, that the definition' and the defined term 'are equivalent in signification' (aequivalent in signiflcando), but normally the statement is either qualified with an 'aliquo modo' (see Ord. T, Pro!., q. 3, OTh.1, p. 139), or limited to the special case of conventional abbreviations (as in Quaest. in Phys., q. 63, OPh VI, p. 568). Beretta 1999, pp makes much with a passage from SL I, 5l, where all relative terms are said to be equivalent in signification to the longer phrases which are their nominal definitions (OPh I, p. 171: 'Non enim quaerendum est omnibus ten11inis quid rei scd tantum quid nominis in multis, quales sunt omnes termini relativi ct nonnulli alii, qllol"ll1ll qlli/ihct al'qlliva/ct ill sigllifh:ando /ongal' orationi', italics arc minc J, hut this vcry chaptcr has hccn dccmcd inauthcntic hy thc cditors of Ockham's

55 98 THE ROLE OF NOMINAL DEFINITIONS 99 Summa (see the Introduction by G. Gal and S. Brown in OPh I, pp. 40*-41*); even though they believe that there is nothing in it which Ockham would not have approved of, not too much weight should be given to the details of the formulations found in it. 10. See above chap. 4, sect. 2, with, in particular, the texts quoted in nn. 12, 14-17, and See, for example, Quodl. V, q. 8, OTh IX, p. 511: '... nec nominibus synonymis correspondet 'pluralitas conceptuum'; as well as the other texts mentioned above in chap. 4, n See the passage from Quodl. V, q. 20, OTh IX, p. 556 quoted above in chap. 4, n. 40, as well as the discussion of Spade's objection tothis argument in n See the passage from Summ. Phil. Nat. I, 3, OPh VI, p. 162 quoted above in chap. 4, n The point was developed in details in chap. 4, sect Quodl. V, q. 19, OTh IX, p. 555: '... proprie loquendo unius nominis habentis definitionem exprimentem quid nominis est tantum una definitio explicans quid nominis... ' (trans!. Freddoso 1991, p. 464, with a minor amendment). 16. Ibid.:'... talis nominis non sunt diversae orationes exprimentes quid nominis habentes partes distinctas, quarum una pars significat aliquid quod non eodem modo importatur per aliam partem alterius orationis'. 17. SL I, 10, OPh I, p. 36: 'Et tale nomen [connotativum] proprie habet definitionem exprimentem quid nominis...' 18. SL III-3, 26, p. 691: 'Connotativa definiuntur per sua subiecta sumpta in recto et per nomina connotatorum sumpta in obliquo, vel per verba.' See also Exp. sup. Elench. II, 16, OPh III, pp As Adams remarks (1987, p. 321), Ockham in SL I, 10, OPh I, p. 36, says that the nominal definition of a connotative frequently contains a term in' recto and a term in obliquo, leaving it open that it might not always be the case CEt tale nomen proprie habet definitionem quid nominis, et frequenter oportet ponereunum illius definitionis in recto et aliud in obliquo', italics are mine); this possibility will be accounted for below (see in particular the passage quoted in nn. 25 and 26). 19. Quodl. V, q. 19, OTh IX, p. 554: 'Definitio autem exprimens quid nominis proprie est de nominibus connotativis et relativis quae significant unum in recto et aliud in obliqud, sicut "album", "calidum", "pater" et "filius" sunt huiusmodi nomina.' 20. See for example SL III-3, 26, OPh I, p See the text quoted above in n Ibid. 23. See SL III-3, 26, OPh I, p. 691: 'Album autem definitur sic "album est corpus habens albedinem".' Ockham also loosely defines 'album' sometimes as 'aliquid habens albedinem' (for example, in SL I, 26, p. 88), but the definition with 'corpus' is more precise. 24. You mightthink, although this is not one of Ockham's examples, of a term like 'gift', which connotes, presumably, both the donators and the receivers, but under different modes. Its Ockhamistic nominal definition, then, should include two distinct grammatical complements. It would read a bit like this: 'something that pa'sses from a donator to a receiver'. 25. See SL II, 12, OPh I, p. 283: '... omnes tales termini [= negativi, privativi, infiniti] sunt vere connotative, eo quod in eorum definitionibus exprimentibus quid nominis debet poni aliquid in recto et aliquid inobliquo, vel in recto cum negatione praevia' (italics are mine). 26. Ibid.:'... et definitio istius termini "non-homo" est ista "aliquid quod non est homo", et sic de ahis.' 27. See above n R. In SL I, 31, OPh I, p. 94, the copula is defined as 'verbum copulans praedicatum cum subiecto'. It can be galhered from Ockham's own practice thai this includcs nol only 'to be', but any other connecting verb such as 'inest' (see, for example, SL II, 11, OPh I, p. 281 where 'inest' is treated as the copula of such sentences as 'Sorti inest albedo'), 'incipit esse' (see SL 1,75, OPh I, pp ), and even verbs like 'vidit' (see SL II, 21, OPh I, p. 321). Some among these, like 'esse', and maybe 'habere', turn out to be purely syncategorematic terms in the sense that they do not have any special significata of their own (see Quodl. VI, q. 29, ad 1, OTh IX, p. 695: '... dico quod unio extremorum propositionis in mente est conceptus syncategorematicus verbi copulativi sive copulantis subiectum cum praedicato'; or again Exp. in Perih. I, 2,4, OPh II, p. 389: 'hoc verbum "est" neque est signum rei'). More on this below in chap. 8, sect Some verbs, Ockham says, signify certain things determinately (Exp. in Perih. I, 2, 2, OPh II, pp ). 'To run' is a good example: it primarily signifies the runners and connotes their running acts. Such verbs are normally equivalent to a combination of 'to be' plus a categorematic participle (SL 1,3, OPh I, p. 11: 'Nam verbum et participium verbi sumptum cum hoc verbo "est" semper videntur in significando aequivalere'). When it comes to logical analysis, the composite form 'to be + participle' is more fundamental (see, for example, SL II, 21, OPh I, p. 321, where Ockham discusses the case of propositions with adverbial determinations: 'Et tunc in tali casu... opportet resolvere verbum in suum participium et in hoc verbum" est "'). 30. See, for example, the passage quoted above in n See SL I, 10, OPh I, p. 35: 'Immo, proprie loquendo talia nomina [= nomina absoluta] non habent definitionem exprimentem quid nominis... ' 32. The incompatibility of Spade's original interpretation with Theses (1) and (2) is straightforward; Spade held both that what he called the 'fully expanded nominal definition' of a connotative term contains no simple connotative term (Spade 1975, pp ), which is the negation ofthesis (1), and that every connotative term is synonymous with its nominal definition (ibid.), which is the negation of Thesis (2). As for Thesis (3), the case is more indirect. A connotative term, according to Spade, frequently has more than one. correct Ockhamistic nominal definition, since not all nominal definitions, he thinks, are 'fully expanded'; some of them, that is, are taken to contain connotative terms which have not yet been replaced by their own nominal definitions (Spade 1975, p. 69). This is not enough, though, to contradict Thesis (3) as Ockham understands it, since it is explicitly held by Ockham that there can be several variants for the nominal definition of a given connotative name, on the condition that they all be strictly synonymous with each other, a condition which Spade readily admits (Spade 1975, p. 66). The problem, rather, is that Spade's notion of synonymy among complex phrases differs from Ockham's. According to Ockham, two different nominal definitions are synonymous variants ofeach other ifand only ifnone ofthem contains a part which signifies under a given mode something which is not signified under the same mode by some corresponding part of the other (see, for example, the passage from Quodl. V, q. 19, quoted above in n. 16). Now, two nominal definitions one of which contains a connotative term which is replaced in the other by its own nominal definition will not always be synonymous in this sense, since, as we have previously.stressed, some part of the replacing nominal definition might very well signify something under a certain mode which is not signified under the same mode by any part of the original definition. Suppose, for example, that a connotative term like 'white' occurs in some not yet 'fully expanded' nominal definition for a given term T. When 'white' is replaced within this definition of T by its own nominal definition ('a body having a whiteness', according to Ockham), there will be a part of the resulting definition for T (namely 'body' in this case) which will signify under a certain mode certain things that might not be signified at all by any part of the original definition (namely the bodies that arc not while). And IIIis contradicts Thesis (3) as Ockham formulates it..n SI. I. 26, OPh I. p. XX: '[)efinitio autcm exprimens quid nominis est oratio cxplicite declarans quid per unam diclioncm im[1l1rl(//iii'... (thc translation and italics arc minc).

56 100 THE ROLE OF NOMINAL DEFINITIONS 101 In what immediately follows, Ockham mentions that a term like 'white' 'signifies the same' (significat idem) as its nominal definition. As we have previously seen, however (in chap. 4, sect. 3), this does not mean that they should be synonymous in the strong sense, but only that every significate of 'white' must be signified in some way by (some part of) the definition, which is a much weaker condition. 34. See above chap. 4, n See for example SL I, 33, OPh I, p. 96: 'Aliter accipitur "significare" communissime quando aliquod signum... aliquid importat, sive principaliter sive secundario, sive in recto sive in obliquo, sive det intelligere sive connotet illud, vel quocumque alio modo significet...' (the italics are mine). 36. See Ord. I, dist. 2, q. 9, OTh II, p. 316: [N.B. - this passage is about denominative and quidditative terms, but the former, for Ockham, are all connotative, and the latter are absolute]: '... omnis conceptus denominativus habet definitionem exprimentem quid nominis, in quo ponitur aliquid in recto et aliquidin obliquo. Tunc quaero de una parte [or: de prima parte, according to ms. E] illius definitionis: aut habet definitionem consimilem exprimentem.quid nominis, aut non. Si non, habetur propositum, quia talis necessario est quidditativus; aut habet talem definitionem exprimentem quid nominis, et quaero tunc de partibus sicut prius, et ita vel erit processus in infinitum vel stabitur ad aliquem conceptum quidditativum praedicabilem de illo de quo primus conceptus denominativus praedicabatur' (italics are mine). That this argument has to do only with the first part of the definition, the term in recto - as explicitly proposed by manuscript E- becomes clear in the last sentence of the quotation (which I italicized). What. Ockham requires is that this term in recto should ultimately be an absolute name predicable of everything of which the defined term is itself predicable (such as 'body' is with respect to 'white', or 'animal' with respect to 'father'). 37. See above n Spade 1990, p See below chap. 8, sect We now see how it can happen, in some cases, that a connotative term ends up with no primary significates at all. Take the figment term 'chimera', for example. Its definition according to Ockham is 'animal composed of a goat and a cow' (SLI, 26, OPh I, p. 88: 'animal compositum ex capra et bove'). The grammatical complements 'goat' and 'cow' in this definition respectively designate the two series of connotata ofthe defined term, and 'animal' plays the role of the term in recto, the extension of which is expected to include that of the defined term. In this case, however, there are no animals inthe world which are related in the required way with both goats and cows. The defined term, therefore, has no primary significates. Note that, contrary to what Spade sees as a problem in this case (1975, p. 72), nothing here requires that all the significates of 'animal' -namely all the animals there are - should be signified in any way by the defined term 'chimera'. Spade only thinks so because he (wrongly) assumes that whatever is signified by some parts of the definition should also be signified in some way by the defined term. 41. Most notably in SL I, 10, OPhI,pp. 35-8, SL I, 26, OPhI, pp. 88-9, SL III-3, 26, OPh I, pp , and Quodl. V, q. 19, OTh IX, p Spade's interpretation, by contrast, could not account for the fact that many among the most paradigmatic examples of Ockhamistic nominal definitions still include connotative terms without Ockham expressing uneasiness about it in any way. 43. See the passage from Quod/. V, q. 19 quoted above in n Among absolute concepts, two necessarily coextensive terms would ipso facto be synonymous. In the case of simple connotative concepts, if they do not have the same connotata, God could always bring it about that their primary signitkates ditrer. And the same is true, a fijrtiori. for any pair composed of an ahsolute concept and <I simple connot <II ive one. 45. SL I, 8, OPh I, pp : 'Possunt enim utentes, si volentes, uti una dictione loco plurium, sicut loco istius totius "omnis homo" possem uti hac dictione a et loco istius totius "tantum.homo" possem uti hoc vocabulo b, et sic de aliis' (transl. Loux 1974, p. 65). 46. SL I, 8, OPh I, p See Tract. de Quant., q. 1, OTh X, pp , esp See on this text (also called De Sacramento Altaris) Stump 1982, esp. pp Ibid., p. 24: '... una dictio ponitur loco orationis vel alicuius compositi ex multis dictionibus...;' and a little further down the same page: '... non sunt praecise habentia virtutem nominis.' See also p. 30: '[...] hoc nomen "instans" et hoc nomen"generatio" ethuiusmodi non sunt nomina finita habentia finitas significationes, quasi significent praecise res quasdam pro quibus habeant supponere, ad modum quo talia nomina "honld", "asinus", "albedo", "nigredo", "calidum", "frigidum" habent finitas significationes.' Note that the pseudo-names in this last passage are contrasted not only with absolute names such as 'homo' or 'albedo', but also with connotative ones, such as 'calidum' ('hot') and 'frigidum'('cold'). Note also that not all pseudo-names, in Ockham's view, are precise abbreviations. In some cases,. they are only loosely connected with the complex expressions which they come from and they can be expanded, then, only in the context of whole sentences, with maybe various explications in various contexts. See, for example, how Ockham deals with the term 'generatio' in Tract. de Quant., q. I, OTh X, p. 31: 'Sed ista propositio "generatio est in instanti" debet recipi sub isto intellectu:.quando aliquid generatur, non generatur pars ante partem, sed totum simul generatur, ut ista propositio brevis "generatio est in instanti" ponatur causa brevitatis loco istius propositionis longae... ' Ockham, in such analysis, comes very close to the modem idea of a 'contextual definition'. 49. Ibid., p. 24: '... sine omni figura et improprietate sermonis non possunt esse extrema propositionis distincta a copula'. 50. Ibid., p. 30: 'Sed sunt quaedam nomina derivativa a verbis vel aliis partibus orationis, vel consimilia talibus, quae non sunt instituta nisi causa metri vel ornatus vel brevitatis locutionis.' See also Exp. in Phys. III, 2, OPh IV, pp , and Quaest. in Phys., q. 63, OPh VI, pp David Chalmers (1999, pp. 84-6) has argued that 'in the interests of efficient functioning of the mental system', Ockham should also admit the possibility of 'internal abbreviations' for complex mental phrases (p. 85; the italics are Chalmer's). This might very well be so, and whether ot not it would be a case of synonymy in the strong Ockhamistic sense, the Venerabilis Inceptor actually seems to do so when he considers the possibility that certain simple mental acts are 'equivalent' to complex propositional acts (see the passages from Quaest. in Phys., q. 6, OPh VI, p. 410 and Exp. in Perih., Prooemium, 6, OPh II, p. 358, quoted above in chap. 2, n. 86, and the corresponding discussion in the body of the text). 51. See, for example, Brown 1996, p. 85 (about Ockham): 'Nominal definitions are, however, essential to grasp the meaning of a connotative term. We could not conceivably be said to have the concept of a bachelor if we did not know that.bachelors are never married men or the concept of a father without knowing that a father is a male parent.' 52. More on this in chap. 6 below. 53. See, for example, SL III-2, 28, OPh I, pp : '...sed ignoratadefinitione exprimente quid nominis non potest quis cum alio disputare. Et ideo quando quis addiscit significata vocabulorum, tunc addiscit definitiones exprimentes quid nominis...;'or SL III-2, 34, p. 570: 'Et est dicendum quod generaliter definitio exprimens quid nominis non potest demonstrari de definito, sed ista praesupponitur omni demonstratione et omni syllogismo.' 54. SI , 14, OPh I, p. 530: '... significalum vocahuli non potesl scire sine definitione exprimenle quid nominis. si hahel quid nominis' (italics arc mine).

57 Ord. I, dist. 30, q. 5, OTh IV, p. 385: '... ad hoc quod unum sit simile alteri non plus facit intellectus quam facit ad hoc quod Sortes sit albus vel quod Plato sit albus'. See Ord. I, dist. 30, q. 1, OTh IV, pp : 'Et ideo concedendum est quod intellectus nihil facit ad hoc quod universum sit unum vel quod totum sit compositum vel quod causae approximatae causent vel quod triangulus habeat tres etc., et sic de aliis, non plus quam faeit ad hoc quod Sortes sit albus vel ad hoc quod ignis sit calidus vel aqua frigida.' See also Quod!. VI, q. 25, OTh IX, p SL 1II-2, 14, OPh I, p. 529: 'Nam inter res dicitur una prior alia: vel quia praecedit earn tempore, vel quia potest esse sine ea et non e converso, vel quia est perfectior ea.' 57. See for example Quod!. VI, q , OTh IX, p For detailed presentations of Ockham's theory of relations, see Adams 1987, chap. 7, Henninger 1989, chap. 7, Beretta 1999, Michon See Quod!. VII,q. 8, OTh IX, p. 729: 'Sed ille ordo [= spatial order] importat solum ipsa absoluta quae non faciunt unam rem numero, inter quae unum ab eodem plus distat et aliud minus, et unum propinquum alteri et aliud distare plus vel minus sine omni respectu inhaerente, ita quod inter aliqua sit medium et inter aliqua non.' It is remarkable that Ockham's rendering of the spatial order in this passage involves at least three distinct ordered things (one of which being closer to - or further from - the other than the third). 59. Ibid., pp : '... dico quod ordo et unitas universi non est quidam respectus, quasi quoddam ligamen ligans corpora ordinata in universo ad invicem, quasi illa corpora non essent ordinata nec universum vere esset unum sine tali respectu secundum imaginationem Simplicii super Praedicamenta'. 60. On the truth-conditions of propositions with connotative names in them, see SL II, 11, OPh I, p Such truth-conditions, Ockham explains, can be given by a conjunction of exponentes. This is not, as we have seen, because only these explicating propositions - the exponentes - are really present in the mind when such a proposition is uttered, but because, just like the nominal definitions in the case of isolated names, they make explicit the ontological import ofthe proposition as a whole: '... quaelibet propositio quae habet talem terminum [= connotativum], est habens exponentes exprimentes quid importatur per talem propositionem' (p. 281; the italics are mine). 61. See below chap. 8, sect. 3. Chapter 6 Cognition and Connotation As to what is going on in the mind, the result of the preceding chapter, admittedly, is negative: nominal definitions, contrary to what Paul Spade and others had thought, are not meant by Ockham to reveal in general how connotative terms are mentally constructed out of more basic conceptual units. They are tools, instead, for the elucidation of the ontological import of such terms, and this role can be fulfilled quite independently of how the connotative names are acquired. The epistemological significance ofsimple connotative concepts, then, still has to be brought out: how are they implemented? And what contribution are they supposed to make to our knowledge of reality? Such will be the object of the present chapter. The starting point, once again, will be provided by Paul Spade's stimulating formulation of the problem (section 1). The basic idea for sorting out the Ockhamistic answers to Spade's questions, as we will see, is that the Venerabilis Inceptor accepts, in a variety ofcases, an intuitive apprehension ofordered n-tuples, which, given the way our mind is built, triggers the formation of simple connotative concepts (section 2). How concepts thus acquired are liable to adequate nominal definitions of the sort described in the preceding chapter, will then become apparent (section 3). This explanation, however, cannot hold good for all connotative terms. Many of them - such as 'father' - are not ascertainable on the direct basis of experience and must, presumably, be introduced in some derivative fashion on the basis ofmore primitive terms. The question will have to be discussed, consequently, whether Ockham's theory of concepts remains or not, in the end, a variety of the socalled Classical View, according to which, remember, the vast majority of our concepts are present in the mind in the guise of complex definitions (section 4). 1. Spade's questions Although Paul Spade has recently granted that Ockham did in fact countenance simple connotative concepts, he still sees this as a mistake on William's part, given the rest of his theory. One of his main motivations for thinking so has to do with nominal definitions; this is what I have examined in detail in chapter 5. But it is not the only one. A second important reason he gives stems from the theory of knowledge. Here is, in his own words, what he takes the problem to be:... there is another reason to be worried about the possibility of simple connotative concepts: they threaten to interfere with the epistemology of mental language... Obviously there are many questions to ask here. What exactly does the mind do to produce these connotative concepts'! [fit does not usc absolute concepts as parts of the connotative concepls it produces. then why arc the absolute concepts required for this mental operation al all'! And exactly what sense are we 10 make out of the 'semanlic col1lpleltily' of un

58 104 COGNITION AND CONNOTATION 105 metaphysically simple concepts?... In my view, any full account of Ockham's theory of connotation must deal with these epistemological issues, since Ockham himself certainly uses connotation-theory for epistemological purposes.' Spade is undoubtedly right: these are questions that should not be avoided. Before I set out to answer them, let me explain a little bit more what their relevance is. Three different related questions surface in the passage just quoted: (1) How is the mind supposed to produce simple connotative concepts? (2) Why are absolute concepts needed in the process? (3) What sense are we to make of the semantic complexity of allegedly simple concepts? Let us start with the third one. Why does it arise? Well, Ockham's connotative terms, whether spoken, written or conceptual, do have in effect a certain semantic complexity, in the ~ense, precisely, that they have a primary signification (they are usually - true of certain things in the world) and a secondary signification or connotation (they - always - obliquely refer the mind to some other things). If we are to speak of simple connotative concepts, consequently, just as we speak of simple connotative words (like 'white' or 'horseman'), we have to reconcile this alleged simplicity with the semantic complexity which is taken to be characteristic of connotative terms. The reason why Spade speaks of 'metaphysical' simplicity in this context is that Ockham, in his later theory at least (which is the one I am interested in here), treats concepts asreal things. They are seen as intellectual acts, and such acts, for Ockham, are real singular qualities of singular minds. My concept of 'horse'.is something real in me, and your concept of 'horse', dear reader, is something real in yourself. Ockham, in fact, does not use any such phrase as 'metaphysically simple', but he does contrast simple concepts in the mind with complex sequences such as propositions or phrases. How, then, can semantically complex concepts still be simple objects? This question, while relevant, is the least difficult of thethree and I might as well indicate at once what the answer will be. To say that a concept is simple in Ockham's sense is not to say that it is devoid of any internal complexity. As we saw in chapter 2, to say that a concept isa simple sign, in Ockham's vocabulary, is merely to say that it is not composed of several parts each one of which is independently endowed with a signification. 2 Compare withspoken words. A wordlike 'father', let's say, is said to be simple, in Ockham's Aristotelian terminology, because it is not composed of several other spoken words. This does not prevent it from having an internal structure. For one thing, it certainly displays some phonological complexity. And even some sort of semantical complexity in so far as it can be broken down into a stem and an ending, each one of which having a distinctive semantical role to play. Yet it is simple if we contrast it with such phrases as 'the father of Socrates' or complete propositions such as 'my father had a moustache', where several components independently have a representative value. The situation needs not be different with concepts. The expression 'metaphysical simplicity', which Spade uses here, is a bit unfortunate in the context.' To say of a connotative concept that it is simple is only to say that it is not composed of several otherindependenlly signifying concepts. Although we still have to explain how this is possible, there is nothing deeply metaphysicalin the issue. Let us now tum to Spade's second question. Ifsimple connotative concepts do not have absolute concepts as their parts, he asks, why are absolute concepts nevertheless required for the formation of connotativeones? Spade, it must be said, raises this difficulty in connection with a particular hypothesis he. examines, according to which simple connotative conceptswould be mere mental abbreviations for their complex definitions. They would indeed be simple intellectual acts, according to this conception, but would be semantically equivalent to some complex expressions - their nominal definitions namely -'.-.. in which no connotative would occur. In view of what we have seen in chapter 5, this hypothesis does not appear any more tenable than Spade'soriginal rejection of all simple connotative concepts, since it continues to assume that all connotative terms are ultimately.eliminable according to Ockham (contrary to what we have labelled as Thesis 1 in chapter 5),4 and that there is normally no interesting semantical discrepancy between a nominal definition and the corresponding connotative term (contrary to Thesis 2). Yet the problem Spade raises with this second question is indeed a real one, for Ockham does consider thatthe formation ofa connotative concept presupposes in general the formation of at least one absolute concept. A clue to this is given in distinction 2 of the Ordinatio, when Ockham strives to show, in Scotus's footsteps, that there must be some quidditative concept univocally applicable both to God and the creatures. One of the arguments he uses in the process explicitly incorporates the claim that the possession of a denominative concept presupposes in general the possession of at. least one quidditative concept. Since denominative terms such as 'courageous', 'white', or 'creative' (which, in the spoken language at least, are concrete terms morphologically related with corresponding abstract ones such as 'courage', 'whiteness', or 'creation') are all connotative according to him, and since quidciitative terms are all absolute, this amounts to saying that the possession of some of the most typical connotative concepts presupposes that of some absolute concepts. Here is part of the passage: For example, I have this concept, that of 'creative being' namely, which I know to be denominative, and therefore a certain concept must be prepossessed to which I attribute this one, saying, for example, that a Gertain being is Gfeative; and it is Gertain that this Goncept to which this one is attributed is not denominative, or if it is, ~ither the process will go on infinitely or it will stabilize at some quidditative concepts.' Anybody who has an applicable denominative concept such as 'white' or 'creative' must also have an absolute concept of which the former is predicable within an affirmative particular proposition (such as 'body' is in the c;lse of 'white', or 'being' in the case of 'creative'). Spade's second question, then, is entirely appropriate: why should it be so? What, in other words, is the rationale for this epistemological priority of absolute concepts over connotative ones? As to the first question, finally - how does the mind produce simple connotative concepts'! - it is obviously the crucial point. Any account of Ockham's theory of concepts has to provide a detailed explanation for it. Fortunately, such an

59 106 COGNITION AND CONNOTATION 107 explanation is available on the basis of Ockham's own texts, and, as we shall see, the answers to Spade's questions (2) and (3) smoothly follow from it. 2. The acquisition of simple connotative concepts How simple connotative concepts originally enter the human mind is one of these issues on which Ockham changed position as he moved from the fictum-theory of concepts to the actus-theory. What I want to present here, as in most of this book, is the later - and final - position he adopted. But let me first say a word about the earlier one. It is very briefly expounded in a single passage of the Ordinatio, distinction 2, which has - to say the least - surprised the commentators who have noticed it. 6 Ockham there is discussing an objection raised against thefictum-theory of concepts: how can this approach account for the formation of syncategorematic concepts, connotative concepts and negative concepts? the objector asks.? Ockham's answer is the same for all three categories: what the fictum-theorist has to say, he thinks, is that all of these are implemented in the mind as a result of its internalization of external language. The mind naturally produces simple intellectual representations of external words, as it does of any other objects it comes in intuitive contact with; then it can, in addition, conventionally (ex institutione, Ockham says) endow these particular mental units with the very grammatical and semantic'll features of the words they represent, and use them accordingly within mental sentences. 8 Strikingly enough, this conception shares one of the most salient features of Spade's original interpretation of Ockham: no simple connotative concept, according to it, is naturally produced by the mind on the direct basis of intuition. 9 Yet itma,rkedly differs from Spade's reconstruction on several counts. Forone thing, it does not identify connotative concepts with their mental complex definitions, as Spade proposed, but with certain inner representations of words, conventionally used in special ways. And it does not require,. consequently, that a connotative term be strictly synonymous with its nominal definition. Whether ornot this approach is compatible with the ineliminability of relational terms from certain nominal definitions (as stated by what we called Thesis I in chapter 5) is not obvious, but it might very well be after all, if the phrase 'connotative concept' in this passage is taken - as seems probable to me - in what Ockham calls its na,rrow sense, according to which it excludes relational terms. lo If so, the explanation given in this passage would not be meant to account for the acquisition of rela,tional concepts, and the conventional connotative words which are supposed to serve as starters for the implementation of simple connotative concepts might be thought to be originally introduced in the external language on the basis of nominal definitions including in some cases both absolute and relational terms. II Be that as it may, this theory was never developed l2 and Ockham eventually renounced it in favour of a more unified approach, according to which the acquisition process for simple connotative and relational concepts is much more intimately related to the acquisition process for simple absolute concepts. This is what we will now be interested in. Let us recall how we normally acquire ahsolute concepts, according to Ockham. The whole process naturally starts when the agent gets acquainted with a physical object - a horse, let's say - through the senses: this is the sensory intuition. The same singular objectis immediately grasped by the agent's intellect: this is the intellectual intuition, which allows the agentto evidently know certain contingent truths about the objectinquestion (thatitpresentlyexists, for example). A generalconceptis then formed: the specific concept of horse in our example, which requires only one encounter with a horse. When the agent later meets with other animals, such as a cow or a dog, he or she forms a generic concept, that of an animal, for instance, or something in between like the concept of a mammal, all this, let me insist, being the result of natural causal processes. 13 The question that now faces us is this: how can the formation of simple connotative concepts fit into such a picture? And the answer I want to put forward is that in many cases - though not in all, admittedly - the process, for Ockham, will be very much the same as the one I just described, except that the intuitive grasping which is required as a starter should be an intuitive cognition not merely of a single individual, as in the example of the horse, but of a plurality of individuals simultaneously present to the agent's perception as a plurality. Although often neglected, some of Ockham's texts are quite telling on the matter. Let us carefully read, for example, the following passage from the Prologue of the Ordinatio: Similarly, an intuitive cognition is such that when some things [please note the plural here) are cognized, one of which inheres in the other, or is distant from the other or is standing in some other particular relation with the other, then straightaway in virtue of this incomplex cognition [note the singular] of those things [note the plural], it is known whether the thing inheres or not in the other, or is distant or not, and so on for other contingent truths... For example, if Socrates really is white, then this cognition of Socrates and of his whiteness in virtue of which it can be evidently known that Socrates is white, is said to be an intuitive cognition. 14 \ The starter, in the latter example, is a simple intuitive grasping of both Socrates and his whiteness simultaneously, and this is said to be enough for the cognizer to know that Socrates is white. This strongly suggests that the simple intuitive grasping of the two objects suffices to bring about the formation of the simple connotative concept 'white'. A related passage from the Quodlibetal Questions is even clearer:... a relative concept is caused by both extremes, posited simultaneously, prior to composition and division... Therefore, the order is as follows: When two whitenesses are seen, then, first, the specific concept of a whiteness is caused in the intellect; second, the concept of a similarity is naturally caused through the mediation of that specific concept. and, I claim, this happens immediately, from the whitenesses themselves, or from the cognitions of them; and only after that, at least in the order or nature, is a proposition formulated. 15 I grasp two ohjects simultaneously, two whitenesses let's say. As usual, I naturally form the relevant ahsolute concept, the concept of whiteness in this case. The important point. however. for our present discussion. is thai prior to any intellectual

60 108 composition or division, I also naturally form, according to Ockham, at least one simple connotative concept, that of 'being similar' in the chosen example. This is a general concept, with at least one of the perceived whitenesses among its primary significates and the other one among its connotata; it refers in the same ways to all couples that are such that one member of them is similar to the other one. 16 It is a simple concept in tne standard sense that no part of it is itself a concept; yet whatit refers the mind to are ordered couples the first member of which is a primary significate of this concept, and the second member a secondary significate. We can thus see why the formation of a simple connotative concept always (1) (2) (3) COGNITION AND CONNOTATION is the gist of Ockham's theory of simple connotative concepts. 109 concepts, indeed, are not even parts of the nominal definitions of the connotative concepts (for example, 'dog' does not occur in the definition of 'black'). But even when they happen to be parts of the definition, it does not make them parts of the connotative concept itself. Suppose I see a white dog; I then form at least two absolute concepts, 'dog' and 'whiteness', and at least one connotative, 'white'. It so happens in this case that one onhe absolute concepts involved, that of 'whiteness' namely; occurs - according to Ockham - in the correct nominal definition of the connotative concept (the definition of 'white' is supposed to be something like 'a body having a whiteness'). But the concept 'whiteness' in this case is no more a part of the concept 'white' than the concept 'dog' is.,its occurring in obliquoin the nominal definition of 'white~ simply warns us that whitenesses - real whitenesses, that is -are the secondary significates ofthis particular concept. And this, in Ockham's view, does not require that 'white' be a complex concept. The answers to Spade's three questions, then, are straightforward: How does the mind produce simple connotative concepts? By simultaneously grasping a plurality of individual things. Why are absolute concepts needed in the process? Because each one of the individual things involved triggers the formation of at least one absolute concept. These absolute concepts, however, are not parts of the connotative ones. How is the semantical complexity of these connotative concepts compatible with their alleged simplicity? Because the 'simplicity' of such concepts merely means that no part of them is independently endowed with a signification. This does not preclude them from signifying some things in a certain way and some others in another way. These answers make it clear why the ability to produce simple connotative concepts must have a major epistemological significam;e in the context of Ockham's theory. Their distinctive contribution to our cognition. is that they are naturally acquired general representations of ordered n-tuples. If our basic stock of general representations was limited to absolute concepts, we might be able to nicely categorize singular things, but not to reason or speculate about their being ordered in the world out there, which, of course, they are, as Ockhamreadily acknowledges. 18 We have repeatedly insisted on it: the connotative concepts which are needed for a correct understanding of the world - especially the relational ones - cannot all be constructed out of absolute. concepts and syncategorematic connectors. There is no reason, in particular, why the group of things which are ordered in a certain way with respectto some otherthings (those that are taller than something else for example, or similarly coloured, and so on) should always constitute a natural kind, or result from some formal combination of natural kinds. If our categorematic concepts were all absolute, we would have a hard time finding our way in this world! It so happens, fortunately, that we are able to simultaneously grasp several things and (according to Ockhamism) to naturally produce, on this hasis. certain general representations that are projectible to all other n-tuples in the world that are ordered in similar ways. From the epistemological point of view, this presupposes that of at least one absolute concept. When the human mind naturally acquires a simple connotative concept such as 'white' or 'similar', it must apprehend at least two distinct objects as simultaneously present, two whitenesses, for example, or Socrates and his Whiteness, or Socrates and Plato, and so on. And it must apprehend them as distinct from one another. Each one, consequently, will trigger the normal natural process that leads to the formation aran absolute concept. IfI see two whitenesses, as in the example from the passagejust quoted, each one of them suffices to bring about in me the absolute concept of whiteness. The simultaneous grasping of the two of them brings about, in addition, the formation of at least one connotative concept, but this does not preclude thenormal formation of the relevant absolute ones. In many cases, there will even be more than one absolute concept involved. Suppose, for example, that you simultaneously see a black dog and a black horse; what the Ockhamistic theory predicts is that you will then acquire (if you do not already have them) the absolute concepts of 'dog', 'horse' and 'blackness', and, in addition, the simple connotative. concepts of 'black' and 'being similar to' (or 'being similarly coloured'); It follows, in particular, that the possession of a simple connotative concept always implies the possession of a quidditative concept which is such that the connotative concept in question is predicable orit in a true affirmative sentence, just as Ockham stated in a previously quoted passage. 17 Suppose, for example, that I have acquired the simple connotative concept 'black' by meeting with a black dog; then automatically I will have acquired on the same occasion,- if I didn't already have it - the concept 'dog'; this concept is such, in this situation, that the sentence 'a dog is black' is true. The simple connotative concept 'black', then, will be anchored, so to say, through the quidditative contept of 'dog'. Note, however,that the presupposed absolute concepts in such cases are not parts of the connotative concept. The concept 'dog' is not a part of the concept 'black'. The absolute concept 'whiteness' is not a part of the concept 'similar', even for those agents who have acquired the latter by meeting with two whitenesses. Similarly, if I see both Socrates and Plato and form the simple connotative concept 'taller' beyause I intuitively see Plato as taller than Socrates- Iwill also standardly form the absolute concept 'man'; yetthe concept :man' is in no way a part of the connotative concept 'taller'. I could have acquired this same relational concept by meeting with two apes instead, or I could have acquired the concept 'black' by meeting with a black horse instead of a black dog, and so on. The meeting with Plato and Socrates, in one example, or with a black dog, in another, merely serve as occa,\'ion.l' for the formation of simple connotative concepts, just as they do for the relevant ahsolute ones. No part-whole relation is involved here. In the latter examples. the ahsolute

61 The adequacy of nominal definitions How such concepts lend themselves to nominal definitions, even though they are not introduced by them, can now be more fully understood. The main thing, once more, is that the triggering intuitions for these simple connotative concepts are graspings of ordered n-tuples. The Ockhamistic picture of the mind supposes that a particular couple, let's say, is naturally treated as a sample for all other couples that are similarly connected; Plato and his whiteness are taken as a (complex) sample for all couples composed of a body and its whiteness; Socrates and Plato are taken as a sample for all couples composed of a taller individual and a shorter one, and so on. What a nominal definition should do in such cases is to make it explicit, in the required way, to which couples exactly does a concept which is so acquired, naturally extend. More specifically, Ockham's theory is committed to the idea that it should be possible in principle, for each simple connotative concept: 1) 2) 3) (A) Formally, the definition should conform to the structure identified in chapter 5 and (normally) contain at least one term in recto and one (or more) grammatically.subordinate term(s}. (B) to identify the natural group of individuals that all the primary significates of this concept belong to (this corresponds to condition Cl of a good nominal definition, in the analysis put forward in chapter 5); to delineate exactly the connoted group (condition C2); to express, by means of the syntax and the syncategoremata, the connection that holds between the primary significates of the considered connotative concept and the corresponding secondary significates (condition C3). The definition that does that must not necessarily be known - or even recognized as such by every cognitive agent who happens to have the relevant connotative concept. 19 The role of an Ockhamistic definition, remember, is not to reveal what is going on in the mind of the cognizer, but to tell us how the world should be for a given concept to apply. What is required by the theory is merely that it be possible to construct such definitions. The real question that arises, then, is how the adequacy of the definition can be assessed in each particular case. If t\ie cognitive agent cannot be expected to spontaneously give us the nominal definitions of his own simple connotative concepts, how are we to know that we get them right? This is not something Ockham has elaborated upon, but a number of criteria do emerge from what we have established so far. First, there are some general conditions to be considered: The nominal definition should be compatible with a sound ontology. In order to play its role of clarifying the ontological import of the defined term, all of its categorematic terms, whether in recto or in obliquo, should clearly apply to entities which are acknowledged as such by a good ontological theory. In the context of Ockham's doctrine, this means, of course, that they should clearly apply to individual entities. (C) (D) COGNITION AND CONNOTATION The nominal definition should be compatible with a sound epistemology. In the case of simple connotative concepts, in particular, this means that the connection which the definition sets down between the primary significates of its term in recto and that of its term(s) in obliquo, should be ascertainable on the basis of direct experience, at least in some cases. Otherwise, the abstraction process could never leave the ground! Taken together, these three constraints make it clear that the adequacy ofa proposed definition for a naturally acquired concept cannot be established by simply confronting it with the defined term, or with what the cognizing agents spontaneously say about it. Given what the definition is for, it is bound to be heavily theory-laden, especially with respect to semantics, ontology, and epistemology. Yet it is plain that the three criteria formulated so far do not suffice for establishing in any particular case the adequacy of some unique definition. A different sort of condition is still needed: The nominal definition should generally agree with how the defined concept is used in fact. Since concepts are not available for public observation, it has to be supposed for this test to be conducted, that a simple concept is usually linked, for a particular agent, with a given spoken word or linguistic phrase. How the agents apply the word in concrete situations and what inferences they make with it will count as evidence for how they use the concept. This indeed seems to be what Ockham does in practice. What is required by this condition, however, can on~y be general agreement. A subject can be mistaken in the application of anyone of her own concepts in some particular occasions - or even in many occasions - and in the inferences she makes with it. Condition D has to be used in constant conjunction with the other three. Ockham, 'as I believe, is committed to all of this, but none of it is very explicit in his works. So I will not pursue the matter any further at the theoretical level. Let me simply illustratehow it can work with a couple of examples. Take, first, the relational concept 'taller'. It is not one of Ockham's own examples, but it is quite revealing, for 'taller' is, presumably, one of those relational connotative conceptsthat can be acquiredon the basis ofdirectexperience according to the doctrine sketched in the previous section. In Ockham's perspective, 'taller' primarily signifies everything which is taller than something, and secondarily signifies - or connotes~ everything with respect to which something is taller. In so far as it is an asymmetrical relational term, it must have a correlate, 'shorter' let's say (or something like it); since correlative terms interdefine each other according to Ockham,20 'shorter' should be one of the key categorematic terms in the complete nominal definition of 'taller'. The definition of 'taller', consequently, would be something like 'a body with respect to which there is something shorter'. Let us see how this formulation faces criteria (A)-(D). For one thing, it does agree with criterion (A) in having the general form Ockham expects from a nominal definition: it includes a categorematic term in recto - 'body' - which applies to all the primary significates of the defined term, and another categorematic term in a grammatically suhordinate position - 'Isomethingl shorter' 111

62 112 COGNITION AND CONNOTATION which applies exactly to the connotata of the defined term; the other components of the definition being syncategoremata.~l With regard to criterion (B), the proposed definition for 'taller' clearly agrees with. what Ockham takes to be a sound ontology: the primary significates of its two categorematic terms are singular concrete things. Bodies certainly are at any rate. As to shorter things, one can tum, if in doubt, to the definition of 'shorter', which, presumably, would be something like: 'a body with respect to which there is something taller'. Despite the striking circularity here, the recourse to this latter definition has the significant advantage of making it dear that the primary significates of 'shorter' - and, consequently, the connotata of 'taller' - are also singular bodies. A complete Ockhamistic nominal definition, remember, is not intended as a reductive or a learning device, but as a way of making conspicuous the ontological import of the defined term. Thatthiscan best be done in this case with the help of two correlative definitions.is not a problem for Ockham.2~ What the example nicely reveals is that when two connotative terms are interdefinable in the way that ~taller' and 'shorter' are taken to be, it is the terms in recto in the two definitions that ultimately indicate the sort of things. that must be signified and connoted by the two terms (since the connotata of each one of them should be the primary significates of the other, and the primary significates of each should be among the significates of the term taken in recto within its own nominal definition).~3 These terms in recto, moreover, should normally be absolute terms,24 and belong, consequently, to the categories ofsubstance or quality. Criterion (C) also applies to the example, in so far as the case is in accordance with Ockham's epistemology. What is supposed, in particular, is that there can be in normal human experience a simultaneous intuitive grasping of two things which is such that one of them is seen as taller than the other one, and the latter as shorter than theformer. It is obv:iously the kind of case Ockham has in mind when he speaks of relational concepts that are 'caused by both extremes, posited simultaneously, prior to composition and division'. 25 In such a situation, the two correlative concepts, as he says, 'simultaneously enter the intellect'26 and can legitimately interdefine each other. The proposed definition, finally, can be tested against criterion (D) if we suppose that the defined concept 'taller' is usually associated 'with a certain word, for example, the word 'taller' for speakers ofenglish. We can check that the speakers normally apply the relevant word in concrete situations to something which is such that &omething else appears to them as shorter than it, and that they are prone to infer from 'xis taller than y' that y is shorter than x, and that x and yare bodies. Another sort of example is provided by the standard case of 'white', with its (by now familiar) Ockhamistic definition: 'a body. haying a whiteness'. What is special here with respect to the previous example is that the defined connotative concept is non-relational, and that it cannot be defined, consequently, by means of a correlate (since it doesn't have any). The adequaoy of the proposed definition, nevertheless, can be assessed in the light of criteria (A)--'-(D) just as in the previous case. The definition, obviously, has the required form (criterion A). It agrees with Ockham's nominalistic ontology in requiring only singular bodies as primary significaies and singular whitenesses as connotata (criterion B). It also agrees with Ockham's epistemology, since the sort of couples that arc descrihed in the definition (hodies with their whitenesses) can intuitively be grasped as such according to him,27 and this would typically be the kind of experience that can bring about the acquisition of a concept such as 'white' (criterion C). The definition, finally, can be tested against the intuitions of the speakers, through the observation of how they apply such words as 'white', 'blanc', or 'albus' in concrete situations, and what inferences they are prone to make with them (criterion D). The simple connotative concepts acquired on the basis of direct experience are not identified, within Ockhamism, with their nominal definitions; they are not introduced by way of them, and they do not have to be synonymous with them. Yet, because of their naturally hierarchized semantical structure, they do have such definitions, and these can, in general,.be made explicit with the' help of a good semantical, ontological, and epistemological framework. What has to be assumed for this - as Ockham certainly does -is that the possession of a given simple concept is usually linked somehow with certain dispositions to apply it in concrete situations, and to use it within inferences. This does not mean that the simple concept itself is reduced to a recognitional capacity or to an inferential role. Given its most important function, which is to contribute in determinate ways to propositional acts, the specific identity of any given concept, whether connotative or absolute, is partly determined by the way its (primary and secondary) significates are fixed. In the case of simple connotative concepts of the type we have discussed so far, as in that of simple absolute concepts, this is done by the natural working of the abstractive process on the basis of experiential' intuitive graspings; The distinctive feature of simple connotative concepts, from this point of view, is that they are specifically triggered by the grasping of ordered n-tuples. This is why, ultimately, anyone of thern can in principle be defined along the lines set down in chapter Ockham and the Classical View It is clear, however, that not all connotative terms can be acquired on the sole basis of intuition. Many of them, surely, correspond to connections between things that cannot be ascertained in direct experience, even when the relevant objects are all physically present to the cognizing subject. Such is the case, saliently, for one of Ockham's favourite examples of connotative term: 'father'. Even ifi simultaneously grasp President George W. Bush and the former President George Bush, let's say, this cannot be enough for me to acquire a simple concept of 'father' if Ido not already have one, since this simultaneous grasping of the two men is not sufficient for me to evidently know that one of themis the father of the other. If most terms in our external languages are connotative, as comes out from Ockham's treatment of the Aristotelian categories, then this situation must hold for a great many of them. In other words, the acquisition process described in section 2 can account, in the context of Ockhamism, for the introduction of only a portion of all connotative terms. This is not to say that it is not important. We have seen, on the contrary, that such a mechanism is indispensable, in Ockham's framework, for the development of appropriate mental representations of the world as it is ordered. But the question stili arises: how are the other connotative terms acquired? And how are they represented in the mind'! Must they nol, in particular. he introduced hy complex arrangements of

63 114 COGNITION AND CONNOTATION 115 more primitive concepts? And if so, aren't we led back to an interpretation of Ockham's theory of concepts as a variant of the so-called Classical View, according to which the vast majority of our concepts ate in fact complex definitions? What we have to find out, to answer this, is what other modes of introduction there can be for connotative terms within Ockhamism, besides the intuitive grasping of n-tuples. As far as I can see, at least two different - but noncexclusive - such procedures can be accepted by Ockham: abbreviation and definition. As to abbreviation, Ockham admits that the speakers of a language can always, if they wish, 'use one locution in place of several'.28 The result of this procedure will not in every case be a connotative term, but it can be ifthe abbreviated phrase is originally structured in such a way as to be capable of occurring in personal supposition as the subject or predicate of a well-formed sentence. If a complex phrase such as 'man necessarily' is abbreviated into a single word (as Ockham admits it can be),29 the resulting abbreviation will not be a true connotative name since the original complex phrase is not capable by itselfofbeing subject or predicate ofa well-formed sentence; such expressions, as Ockham says, 'do not have precisely the value of a name'.30 A phrase such as 'black horse', on the other hand, can occur as subject or predicate; if we should abbreviate it into a single word, the resulting neologism -'blorse', let's say - would be a perfectly good connotative term. When Ockham speaks of this abbreviative device in Summa Logicae I, 8, he presents it as holding between linguistic units: a complex spoken or written phrase, he submits, can always be abbreviated into a single word. But there is no reason why it could not hold between a complex mental phrase (for example, the complex mental expression corresponding to the English words 'black horse') and a single spoken or written word. In such cases, the external connotative term would be represented inside the mind in the guise of a complex concept, and it would be strictly synonymous with this concept. Nominal definitions, on the other hand, can also be accepted within Ockhamism as a device for introducing connotative terms. Nothing prevents us from constructing complex expression having the required Ockhamistic form for a good nominal definition and, then,. to conventionally strike a word of which this complex expression would be the definition. Such cases would importantly differ from the previous ones in that the introducedterm wouldnot necessarily be synonymous with the complex phrase that is used to introduce it, since, as we repeatedly insisted, an Ockhamistic nominal definition is not synonymous, usually, with the term it defines. The definition indicates what the primary and secondary significates of the newly introduced term are, but some of its components can signify or connote certain things that are neither signified nor connoted by the defined term. 'Father' is a case in point. Since it cannot be acquired on the basis of direct intuition, and since it has, according to Ockham, a good nominal definition ('[male] animal having a child'), we may conjecture that it can be introduced in a language by way of this definition. If so, it will not signify every single thing that is signified by at least one component of the definition (it will not signify all animals, for example), and it will not, consequently, be synonymous with the definition in Ockham's strong sense of syrtonymy. It will not be said, in particular, to be subordinated to its mental definition, in Ockham's technical sense of 'subordination'.31 Yet it would have a determinate meaning in the external language, while being represented in the mind by a non-synonymous complex phrase. The outcome is that whether they are introduced by abbreviations or by definitions, simple spoken or written connotative words of the sort we are now discussing should be represented in the mind in the guise of complex conceptual combinations. Although a bit more complicated, this comes very close to what the so-called Classical View is taken to hold, except that when nominal definitions are involved, strict synonymy is not required between the simple external word and the corresponding mental sequence. Despite this non-neglectible difference, it still will be supposed in the Ockhamistic framework that a large part of the simple categorematic terms of our external languages receive their meanings from complex constructions out of more primitive simple concepts. We are not brought back to Spade's picture of Ockhamism for all that. The main problem with this picture was that the only categorematic concepts it accepted as primitives in the Ockhamistic constructional processes were the absolute ones. It would have followed that the meanings of all relational terms should be constructible on the sale basis of absolute terms (plus syncategoremata). Which was doomed to failure. What we arrive at now is entirely different. Among the simple concepts naturally implemented in the human mind on the direct basis ofexperience, there are also some connotative terms for Ockham, relational ones especially. And this allows for a much more powerful natural basis than Spade and his followers thought, with a much better prospect for Ockham's nominalistic program thus reinterpreted. In a recent paper, one of Spade's pupils, Yiwei Zheng, has expressed the worry that as a result of my acceptance of simple connotative concepts in Ockham's epistemological basis, I'd be led to abandon 'the idea that Ockham's mental language has a recursive semantics', an outcome which Zheng sees as 'clearly undesirable'.32 But why should I be so led? What a recursive semantics requires is that there be both primitive and derivative terms in the considered language and that the semantical properties.of the derivative ones be obtained from that ofthe primitive by means of a limited number of (indefinitely) repeatable constructional rules. This is indeed what we arrive at. The difference with Spade's original proposal in this respect is that Ockham's primitive basis, as I see it, is extended to include a number of connotative terms (those that can be acquired on the basis of intuitive graspings) in addition to the absolute ones. Even though they are definable, these basic connotative terms are not reducible to more primitive concepts, since their definitions can circularly involve certain terms which they themselves serve to define. But many other connotative terms are still taken to be engendered through constructional devices (abbreviations or nominal definitions) which can be governed by recursive rules. Just how many simple conrtotative concepts can there be, or what their proportion should be, is left open in Ockham's own texts. All we can say is that there should be at least one for each sort. of connection between individual things which is independently ascertainable ort the basis of direct intuition; a finite quantity presumably. The other ones are to be constructed somehow, and, as far as we know, an infinity of them might very well be possible... The important point for Ockham's programme is that the admittance of simple connotative concepts into the basis of the epistemological system does not complexit'y the ontology, not more at any rate than what is acceptable for him. Simple connotative terms have primary and secondary significates, but given the way these terms are acquired. all of these arc singular things. Spade and many other

64 116 COGNITION AND CONNOTATION 117 commentators have been presupposing that the ontological elimination of special entities from the categories other than substance and quality was only possible, in the context of Ockhamism, if all the terms in these categories '- and especially all the relational ones - could be reduced through constructional devices such as definitions, to absolute terms from the categories of substance and quality.33 But this is not how Ockham's nominalism works. The ontological import of a given term, for the Venerabilis Inceptor, depends on what entities it is supposed to signify, whether primarily or secondarily Irreducible connotative concepts, therefore, can be admitted if it can be shown ~ through their nominal definitions - that they signify or connote nothing but singular substances or qualities. Notes 1. Spade 1996, pp See above chap. 2, sect See also Zheng 1998, whose discussion of Ockham's theory of connotative concepts equally builds upon this distinction between metaphysicaland semantic simplicity. 4. See above chap. 5, sect Ord. I, dist. 2, q. 9, OTh II, pp : 'Verbi gratia, habeo istum conceptum, puta esse creativum, quem scio esse denominativum, et ideo oportet praehabere unum conceptum cui istum attribuo, puta dicendo quod aliquid ens est creativum; et certum est quod iste conceptus cui iste attribuitur non est denominativus, vel si sic, erit processus in infinitum vel stabiturad aliquem conceptum quidditativum' (italics are mine). 6. See Ord. I, dist. 2, q. 8, OTh II, pp Normore 1990, p. 59, finds the doctrine expounded there simply'astonishing'. See also Adams 1987, p Ord.. I, dist. 2, q. 8, OTh II, p. 282: 'Quartum dubium est de conceptibus syncategorematicis et connotativis et negativis: unde possunt sumi vel abstrahi?' 8. Ibid., pp : 'Possunt autem tales.conceptus imponi vel conceptus abstrahi a vocibus... Tunc ab istis. vocibus sic significantibus abstrahit intellectus conceptus communes praedieabiles de eiset iinponit istos 'conceptus ad significandum illa eadem quae significant ipsae voces extra. Et eodem modo et de talibus format propositiones consimiles et habentes consimiles proprietates quales habent propositiones prolatae.' For a detailed analysis of this approach in so far as it applies to syncategorematic concepts, see below chap. 7, sect Ibid., p. 285: 'Ad quartum dubium dicerent [or dico, according to ms. A and F] quod conceptus syncategorematici et connotativi et negativi non sunt conceptus abstracti a rebus ex sua natura supponentes pro rebus... Et ideo dicerent [or dico, according to ms. A, B, C and F] quod nullus [est] conceptus syncategorematicus nec connotativus nec negativus, - nisi tantum ex institutione... ' 10. See Exp. sup.. Elench. II, 16, OPh III, pp : 'Large omne nomen vocatur connotativum in cuius definitione ponitur aliquid in recto et aliquid in obliquo, vel verbum vel alia pars orationis... Isto modo accipiendo hunc terminum "nomen connotativum", est in plus quam "nomen relativum"; quia omne nomen relativum est nomen connotativum... Aliter accipitur nomen connotativum stricte... Omnia autem nomina connotativa, large accipiendo connotativum, praeter ista quae stricte vocantur connotativa, sunt nomina relativa' (italics are mine). I deem it probable that Ockham uses the narrow sense in the passage from the Ordinatio we are currently considering, because it is the simplest interpretation in the conlexl. Except in Ockham's own logical works, where he remodelled the notion of a connotative lerm and gave it both a wider extension and a greater weight, it was usual al Ihe time 10 distinguish l'onnolalive from relative terms (see, for example, the Pseudo-Carnpsall, Logica 9, ed. Synan 1982, p. 102:,... differencia est inter connotativa et relativa...)'. It is relevant to note that this particular passage of the Ordinatio was written early in Ockham's career and that the technical word 'connotativum' is introduced in it via an objection. In Exp. sup. Elench. II, 16, OPh III, p. 302, Ockham makes it explicit that relational terms occur in the nominal definitions of some non-relational connotative terms: 'Ex quo patet distinctio inter nomina relativaet connotativa. Sequitur autem ex isto quod non omne nomen in cuius definitione exprimente quid nominis ponitur nomen relativum est nomen relativum.' This conventionalist theory of connotative and syncategorematic concepts is allusively referred to again in a few other passages. See in particular Ord. I, dist. 3, q. 2, OTh II, p. 403: 'Tertio, dico quod [Deus] est cognoscibilis a nobis in conceptu aliquo modo simplici ad placitum instituto ad significandum, et hoc in conceptu connotativo et negativo sibi proprio'; and - more surprisingly, since this is supposed to be a later text, but also more restrictively, since it bears only upon mental syncategorematic terms - in Quodl IV, q. 35, OTh IX, p. 471: '... signa mentalia ad p1acitum significantia..., puta syncategoremata mentalia'. See the summary of the process given in SL III-2, 29, OPh I, p. 557, with, in this case, the example of a man as the known object: '... 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... Deinde apprehenso alio animali ab homine vel ahis animalibus, elicitur una notitia generalis omni animali...' Ord. I, Prol., q. 1, OTh I, p. 31: 'Similiter, notitia intuitiva est talis quod quando aliquae res cognoscuntur quarum una inhaeret alteri vel una distat loco ab altera vel alio modo se habet ad alteram, statim virtute illius notitiae incomplexae illarum rerum scitur si res inhaeret vel non inhaeret, si distat vel non distat, et sic de aliis veritatibus contingentibus '" Sicut si Sortes in rei veritate sit albus, illa notitia Sortis et a1bedinis virtute cuius potest evidenter cognosci quod Sortes est albus, dicitur notitia intuitiva' (italics are mine). Quodl. IV, q. 17, OTh IX, p. 386: '... nam conceptus relativus causatur ab utroque extremo simul positis ante compositionem et divisionem... Unde iste est ordo, quod visis duabus albedinibus, primo causatur in intellectu conceptus specificus albedinis; secundo, naturaliter mediante ilio conceptu specifico causatur conceptus similitudinis, et hoc dico immediate ab ipsis albedinibus vel a cognitionibus earum; et post saltern ordine naturae formatur propositio' (transl. Freddoso and Kelley 1991, pp , slightly modified). In the particular example under consideration, the relevant connotative concept - that of 'being similar' - happens to be symmetrical, so that, presumably, all of its primary significates are also found among its connotata (everything which is similar to something is such, that something is similar to it), but this cannot be seen as a necessary feature of all simple relational concepts. See the text from Ord. I, dist. 2, q. 9, quoted above in n. 5. See above chap. 5\ sect See above chap. 5, sect See SL III-3, 26, OPh I, p. 690: 'Et tali definitione relativa mutuo se definiunt.' See above chap. 4, nn , and chap. 5, sect. I, Thesis 1. 'In respect to which there is' can be treated in the proposed formulation as a single syncategorematic asymmetrical preposition or copula, just like 'having' in Ockham's detinitions of 'white' ('a body having a whiteness') and 'father' ('an animal having a child'). This point, of course, requires further elaboration, but, as I said earlier, Ockham docs not provide any detailed theory for such syncategorematic connectors. More on this in chap. X, sect..1 hclow.

65 Cf. SL 111-3,26, OPh I, p. 690: '... non est inconveniens si mutuo definiant se'. 23. Thisfollows from conditions Cl and C2 in chap. 5, sect. 2 above. 24. See above chap. 5, n. 36. In the examples presently under consideration, the relevant absolute term in recto is 'body' in both the definition of 'taller' and that of 'shorter'. It is true that Ockham sometimes lists 'body' (corpus) as a connotative term (for example, in SL I, 10, OPh I, p. 37, SL I, 44, OPh I, p. 139, SL I, 45, OPh I, p. 142), but he uses it elsewhere as an absolute term (for example, in SL 111-2,14, OPh I, p. 530, where 'corpus' is attributed a real definition - 'res composita ex materia et forma' - typically suitable for an absolute term). The word, therefore, has two different senses in his writings. When it occurs in recto within a nominal definition (such as that of 'albus'), it should be taken in the second of these. 25. Quodl. IV, q. 17, OTh IX, p. 386; the text is quoted above in n SL 111-3, 26, OPh I, p. 690: '... relativa sunt simul in intellectu'. 27. See the text from Ord. I, Prol., q. 1, quoted above in n SL 1,8, OPh I, p See above chap. 5, sect Ibid. 30. Tract. de Quant., q. 1, OTh X, p. 24. See above chap. 5, nn A word is subordinated to a certain concept in Ockham's vocabulary when it is imposed to signify the very things that are signified by this concept, and to signify them in exactly the same way as the concept does. See SL I, 1. More on this in chap. 9 below. 32. Zheng 1998, p ; See, for example, Spade 1998, p Chapter 7 Concepts as Similitudes Faithful to the Aristotelian tradition, Ockham holds that a concept is a likeness of whatever it represents, a similitude.! Many commentators have seen this as a problem: what can it mean in the context of Ockhamism? Is it a mere far,;on de parler? Or even an incdnsistency on William's part? Or is there some philosophical interest, within his nominalism, to this notion of the concept resembling its significata in some sense? Marilyn Adams, for one, devoted an important development to the question in her landmark book of 1987: 2 'I do not see', she concluded, 'how Ockham can specify, either on the objective-existence theory or on the mental-act theory, a similarity relation that can constitute the natural signification relation for general concepts such as "animal" and "man".'3 A couple of years later, Pierre Alferi went further and claimed that strictly speaking, Ockham is inconsistent in accepting the vocabulary of similitude within his mature theory ofthe concept, and that he yields in so doing to an unrigorous way ofspeaking, a mere remnant of an old doctrine that he had himself rendered obsolete. 4 Joel Biard even suggested in 1989 that Ockham, in fact, abandoned the vocabulary ofresemblance when he turned to the actus-theory.5 While admitting in a later work that this was somewhat of an exaggeration, he still strongly stressed the prominence ofcausality over similitude in Ockham's final theory of natural signification. 6 Both Cyrille Michon and I have discussed the point in some detail in the 1990s, and conc1udedthat similitude was always considered indispensable in Ockham's epistemology;7 but we still underestimated, I am afraid, the interest and importance of the idea in Ockham's own eyes. I would like to revisit the whole issue in the present chapter. As in most of this book, 1 will concentrate on Ockham's mature theory, where concepts are identified with mental acts. We will see first that the notion of conceptual similitude is still importantly present in this phase of his thought, with significant philosophical roles to play (section 1). In the absence of explanations from Ockham himself about what this means exactly, I will propose a tentative interpretation (section 2), and apply it to the main categories of general concepts countenanced by the actus-theory (section 3). This will put us in a position, finally, to discuss in some details two intriguing puzzles recently raised about Ockham's theory of absolute concepts (section 4). 1. Similitude sustained That the concept is a likeness of the thing represented is an integral part of Ockham's former theory. How can a purely ideal object such as a jictum represent anything within a given mind? The answer was that it is a kind of intellectual picture: I III

66 120 CONCEPTS AS SIMILITUDES seeing something outside the soul, the intellect fonns something similar within the mind, in such a way that if it had a productive capacity, just as it has a representative capacity, it would produce such a thing outside itself in real being, which would be numerically distinct from the prior one... And this [= thefictum in the mind] can be called a universal, for it is a portrait and it indifferently refers to all the [relevant] external individuals; and it is in virtue ofthis similitude in objective being that it can supposit for external things...8 Elizabeth Karger, in commenting on this passage, suggestively says that Ockham's fictum is 'a sort of purely ideal blueprint of the thing'.9 It has what could be called an intentional resemblance with what it represents, and. this resemblance accounts for the semantical features of the concept, its capacity for supposition in particular. This seems unproblematic within the fictum-theory because the concept, in this framework, is supposed to have no other mode of existence, precisely, than that of an intelligible portrait. When Ockham abandoned the fictum- in favour of the actustheory, however, he did not renounce the idea that the concept is a similitude. Several texts of the mature period testify to it. In Quodlibet IV, 35, saliently, the Venerabilis Inceptor, having criticized the fictum-theory, reaches the following conclusion: Therefore, I claim that both a first intention and a second intention are in reality acts of understanding, since whatever is preserved by appeal to a fictive entity can be preserved by appeal to an act ofunderstanding. For like a fictive entity, an act ofunderstanding, (i) is a likeness ofan object, (ii) is able to signify and supposit for things outside the soul, (iii) is able to be the subject or the predicate in a proposition, (iv) is able to be a genusor a species, etc. to The argument is straightforward: all the representational and semiotical functions we might want to attribute to a concept can be fulfilled by the act of intellection as well as by the fictum; since the act is indispensable anyway, as everybody admits, the Razor Principle requires the elimination of the fictum. 11 Now, among the conceptual functions, the first one Ockham mentions is similitude: an act, he says in so many words, can be a similitude of the object represented.. This is not a slip or an isolated statement. One can turn, for example, to another passage where Ockharn elaborately discusses the fictum- and the actus-theory and clearly decides in favour of the latter: the Questions on Aristotle's Physics. Questions 1-7 of this work are dedicated to the ontological status of the concept. Ockham, there too, explicitly subscribes to the idea that 'the concept is a similitude of the external thing'.12 As Cyrille Michon has remarked, he even turns the point into an argument in favour of the actus-theory.13 Of the seven arguments against the.fictum listed in question 1, the seventh - and longest - is the following: Seventhly, such an ideal picture [idolum] would differ more from a thing, than whatever thing from another one, since a real being and a rational being differ more from one another than any two real beings; this is why such a picture would less assimilate to a thing, and consequently be less able to. supposit for a thing, than the intellection [or intellectual act] which assimilates more to a thing; and the ideal picture will less be common to external things, and it will less have the status of a universal than the intellection. But such a picture is called for for no reason except for suppositing for a lhing. or to be a component of a proposition, or to be common to external things etc., and all of this suits the intellection better than the ideal picture. This ideal picture or fictum, therefore, is superfluously posited.1 4 The possibility for a general concept to stand for real external things within a proposition is here said, once more, to hang upon the resemblance of the concept with the things in question. But an act,.ockham says, can be more similar to any real thing than afictum can, since it is itself a real thing. It is true, as Michon stresses, that Ockham had put forward a reply to this argument in a previous discussion of the fictum- versus the actus-theory.'5 This was in the Prologue to his commentary on Aristotle's Perihermeneias, a text in which he still seemed to hesitate between the two approaches. His rejoinder, then, was that a real resemblance is not needed in order to account for the conceptual functions; a resemblance 'in intentional being' is what is relevant for the concept to be able to supposit for a thing, or to be common, etc. 16 Yet, the conclusion he reaches in the Questions on the Physics, is that a real resemblance is even.better suited. And the whole discussion, anyway, presupposes on both sides that thelikeness of the concept with the thing - whether it is a real or a merely intentional similitude - is the condition of possibility for the referential functions of general concepts to be fulfilled. In the Quodlibetal Questions, moreover, the idea that the concept is a similitude is the basis for another important philosophical argument. Ockham uses it as the main premise to show that there are no simple singular concepts in the human mind. Since the representational function of the concept depends upon its resemblance with certain real things, he argues, any given concept must equally represent all the individuals which are maximally similar to each other, and it cannot, consequently, singularize any of them in particular. See the following extract: Fourthly, I claim that our intellect cannot have any such proper and simple concept with respect to any creature, either with or without a vision of the creature. And this is because each such cognition or concept is equally a likeness of, and equally represents, all exactly similar individuals, and so it is not more a proper concept of the one than of the other. 17 The same argument is found in another passage of the Quodlibetal Questions, which is even more interesting for our present purpose: Third, I claim that the first (by a primacy ofgeneration) simple abstractive cognition is not a cognition proper to the singular but is sometimes, indeed always, a common cognition... [This assumption] is evident from the fact that no simple abstractive cognition is more a likeness of one singular thing than of another exactly like it; nor is it caused by, or apt by nature to be caused by, [just one] thing. Therefore, no such cognition is proper to the singular thing; rather, each such cognition is universap8 The 'first abstractive cognition', in Ockham's vocabulary, is the one that is generated in the mind on the heels of the singular intuitive cognition. I see a nuthatch for the first time, let's say. Once the intellectual intuitive cognition occurs, a concept is immediately generated, which abstracts from the existence of this concrete small bird in front of me. What Ockham is now telling us is that this first abstractive cognition is already 1(enem/. It is not a singular representation of this particular

67 122 nuthatch, but a concept of everything - even every possible thing - which is maximally similar to this nuthatch, in short a general concept of nuthatch. The reason given for this is that an abstractive cognition represents something in virtue of resembling it. What accounts for the representational import of a concept is similitude, not causality. As we now see, this is a well-thought out thesis in Ockham, even in the context of his later actus-theory. 2. Acts and similarities CONCEPTS AS SIMILITUDES act - which is a quality of the mind in Ockham's ontology - could not be said to be similar to anything. Nor could the mind itselfbe said to be qualitatively similarto the represented thing, since it does not in general acquire through representation a quality of a sort which is exemplified by the thing represented: the soul, after all, does not become red when it forms a conceptual representation ofa red thing! A second sense of similarity that should be excluded from the present discussion is the one Ockham presupposes when he speaks of two things being 'maximally similar' to each other: simillimae. As I have explained elsewhere,24 the superlative 123 simillimus amounts to a technical term in Ockham's vocabulary: it usually applies to things which belong to the same species specialissima. 25 Two men, for example, will be said to be maximally similar to each other,26 or it could be two nuthatches, or two horses. Or even two whitenesses: maximal similarity in this technical sense is not restricted to substances. And it is not a degree of qualitative - or accidental similarity in the Aristotelian sense identified above. What it presupposes is a general notion of essential similarity, understood as a graded relation. A horse and a dog, for example, have a non-maximal essential similarity to each other, because, although they do not belong to the same species specialissima, they both are animals; this does not depend upon any quality that they might have, but directly upon what they intrinsically are, their (singular) essences. A whiteness and a redness, in this sense, also have a noncmaximal essential similarity to each other, for they both intrinsically are colours. Whether maximal or not, essential similarity is crucial to Ockham's ontology and epistemology. Yet it is obviously not the relation we are looking for between a concept and whatever it represents, since a concept does not represent something in virtue of belonging to the same species or the same genus as the thing. Not only is the concept of horse not a horse, but it is not a mammal either, or an animal, and not even a substance for that matter. There is a third sort of resemblance that must imperatively be put aside at this stage, although some commentators seem to have it in mind when discussing these issues. It.is the idea of a perceptible image that one could look at, like a colour photograph or a realistic picture. This model for understanding similitude, I suspect, plays a non-neglectible role in the difficulty many scholars encounter with the idea that a mental act could have been accepted as a likeness of the represented object. How, after all, could an intellectual act be anything like a photograph? For one thing, the intellectual act, for Ockham, does not need to be looked at by the mind in order. to represent; it does not display, moreover, any perceptual features, like photographs or statues do. What we should conclude, however, is not that Ockham's idea of conceptual similitude is muddled, but that this is not the right model for it. Other possibilities are to be explored. Let us return to Biard's incisive question: 'what does a resemblance between a thing and an act mean?' Above all, we should not be misled by the term 'act' in this formulation. As was explained in chapter 2, a mental act for Ockham is not an action in the modern sense, or a gesture, or a movement. It is more like an actualized state of the mind, a posture it takes. It does not, in particular, have to be dynamic. Consider, for a comparison, the manual grasping of an object. This is an act, in Ockham's sense. It is not identical with the hand itself and can be classified as a quality of the hand just as the intellectual act is classified as a quality of the mind. The grasping. however. is not necessarily a gesture or an action. It can simply How is this to be understood? The question has been sharply raised by Joel Biard in particular. In the move from the fictum- to the actus-theory, Biard contends, 'the idea of similitude loses some of its sense. For we might admit a resemblance between a thing and a concept endowed with an esse obiectivum, but what does a resemblance between a thing and an act mean?' 19 Objections, indeed, had already been raised in Ockham's own times. The Dominican Crathorn, a harsh critique of Ockham's theory of mental language, was ironic on the point as soon as the early 1330s in his commentary on the Sentences. It is impossible, he complained, that a quality in the mind should simultaneously resemble everything there is, as the concept of 'being' (ens) would be supposed to in Ockham's doctrine.2o And how could the general concept of 'quality' (qualitas) be itself a mental quality naturally similar to every quality in the world? How could th~ concept of 'colour' be a quality naturally resembling whiteness, redness and green all at once?21 How, we might add, could the mental act correspondingto the concept of 'animal' equally resemble a butterfly and an elephant, while having with both a greater resemblance than with an orchid or a hammer? These are disturbing questions. Yet what they primarily suggest, I submit, is that our understanding of Ockham's own idea of similitude is still inadequate. Since he is not in the habit of writing carelessly on topics that matter to him, and since he obviously takes with great seriousness the role of similitude in conceptual knowledge - it is, after all, supposed to account for generality - the task that is incumbent upon us at this point is to try to figure out what he means by'similitudo' in such contexts, even though he does not bother to explain it much himself. A number of familiar notions ofresemblance must be discarded as irrelevant at the onset. One sense which does notapply, for instance, is the standard Aristotelian one, according to which a thing A is similar to a thing B ifa andb sh:;tre a common quality, or if, at least, they both have a quality of the same sort. This shirt and the horse Bucephalus, for example, resemble each other in that they both have a quality which is a whiteness. This is Aristotle's sense in Categories 8, lla 15-19: 'strictly speaking', Ockham says in commenting on thispassage, 'nothing is said to be similar or dissimilar to something else except according to a quality'.22 This acceptation was common in the Middle Ages, and Ockham himself frequently uses it;23 yet it is much too restrictive for our present concern. Only substances can be said to be similar to each other in this sense, for only substances have qualities. A whiteness, for example, would not be similar to another whiteness, according to this restrictive use, since whitenesses, being themselves qualities, do not have further qualities with respect to which they could be said to be similar in this strict sense. A fortiori, a mental

68 124 CONCEPTS AS SIMILITUDES 125 correspond to a static position of the hand, which makes it easier to understand how it can be a likeness of the thing grasped. Even a dynamic gesture, actually, can mimic an object in certain cases. I can sketch the form of a human body, or of a house, with a movement of my hands. A fortiori, nothing prevents a static actualized state of my hand to be seen as a likeness ofsome external thing. Suppose I grasp a ball, ora pen, and let the grasped object be removed without any change in the position of my hand. What is left is a likeness of the ball or of the pen. In Ockham's vocabulary, the remaining act of my hand- its actualized state or position - is a similitude of the object it previously grasped. It is, admittedly, a rough similitude in this case, but a similitude nevertheless. This manual act, moreover, is to the same degree a similitude of any other object which is sufficiently similar to the original ball or pen. In so far as it is a similitude, this posture of my hand pertains to a multiplicity of possible objects. If it could play the role of a sign within a proposition, it would be a general sign, able to supposit for all the objects that have the relevant shape. This comparison, I believe, brings us very close to Ockham's point about the similitude of the concept with the represented thing. What we have in the case of the grasping hand is a physical analog of the distinction between the intuitive and the abstractive acts. As long as the ball is there, the act of my hand can be seen as the singular grasping of this singular ball that causally shapes the posture ofthe hand. This is the equivalent ofan intuitive grasping. Yet I can also consider the position of the hand in abstraction from the actual presence ofthe ball (whetheritis stillthere ornot), andthen I have an analog ofthe abstractive act: a general representation of any object that would tightly fit into the hand when it is so placed. This could legitimately be described as a manual concept! If my hand was then productive, it would engender a ball similar in size and shape with the original one, which is how Ockham originally characterized intentional similarity.27 In short, to say, in Ockham's vocabulary, that an act resembles a thing raises no special difficulty. The act of apprehension by which the mind grasps something must be a similitude of this particular thing and of all the other ones that are sufficiently similar to it, exactly like the act of apprehension by which my hand grasps a ball must be a similitude of this particular ball and of all the other objects that are relevantly like it. We can see why similitude in such cases is the condition of possibility for generality. In the direct grasping corresponding to the intuitive act, what is relevant is the actual causal interaction with a given object. But causality by itself is a link between determinate singular objects or events. What transforms the grasping act into a general representation is that it fits the shape of the grasped object, thus resembling it to some extent; in so far as it is a likeness of the singular object which is grasped, it is also a likeness of any other object which is relevantly like it. While causality is strictly singular, generality naturally comes along with likeness. The required notion of similitude, however, is quite distinctive. It is neither qualitative resemblance in the Aristotelian sense (as between two red things, or two round things), nor essential similitude (as between two nuthatches, or between a dog and a horse), nor perceptual likeness (as between a photograph or a statue and the pictured object). The relevant Ockhamistic connection typically follows upon a causal link and thus differs from both qualitative resemblance and essential similitude by being asymmetrical: a foot track, in this sense. is a similitude of the foot, but not conversely. And it differs from the photographical representation by being a wider and looser relation, more akin to what John Buridan later called a similitude 'by fitness' (secundum convenientiam), which he described as an appropriate fitting of something on to something else. 28 Photographicallikeness is a special case of this, admittedly, but the relevant appropriateness can vary widely, according to the sort of effect which is involved: the appropriate fitness of a foot track to a foot is quite differentfrom that of an offspring to its parents, or of a heated thing to the fire that heats it. Yet these are all cases of this asymmetrical relation of similitude we are now trying to circumscribe, and which holds presumably, in one form or another, between any effect and its cause. 29 Although asymmetrical, however, and closely related with causality, this connection is not to be identifiyd with causality, since it relates something not merely to its particular cause, but to everyotherthing which is relevantly equivalent with this cause as well. It is not itself a transitive relation for all that (a second-order conceptual representation of the concept of 'horse', for example, is not a conceptual representation of horses), but its co-domain is openly generated by another - and quite different - similarity relation, which is both transitive and symmetrical, an equivalence relation in the technical sense. A foot track, for instance, is a similitude, in the required sense, of every (possible) foot which is relevantly equivalent with the one that caused it; the posture of a hand is a similitude, in the required sense, of every (possible) singular object which is relevantly equivalent with the one that originally caused it. This is, I believe, Ockham's basic intuition on the matter. 3. Varieties of conceptual representation It remains to be seen how this model can be applied to intellectual acts in such a way as to yield the required results within the framework of Ockhamism. Ockham holds, in particular, that a single encounter suffices for the formation of a general concept corresponding to a species specialissima, while several are needed for generic concepts. 30 The idea of similitude just developed does account for the representational generality of the conceptual act, and for the recognitional capacities of the intellect, but will the resulting generality always be the right one? Given this understanding of intentional similitude, how are we to explain, in particular, that our intellect reaches specific concepts at one sitting? And how can our comparison with the manual act shed any light on the formation of more general concepts such as that of 'animal', or 'colour', or 'being'? These are difficult questions. I propose to tackle them in a progressive way by successively reviewing the main sorts of simple categorematic concepts countenanced by Ockham: absolute specific qualitative concepts such as 'whiteness' or 'redness' (section 3.1), specific concepts of substances such as 'horse', 'nuthatch' or 'man' (section 3.2), connotative concepts such as 'white', 'taller' or 'similar' (section 3.3), and generic concepts such as 'colour', 'coloured', 'animal' or 'being' (section 3.4). We will try to assess in each case how the signification of the concept - the things it signifies - can be connected with its being an intentional similitude in the sense put forward in the previous section. This. in turn. will shed it revealing light on what the functions of concepts are supposed to be in Ockham's theory.

69 126 CONCEPTS AS SIMILITUDES Absolute specific quality concepts Let us start with the most specific qualitative concepts, like that of a precise colour, a specific brand of red, let's say - which, for the sake of commodity, I will simply call 'redness' in the rest of this discussion. What is striking here is that we cannot appeal to an isomorphism between the concept and the singular rednesses it represents, since a plain redness, typically, has no phenomenal structure. When the intellect forms its first concept of 'redness', it has to do so, in Ockham's doctrine, on the basis of a single encounter with something like a simple object, a given instance of redness. Similitude in this case cannot be structural isomorphism. And of course, it cannot be qualitative resemblance either: the mind does not really become red. In Ockham's doctrine, the first intellectual analog of redness that occurs within the mind must be the posture the intellect takes on when the agent really apprehends some redness. This is the intuitive act. And it is a similitude in the sense identified above. 3l It is normally caused in the mind by a direct encounter with some redness; yet in so far as it is a similitude, it pertains to all rednesses if - as Ockham is committed to suppose - a distinctive intellectual posture is adopted whenever the agent comes across a redness. This original act, according to the theory, leaves in the mind a derivative trace which is also an intellectual analog of all rednesses, but which can be reactivated in the absence of any redness. This one is the concept. It is also a posture of the intellect: this precise posture, namely, that it becomes.able to take on if and only if it has previously met with at least one redness. Since any redness will do as its trigger, this concept can be said to be a similitude of all of them equally, and it will thus enable the agent to correctly re-identify new rednesses when they occur: the relevant equivalence relation between the represented things in this case is the cospecificity - or maximal essential similarity - of all rednesses among themselves. We thus reach the required result: the specific concept of redness is a similitude pertaining in the same way to all rednesses, and to nothing else; it can be caused in the mind by a single encounter with some particular redness in optimal conditions. So far so good! 3.2 Specific concepts ofsubstances The case of substantial specific concepts, like 'horse' or 'nuthatch', is more delicate. The problem is that while qualities are immediately grasped in themselves according to Ockham; this is not so for substances: 'it is by the accidents that individual substances are known to us, because the accidents are cognized directly and in themselves by our intellect, but not the substances'.32 How, then, can an adequate general concept of a given species of substances be formed upon a single encounter? One thing is sure: we cannot expect, on such a doctrine, that the intellect which forms the concept of 'nuthatch', let's say,.should ipso facto develop in so doing a deep understanding of the essences of nuthatches, of their internal structure or of the necessary conditions something must meet in order to be a nuthatch. Ockham is certainly not supposing that the mind is endowed with some sort of metaphysical X ray which allows it to grasp the intern;l1 essence of a thing on the basis of a single perceptual encounter. In so far as the general concept of nuthatch is a similitude, therefore. it cannot he a similitude of the internal structure of the nuthatch. since this structure is not disclosed at first sight to the human mind. On the other hand, substantial concepts are not inferred for Ockham; they are not constructed as complex descriptions or nominal definitions of the form: 'whatever is the substratum for such or such perceptible accidents'. Ockham makes it clear that we have, in his view, simple intuitive and abstractive cognitions of substances as well as of accidents. 33 His whole epistemological theory could not fly without this assumption. The acknowledged impossibility of grasping a substance in itselfis not seen by him as a complete impossibility of intuiting the substance: even if through the accidents, we do have simple intuitions of substances, and simple substance concepts are acquired on this basis. The theory of the cognition of substances Ockham supposes is thus approximately the one John Buridan will make explicit a few years later. 34 Everybody admits, Buridan says, that the cognition of accidents is necessary in this world for the cognition ofsubstances. It should not be concludedfrom this, however, that substances are not apprehended by the senses, the imagination or the intellect. The substance is grasped, indeed, but 'confusedly', along with its accidents: 'it is by a direct representation that a substance with its accidents confusedly represents itself to the sense and, through the sense, to the intellect, which can extract it'.35 When I perceive a nuthatch, according to this theory, what I grasp - both by the sense and the intellect - is not merely a certain shape and a certain arrangement of colours, but the substance which is so shaped and coloured. This substance, admittedly, is not apprehended in its bare purity ('in se', Ockham says), but it is nevertheless present in person to the receptive mind; it is the immediate object of an intuitive act and of the abstractive act that follows upon it. This abstractive act, then, truly is a simple concept of some sort of substances. Yet, if it is to be a similitude, it cannot be a similitude of the internal structure of the perceived individual (which is still hidden to the mind), but only of its outward appearance. Since Ockham takes this concept to be a likeness not only of the particular individual that triggered its formation, but of all those who belong to the same species specialissima as well, he must presuppose that the perceptible appearance is in general a reliable indicator of the species. The shape and colour of a given bird or of a given flower, for example, usually are reliable indicators of the species they belong to. Ockham's presupposition is that this is the general and normal situation with respect to perceptible substances. The natural working of our mind must rest on this relative regularity in the connections between substances and the typical accidents which shape their outward appearance. Without such regularities, the human mind could not form reliable intellectual similitudes of external substances, at least not on the basis of a few encounters. If this is so, the idea of similitude which is required for specific substantial concepts is very closely related to what we have found in the previous case. It corresponds to the assimilation of the intellect to external objects by means of its capacity to recognize perceptible features, and it is rooted in the typical posture the intellect takes on whenever it comes across a substance of a certain specific sort. The agent, for example, sees a nuthatch. Her intellect, then, takes on a posture of assimilation with respect to the perceptible features of this hird. Once ahstracted from the concrete presence of this particular nuthatch, the remaining intellectual posture is supposed to he the one that would typically result from encountering any

70 128 CONCEPTS AS SIMILITUDES 129 other nuthatch. The difference with simple qualitative concepts is that the ones we are now considering pertain not to the qualities as such, but to the things that have these qualities. It must be supposed, accordingly, that our intellect is innately capable of (at least) two sorts of different assimilating postures, one that is actualized when it attends to one or more qualities, and the other one when it reacts to the things that have these qualities. Those must be something like a priori categories of our intellect. This is not said by Ockham in so many words, of course, but I do not see how he could avoid supposing it. Or how we could, for that matter. The idea, in short, is that the intellect sometimes is in substance"mode. When it is, it assimilates itself to some external singular substances via the perceptual scheme which typically corresponds to this specific kind of substances. This amounts to say that the intellect then forms an abstract similitude of a given individual substance which it has encountered and of all the other ones which present the same typical perceptual features. Does this mechanism yield the right result? Namely the formationof simple substance concepts corresponding to natural kinds, such as the concept of 'horse' or 'nuthatch'? The answer is 'yes', if we suppose, as Ockham did, that each natural species distinctively displays certain typical perceptual features. This is not as naive as it might seem. Ockham takes it to be a commonly accepted principle that if two individuals are of the same species, they will always produce highly similar effects on equivalently disposed patients. 36 It follows from this that two individuals which belong to the same species specialissima should produce, in normal circumstances, highly similar perceptual effects upon human observers. And we can work here, moreover, with a very wide notion of perceptual feature, including, for example, not only shape and colour, but also the typical sounds emitted by the object, its perfume, its ways of moving, and so on. Perception, in this context, can even be extended to empathy, that is, the noninferential apprehension of the perceived object as being itself a subject for sensations, emotions, and the like. The concept of a nuthatch, for example, will be the intellectual act that results from the typical posture the mind takes on when it apprehends, in optimal conditions, a moving and tweeting nuthatch, and apprehends it as a living being. It could still happen, of course, that the correlation fails in special cases between essence and outward appearances. We can imagine a perfect imitation of a nuthatch withrespect to all its perceptible features, which wouldreally be a teleguidedrobot, can't we? Ockham does not discuss. such cases, but his theory of substance concepts cannot rule them out as impossible and should therefore take them into account. What he has to say, obviously, is that the nuthatch-robot does not belong to the extension of our simple concept of 'nuthatch', the one we are supposed to have naturally acquired on the basis of our previous encounter with a real nuthatch. And this has far-reaching consequences. It means that although intentional similitude is what opens the way for generality in representation, it does not adequately determine by itself the extension of our simple substance concepts. The extension of such concepts in Ockham's framework should contain only individuals which really belong to the same species as one another, individuals, that is, which have basically equivalent causal powers. This is how substance concepts can play their cognitive role in our mental apparatus: the falling of an individual x under the concept of 'nuthatch' is supposed in etlect to warrant the conclusion that x has all the causal powers a nuthatch normally has. Intentional similitude, on the other hand, is what helps us recognize, on the spot so to say, whether something falls or not under a given concepty What the case of the nuthatch-robot reveals is that intentional similitude can in special conditions be misleading. And if it can be misleading, then it cannot constitute the sole criterion for what belongs in the extension or not. Another ~ and more fundamental- function of substantial concepts must be at play here: the correct guiding of expectations with respect to the causal powers of things. The notion of intentional similitude, then, turns out to be basically the same for substance concepts as we found it to be for qualitative concepts, but its semantical role is now quite different. To say that a substance concept is a similitude of what it represents is merely to say that it is a typical posture the intellect becomes able to take on in virtue of having been in touch - through the senses ~ with individuals of a certain sort. And the relevant equivalence relation between the represented things is, just as in the case of qualitative concepts, maximal similarity (or cospecificity). The great difference is this: the cospecificity in question is now between substances rather than qualities. And since substances are not cognized in se but only through their accidents (at least in this world), the intentional similitude relation cannot determine alone the extension of substance concepts as it does for qualitative concepts. Intentional similitude still comes out as an indispensable feature of simple substance concepts, and itshould provide in normal situations a quick and reliable recognitional device for categorizing substances. But since the causal powers of a substance are not all immediately apparent in perception, a discrepancy remains possible in principle, given the basic function of substance concepts, between the individuals they resemble and those they signify. 3.3 Simple connotative concepts Although there might be a few exceptions, simple connotative concepts in this regard are more likely to be like qualitative concepts than like substance concepts: in most cases, intentional similitude will decisively determine their correctness of application. Let me explain. What characterizes simple connotative concepts, as was seen in chapter' 6, is that they are naturally acquired general representations of ordered n-tuples. They can be considered,in other terms, as mental models for ordered groups of entities out there in the world. Thus understood, they are similitudes of these ordered groups, in the by now familiar sense that they correspondto the typical postures the intellect becomes able to adopt as a result of having been in intuitive contact with such ordered n-tuples in the way described in chapter 6. Different sorts of cases can occur according to whether the primary and secondary significates of these primitive connotative concepts are substances or qualities, but their correct application, most of the time, will solely depend upon the perceptual features of the relevant situations, as we can check by successively reviewing the main possibilities. Case 1: Both the primary and secondary sign(ficates are qualities This is exemplified, for instance, by the situation Ockham mentions in Quodliht'fa IV. 17. of someone who simultaneously grasps two whitenesses and forms as a result

71 130 CONCEPTS AS SIMILITUDES 131 a certain concept of similarity.38 The relevant connotative concept of similarity' in this example does not connect the substances that have these whitenesses (as the standard Aristotelian notion of similarity would do), but the whitenesses themselves, and it clearly must be ascertainable in direct experience, just as the basic qualitative absolute concept of whiteness is. It willcorrectly apply to every couple of singular colours which are spontaneously recognized as similar in optimal conditions of observation. Unless surrounding conditions - of light, for example - are unfavourable, intentional similitude in such a case leaves no room for error. Case 2: The primary significates are substances while the secondary significates are qualities (or vice versa) A paradigmatic example of this is the connotative concept 'white', which primarily signifies' certain substances - those material bodies that have a whiteness - and secondarily signifies, or connotes, certain qualities, the whitenesses themselves. Although substances are involved here, the ascertainability of such concepts seems as strong as it was in the previous case if the conditions of observations are optimal. For the recognition of a certain substance as being white in no way depends upon a correct identification of what this substance intrinsically - or essentially - is. It is sufficient that our mind be able to intuitively grasp a certain quality - a whiteness in this example - as being inherent to a substance, which is a capacity Ockham explicitly acknowledges. 39 As was explained in the previous chapter, an absolute substance concept will normally be formed on the same occasion (for example, the concept of a man or of a horse), but the empirical application of the connotative concept 'white' will not in general depend upon the adequacy of this substance concept. In so far as intentional similitude accounts for the recognitional abilities of the agent, it should be just as reliable in the case of 'white' as it is in that of 'whiteness'. Case 3: Both the primary and secondary significates are substances Two situations must be distinguished here, and they yield different results with respect to the role of intentional similitude. 3a) The first one occurs when a simple connotative concept applies to certain couples - or n-tuples - of substances in virtue of certain observable qualities that they have. 'Being taller' would be a good example: both its primary and secondary significates are substances (respectively: the taller bodies, and the shorter ones), but they both are so signified by it in virtue of their external observable features. The empirical application of such a concept to a given couple of substances, then, does not depend, any more than in cases 1 and 2, upon the corn~ct identification of their essences. 3b) The other situation is when the connotative concept applies to certain couples - or n-tuples - of substances in virtue of what they intrinsically are. An example here would be the relational concept 'being essentially similar' in so far as it applies to substances. Such a concept, presumably, must be among those simple intellectual representations the human mind is naturally able to form, according to Ockham. If a concept like 'man', for instance, can naturally be abstracted and spontaneously applied to both Socrates and Plato, as Ockham claims,40 the mind, surely, can form from this the relational concept of 'being essentially similar', just as it can form the concept of 'being qualitatively similar' from having met with two white things (or two whitenesses).1f so, such a concept will be, according to the tack we are presently following, this typical posture of the intellect that it can adopt as a result of having met with two different individuals to which it spontaneously applies the same absolute substance concept. But since the latter recognitional application can in principle be mistaken in special cases - remember the nuthatch-robot - so can the judgement that x and yare essentially similar to each other. And for the same reason: the internal essences of things, although they are intuited, are not cognized in se, but only along with their accidents. We thus have, finally, a category of primitive connotative concepts which, in this regard, are more like simple substance concepts than like simple qualitative concepts. The presumption, however, is that this category has very few members. Apart from 'being essentially similar' - or 'essentially different' - I can think of no other connotative concept which could plausibly be naturally acquired on the basis of intuition, while connecting only SUbstances independently of their accidents. Be that as it may, the net result ofthis review is that we have in principle two sorts of simple connotative concepts (even if unequally represented), just as we had previously found two sorts of simple absolute concepts: those that signify substances only, and those that are true or false of certain n-tuples (whether substances or qualities) in virtue of their connection with empirically ascertainable qualities. Intentional similitude - understood as the typical posture the intellect adopts when it meets with certain objects - accounts in both cases for the recognitional capacities of intellectual agents. And it depends in both cases upon the perceptual features of the things represented. Because of this, precisely, there can in principle be a discrepancy in the former case (merely substance-related connotative concepts) between what the conceptresembles and what it signifies, as there can be in the case of absolute substance concepts, while this is excluded in principle for quality-related connotative concepts, where intentional similitude directly guarantees adequacy, as it does for absolute quality concepts. 3.4 Simple generic concepts The intellect, according to Ockham, can also acquire more general concepts of various kinds on the sole basis of experience, without engaging into compositional activity: absolute substance concepts, for example, such as 'bird' or 'animal', or absolute qualitative concepts such as 'colour', or again general connotative concepts such as 'coloured'. What is required for their implementation in the mind is that the agent should have been in intuitive contact with representatives of more than one species of the same genus}1 When I meet with a nuthatch, the first general concept I acquire is a specific concept, that'of 'nuthatch' in this case: if I then meet with a chickadee - which is somewhat similar to a nuthatch, but not maximally so - I naturally acquire a more general concept, which applies both to nuthatches and chickadees,42 and so on. We have been wondering with Crathorn, at the beginning of the present chapter, how such gcncril: concepts could still plausibly be seen as similitudes of what they

72 132 CONCEPTS AS SIMILITUDES 133 represent, especially as they get more and more general. How, for example, can my concept of 'animal' equally resemble a nuthatch and an elephant, but not a pine tree or a chariot? The notion of intentional similitude reached in section 2 above no\\( allows us to provide a satisfactory answer to these Crathornian worries. A generic concept simply is this characteristic posture that the intellect becomes able to adopt after having met with several individuals of different species. The Ockhamistic conjecture, here, is that as a result of having seen a nuthatch and a chickadee, I must have acquired a new mental capacity, that of adopting a certain determinate intellectual posture that can be reactivated by certain individuals (small birds, let's say), but not others. Such a recognizable posture will be an intentional similitude of everything it pertains to in this way, exactly in the same sense that the original specific concepts dealt with in the previous subsections are intentional similitudes. My concept of 'animal' will thus be said to be a similitude of nuthatches and elephants (among other things), but not of pine trees and cars, because it is an intellectual posture I have acquired the capacity of adopting by meeting with many different animals, and that can typically be reactivated, in normal circumstances, by nuthatches and elephants among others, but not by pine trees and cars. In this conception of intentional similitude, there is no limit in principle to the possible generality of intellectual representation. The more general concept, that of 'being' (ens), is a similitude, Ockham says, ofan infinite number ofpossible objects, every thing indeed that can exist. 43 There is no way to understand this if we stic~ to either the Aristotelian or the picture conceptions of resemblance. But the statement loses its halo of mystery when we realize that an intellectual state can legitimately be called a similitude, in Ockham's vocabulary, of whatever could have produced its original implementation in the mind and can reactivate it in normal favourable circumstances. The concept of 'being', admittedly, constitutes a special case for Ockham, since, contrary to' all other generic concepts, which require several encounters, it is supposed to be immediately impressed upon the mind by any single encountered thing, simultaneously with the corresponding specific concept. 44 Yet the sense in which it is an intentional similitude is exactly the same as in the other cases we have reviewed: it is the typical posture the mind is able to adopt in virtue of having met with any existing object. Since there are absolute substance concepts, absolute qualitative concepts, and connotative concepts among our generic mental terms, just as there are among the specific ones, the connections will be the same between signification and intentional similitude in these new cases as they were in the previous corresponding ones. In qualitative generic concepts, in particular - whether absolute or connotative signification should automatically coincide with similitude: the generic qualitative concept of 'colour' signifies whatever qualities it resembles (in thti relevant sense), and the connotative concept of 'coloured [thing]' signifies whatever couples of substance and quality that it resembles (in the relevant sense). The absolute generic substance concept of 'bird', on the other hand, signifies every bird and normally resembles (in the relevant sense) every bird as well, but since the intentional resemblance in such cases is based on outward appearances, a discrepancy between signification and similitude is possible in principle, just as it was for the more specific concept of 'nuthatch' or for that of 'being essentially similar' (as applied to substances). Such discrepancies are rendered possible in both generic and specific concepts of substances by the fact that concepts, as we may now conclude, have two main functions in Ockham's epistemology. They account, on the one hand, for the recognitional capacities of the intellect,45 and they determine, on the other hand, via their signification and supposition, the truth-conditions of the mental propositions in which they suitably occur. 46 Being based on the outward appearance of things, intentional similitude has primarily to do with the recognitional function. It is to be supposed, within the Ockhamistic framework, that when the intellect readopts a determinate posture previously rendered familiar by some perceptual experience, it has some awareness of this. And this awareness, precisely,. is what prompts it to classify a newly met object within the same species or genus as some previously encountered ones. Intentional similitude, in other words, accounts for our grouping dispositions. But grouping things together typically determines our expectations towards them: identifying something as a trout on the basis ofits external appearance 'raises expectations as to its behaviour, its taste, and so on. Whether these expectations are fulfilled or not depends upon the objective truth-values ofourmental propositions (of our thinking, for example, that this thing in the water is a trout). In so far as some of our naturally acquired concepts are supposed to help us classify things according to their essential causal powers - substance concepts, namely - a discrepancy is bound to be possible in principle between the grouping dispositions which are associated with these concepts and what their contribution is supposed to be to the truth-conditions of our beliefs and expectations, between, in other words, similitude and signification. Fortunately, however, the outward appearance of things normally is, under good conditions of observation, a reliable indicator of what they intrinsically are and of what causal powers they have. A nuthatch normally looks like a nuthatch, and what looks like a nuthatch normally is a nuthatch. Ockham, on the whole, is quite justified in thinking that our simple concept of 'man' normally allows us to correctly judge whether something is a man or not. If true of specific concepts such as 'nuthatch' or 'man', this probably holds too - although to a lesser extent, admittedly - of our naturally acquired generic concepts of substances such as 'bird' or 'animal'. 4. Two problems about absolute concepts Our proposed elucidation of intentional similitude now puts us in a favourable position to discuss two subtle difficulties that have been recently raised against Ockham's nominalistic theory of absolute concepts, one by Gyula Klima (section 4.1), and the other by Deborah Brown (section 4.2).47 They don't seem to have much to do with each other at first sight: Klima's puzzle pertains to the simplicity of absolute concepts, while Brown's concerns the purported absence of synonymy in Ockham's mental language. Yet a correct understanding of intentional similitude and its role turns out to be decisive in both cases for their solution. 4./ Klima's ohjcctiol1 Oyula Klima is currently one of the most penelrating analysts of the philosophical import of late medieval semantics and onlology:*h In a reccnl disl'ussion on his

73 134 CONCEPTS AS SIMILITUDES 135 website, he has claimed that Ockham's special brand of nominalism must fail, because, precisely, of how it copes with absolute concepts. 49 There is a crucial requirement, Klima holds, that should be satisfied by any sound theory of conceptual representation, but that Ockham's semantics for absolute concepts fails to satisfy. The requirement in question is the following: 'our concepts should represent particulars belonging to the same kind in that respect in which these particulars should be similar in order to belong to the same kind. Otherwise a concept could not represent all individuals of the same kind on the basis of acquaintance only with a limited number of them.' Once this requirement is made plain, it is obvious that Ockham's account of absolute concepts cannotmeet it since absolute concepts, for him, as Klima points out, do not represent their objects 'in respect of something'; they simply have them in their extension, period. Now, Klima is certainly right on this last point, as our own discussion of intentional similitude confirms. Absolute concepts are indeed similitudes of their objects for Ockham, but they cannot be said to signify these objects in respect of the properties according to which they are similitudes of them. This is especially clear in the case of simple substance concepts such as 'nuthatch' or 'man'. In so far as these intellectual representations are similitudes of nuthatches or men, this must be, as we have explained, on the basis of the perceptible or phenomenal features of the represented objects. Yet a substance concept of this sort is not a mental sign of these features, but only ofthe cospecific substances that display them. It is not analysable, as a complex sequence of simpler - qualitative or substantial ~ concepts, and it does not connote in any way the perceptible qualities in virtue of which it resembles its objects. Being an absolute concept in Ockham's sense, its proper semantical import is restricted to its extension. And this extension contains only singular substances, each one of which being a member of it to the same degree as any other. 50 Intentional similitude opens up the way to generality in representation, but the internal structure of the representation, in so far as intentional similitude requires such a structure, must not be ideptified with its semantical content. Seen as a posture of the mind, an absolute concept might well be a complex psychological state - how, otherwise, could it be a similitude of its objects with respect to their perceptual features? - yet it is to be counted as a simple concept in so far as none of its parts is itself a concept. This, indeed, is the very reason why a discrepancy becomes possible in principle in the case of substance concepts between. what they resemble and what they signify. Although Ockham's absolute concepts might resemble their objects in respect of their outward appearance, Klima is right that they do not signify them in respect of anything. It is true, then, that Ockham's theory of absolute concepts does not satisfy Klima's requirement. But why should it? What Klima presupposes here is that since all the significates of an absolute concept are supposed to resemble each other, they must do so in some respect. Which is why, in his view, the absolute concept in question must signify them in this very respect. Butthis is a notion Ockham does not share, and does not need to share. What determines the extension of an absolute substance concept, for Ockham, is not any connection that its members have with another domain of objects (respects, features, qualities, or the like), but the relation thl'.v have with each other. And this relation, as we have seen,must he an equivalence relation, both symmetrical and transitive. It is, more precisely, what we have called 'essential similarity': maximal essential similarity in the case of specific concepts such as 'nuthatch', and less stringent essential similarities for generic ones. There is nothing in respect of which all nuthatches have to be similar in order to fall under the absolute concept of 'nuthatch'; they simply have to be essentially similar to each other - or 'cospecific', if one prefers. For this approach to yield the right result in the case of specific concepts, it has to be assumed - as Ockham certainly does - that each individual belongs to one, and only one, species specialissima. Once a certain sample s is present to the mind, the group of all individuals which are of the same specific kind as s is thus uniquely determined, and so is, consequently, the extension of the simple specific concept naturally formed by the intellect on this occasion. This does not mean that the agent will automatically be able to recognize as such any new instance of this particular kind that she might come across in the future, without any possibility of mistake. Two quite different sorts of similarities are involved in Ockham's theory of absolute concepts: essential similarity among their significates on the one hand, which determines the extension of the concept, and intentional similitude on the other hand, which accounts, I take it, for the recognitional dispositions of the agent. The latter,admittedly, must depend upon certain observable features of the objects, and it is true that Ockham takes it to be normally reliable, buta discrepancy always remains possible in principle, as we have insisted, between what the agent is disposed to recognize as falling under a given absolute substance concept of his, and what does in fact fall under this concept. Falling or not under a given absolute specific concept is a matter of truly being cospecific or not with certain salient instances of this concept, and not a matter of having such and such recognizable features. A non-trivial picture of the working of the mind is involved here, no doubt. But not an implausible one. It seems reasonable to suppose that our intellect is so constructed -either by God or by evolution - as to form spontaneously, upon any single encounter, a general mental sign which is true of all individuals of the same kind as the encountered one, whether we are easily able to recognize them or not. Since being of the same kind involves having highly equivalent causal powers, such a mental device should tum out to be most useful in the long run for guiding expectations, and one especially interesting brand of them in particular: conditional expectations. Once I have acquired the specific concept of a nuthatch, for example, I know that all the significates of this concept will essentially have the same basic causal powers as the original samples do, even if these causal powers are not yet known to me. What I know, in other terms, is that if one of my salient samples of nuthatch turns out to essentially have a given causal power, all the other significates of this mental concept will have it as well, precisely because'all agents of the same specific kind [species specialissima] are productive ofeffects ofthe same sort',51 This is how Ockham can admit that true generalizations can sometimes be reached on the basis of a single experiment. 52 Since absolute concepts are supposed, in addition, to come along with certain recognitional capacities (in virtue of intentional similitude), the ability to produce such true generalizations, and to act on them, is undoubtedly an invaluable advantage. even if, as we have acknowledged, the recognitional aptitude should fail in exceptional cases (think of the nuthatch-rohot again!).

74 136 CONCEPTS AS SIMILITUDES 137 The situation, admittedly, is somewhat more complicated for generic concepts, and Ockham does not give much indication as to what concepts exactly we are expected to acquire from having met with a nuthatch and a chickadee, let's say, or a nuthatch and a duck, or a nuthatch, a duck and a trout, and so on. But the principle, basically, should be the same. What we are to suppose is that at every level of generality, the significates of an absolute concept are uniquely determined by some degree or other of essential similarity among them, according to what function exactly the concept in question is naturally designed to fulfil within mental propositions, given the way our minds are functionally organized. The main point is that whether they are specific or generic, absolute concepts do not have to signify their objects in respect of anything. Klima's requirement, then, is simply bypassed. 4.2 Brown's puzzle In a stimulating paper published in 1996, Deborah Brown has claimed to identify a tension within Ockham's theory of absolute concepts which, she says, threatens the internal coherence of his whole philosophical project.53 The problem is that absolute concepts, in Ockham's view, are acquired in a purely passive way, through a process described in strictly causal terms. If so, there seems to be no reason why someone might not independently - and unknowingly - acquire two distinct absolute concepts for the same individual: 'One would think that a person could easily acquire the concept "Marcus" in an encounter with Cicero on one day's visit to the Forum and the concept "Tullius" from another encounter with Cicero without realizing that she was encountering one and the same individual.'54 Such a possibility, Brown says, can be generalized to any' absolute term whatsoever, and it bluntly'conflicts with two principles of Ockham's theory ofmental language: (i) that concepts are individuated only by their signification, and (ii) that there is no synonymy in mentalese. What is questioned here, ultimately, is the compatibility of Ockham's causal account of the implementation of absolute concepts with the fundamental semantical role he attributes them. Before getting to the heart of the matter about this difficulty, we need to clear up a couple of preliminary points. First,Brown's Cicero example is ill-chosen. Ockham, as we have seen, is explicit that the human mind cannot form simple proper singular concepts, precisely because concepts are intentional similitudes. 55 The singular 'Marcus' concept formed by the agent upon her first encounter with Cicero, consequently, will not be an absolute concept (no complex concept can be absolute in Ockham's theory), but a complex connotative one, and so will be the singular 'Tullius' concept formed on the second visit to the Forum. The resulting coextensivity of two distinct complex connotative concepts is no problem for Ockham: since these concepts differ as to what they connote - different circumstances of encountering, for example - they will not wholly signify the same things under the same modes, and will not, therefore, be synonymous. Secondly, it is not the case that Ockham's concepts are to be individuated solely by their semantical properties, as Brown assumes. As I have argued in chapter 3, two distinct intellectual acts within a given agent can be counted as tokens of the same concept, in Ockham's doctrine, only if they both belong to a single causal chain originally anchored in one determinate intuitive encounter. 56 Concepts. in this view. are individuated by their causal history as well as by their semantical properties. It follows that if an agent should unknowingly originate on different occasions two independent causal chains of absolute conceptual acts with the same extension, those would be counted as two different ooncepts in the Ockhamistic doctrine. These remarks, however, do not suffice to solve Brown's puzzle. They merely help to locate it more precisely. The problem arises orily if it should be possible for a human agent to form two distinct, but coextensive, general absolute concepts. And this, if possible, would not conflict with two Ockhamistic theses, but only with one, namely that there are no synonyms among the simple concepts of a given agent. The question, then, is the following: should it be admitted as possible, given Ockham's causal theory of concept acquisition, that an agent unknowingly acquires two distinct absolute concepts of exactly the same objects? If what we have said earlier about a possible discrepancy between what a substance concept signifies and what it resembles, is correct, then the answer to this question is: yes, it is possible in principle for an agent to form two distinct and coextensive absolute substance concepts, but this, from Ockham's point of view, must be quite exceptional. The normal situation, Ockham thought, is that in good conditions of observation, a member of a given species can be recognizeq as such on the basis of its outward appearance. The case of general substance concepts in this regard is very different from the Cicero example Brown started with. Even if Brown's agent had not recognized Cicero on her second visit to the Forum (because, let's say, he had grown a beard in the meanwhile, and lost all his hair), she would normally have recognized him as a human being without any problem if she was close enough and the light was good. It might be very' difficult to say whether this nuthatch in front of me is the same individual I saw yesterday, but it would normally be much easier, if the bird is close enough and the light is good, to identify it as being of the same species. The point, here, is that what the agent acquires when she first meets with a nuthatch, let's say, is' a capacity to adopt a certain intellectual posture. This posture, which is said byockham to be a similitude of nuthatches, will be reactualized from then on in two sorts of circumstances: it will occur, even in the absence of any nuthatch, as a component of certain complex propositional acts of the agent, the truth-conditions of which, then, will have to do with nuthatches; it will be reactivated, on the other hand, on any new encounter with a nuthatch (under favourable conditions of observation), thus bringing about in the agent's mind the recognitional judgement that this is a nuthatch. Absolute substance concepts can be identified with such causal chains of intellectual acts and habitus. If everything is normal, then, once an agent has acquired a certain absolute concept in the guise of a determinate causal chain of this sort, she will not independently ~cquire a new one with the same extension. A simple concept being an intentional similitude normally associated with a reliable recognitional capacity, the formation of two distinct but synonymous absolute concepts must be unusual according to this doctrine. Even though unusual, however, it is true that it cannot be ruled out. Since we have admitted the possibility in principle of a discrepancy between signification and similitude in substance concepts, we should also admit the possibility that the recognitional aptitude fails. even under favourahle conditions of ohservation. if the world docs not collahorate. Individuals of a given species might sometimes

75 138 CONCEPTS AS SIMILITUDES 139 misleadingly look different,5? and. individuals of two different species might misleadingly look alike (like water and XYZ in Hilary Putnam's famous tale of the Twin Earth).58 Brown's consequence, then, reappears, albeitin a milder form: even if the case is exceptional, a human agent might still end up with two distinct but synonymous simple concepts in his mental apparatus. This does not constitute, however, such a deep threat to the consistency of Ockham's doctrine as Brown thinks it does. The non-synonymy thesis concerning simple concepts, although characteristic, is not nearly as central to Ockham's system as it has been taken to be. It is true that the Venerabilis Inceptor reaffirms the principle on several occasions: 'there is no multitude of concepts corresponding to the multitude of [spoken] names'.59 But one has to look carefully at what use he really puts it to. His goal in the passages where it is invoked (Summa Logicae I, 3, and Quodlibeta V, questions 8 and 9) is to identify the grammatical categories that have to be postulated in mental language. The important principle at work in this context is the positive one:, any distinction among grammatical categories which is needed for the sake of signification must be present in mental as well as in spoken or written languages. The negative converse principle, according to which the grammatical distinctions which are not semantically relevant are not to be found in mentalese, is much less strongly asserted, and obviously considered less important. About the distinction between participles and verbs, for example, Gckham is content to conclude that 'there does not seem to be any.great necessity to postulate such a distinction among mental terms'.60 And the corresponding negative general rule is presentedmerely as a convenient guideline: 'for we can conveniently eliminatefrom mental names all of those grammatical features with respect to which spoken names can differ, while remaining synonymous'.61 What we have here is a methodological principle of parsimony, the import of which, when it does apply, is the weak conclusion that it is not necessary to postulate a certain distinction in mental language,62 and not the strong conclusion that it is imperative not to postulate it. Gckham's application of these principles, moreover, has to do in effect only with the admission or rejection of whole grammatical categories in mentalese. The corresponding thesis about the rejection of simple synonymous concepts, while clearly asserted, occurs in virtually all cases as a mere comparison, usually introduced by 'sicut':63 what Gckham means to say in all these cases is that just as there is no need to postulate a multiplicity of synonymous concepts in mental language, there is no reason either to postulate the presence in it of those general grammatical distinctions that have no semantical relevance. I do not see any great problem, consequently, in adopting a relaxed attitude towards Ockham's nonsynonymy thesis about simple concepts, and seeing it merely as what he took to be the normal situation. That exceptions should be admitted as possible in exceptional cases does not seem to be deeply troublesome. There are two reasons, it seems to me, why some commentators - Brown among others - have thought otherwise, and have taken the non-synonymy thesis to be a strong and central claim in Gckham's thought. 64 Both of which, I contend, are wrongheaded. First, it has been thought that the very notion of synonymy in Gckham analytically requires that synonymous terms be subordinated to the same concept. 65 But this is not so. Gckham's preferred sense of synonymy is what he calls the hroad sense, according to which two terms arc synonymous if and only if whatever is signified by one of them under a certain mode is also signified by the other under the same mode. 66 It follows, of course, that if two conventional terms are subordinated to the same concept, they become ipso facto synonymous. But such a common subordination is by no means analytically required by Gckham's definition of synonymy. For one thing, two different speakers might end up with synonymous spoken terms if each one of them has subordinated its chosen word to one of his own concepts, and if the two concepts involved - although numerically distinct, being, ex hypothesis, qualities of different minds - do happen to signify the same things under the same modes. These two concepts, moreover, will clearly be synonymous to each other in this (quite common) case, even though none of them is subordinated to anything. That a similar situation accidentally occurs within the mind of a single speaker can certainly not be analytically ruled out in Ockham's vocabulary. The second consideration that brought many people to think of the nonsynonymy thesis as being of importance in Gckham's system, is somewhat deeper. It is that mental language, they think, must be a logically ideal language if it is to fulfil the role Ockham expects from it of accounting for the semantics of spoken languages. 6? 'What we want from a proponent of mental synonymy', Richard Gaskin writes, 'is a semantical (i.e. meaning-theoretical) role for the alleged existence of synonyms.'68 Thatthere is no such role, however, is no problem for Gckham, because his theory of mental language, as should by now be clear to the reader of this book, is not primarily meant as the construction of a logically ideal language, but as a correct description of what really goes on in the human mind. Reaching a sound epistemology with no undesirable ontologicalcommitmentis, in Gckham's view, the predominant constraint on the theory of mental language. This.theory is also expected to provide in the process an appropriate mentalistic foundation for. the semantics of conventional languages, but it is not a flaw that some of its theses should have no semantical role to play, if they are called for by a satisfactory account ofconcept acquisition. Gn the matter of synonymy, what the doctrine of concepts as similitudes leads to is the double conclusion that (a) the human mind is functionally adapted to nature in such a way that there is normally no strict synonymy among the simple concepts of a given agent; but (b) su~h synonymy can exceptionally happen in special situations ifnature does not collaborate, particularly in the case ofabsolute substance concepts. That (b) should be called for by considerations about the recognitional role of intentional similitude rather than by purely semantical reasons is quite all right given the intent of the theory. There is no serious threat here to its internal coherence. Notes 1. The idea comes from chapter 1 of Aristotle's Perihermeneias, 16a8. 'Similitudo' was the Latin word used by Boethius in the translation of this passage. 2. Adams 1987, pp Ibid., p Altcri 19X9, passim. See in particular pp I have discussed Alferi on this in Panacdo Ili9()b. esp. pp Biard llillli. pp Biard IlJIJ7. pp. 21i-.ll.

76 140 CONCEPTS AS SIMILITUDES , See Michon 1994, pp ; Panaccio 1992a, pp Ord. I, dist. 2, q. 8, OTh II, p. 272: '... intellectus videns aliquam rem extra animam fingit consimilem rem in mente, ita quod si haberet virtutern productivam sicut habet virtutem fictivam, talem.rem in esse subiectivo - numero distinctam a priori produceret extra... Et illud [fictum] potest vocari universale, quia est exemplar et indifferenter respiciens omnia singularia extra, et propter istam similitudinem in esse obiectivo potest supponere pro rebus extra...' (italics are mine). The same point is made, for example, in Ord. I, dist. 2, q. 8, OTh II, p. 279, Ord. I, dist. 3, q. 3, OTh II, p. 428, and Ord. I, dist. 13, OTh III, pp Karger 1994, p Quodl. IV, q. 35, OTh IX, p. 474: 'Ideo dico quod tam intentio prima quam secunda est vere actus intelligendi, quia per actum potest salvari quidquid salvatur per fictum, eo quod actus est similitudo obiecti, potest significare et supponere pro rebus extra, potest esse subiectum et praedicatum in propositione, potest esse genus, species etc., sicut fictum'.(trans!. Freddoso and Kelley 1991, p. 390; italics are mine). See above chap. 2, sect. 2for a detailed discussion of this argument. Quaest. in Phys., q. 2, OPh VI, p. 399: 'Conceptus est similitudo rei extra.' See Michon 1994, p Quaest. in Phys., q.1, OPh VI, p. 398: 'Septimo, quia tale idolum plus differret are quacumquequam quaecumque res differt ab alia, quia ens reale et ens rationis plus differunt quam quaecumque duo entia realia, et ideo tale idolum minus assimilatur rei, et per consequens minus potest supponere pro re quam intellectio quae plus assimilatur rei; et ita minus erit communis rei extra et minus habebit rationem universalis quam intellectio. Sed tale idolum non ponitur propter aliud nisi ut supponat pro re vel ex eo componatur propositio vel ut sit communis ad res etc., et ista verius competunt intellectioni quam tali ido10. SuperfIue igitur ponitur tale idolum vel fictum' (italics are mine). See Michon 1994, pp Exp. in Perih., Prooemium, 10, OPh II, pp : 'Ad aliud potest dici quod tale fictum seu idolum plus distinguitur a re extra quam quaecumque res ab una alia; tamen in esse intentionali magis sibi assimilantur, in tantum quod si posset produci realiter sicut potest fingi, esset vere consimile realiter rei extra. Et propter istam rationem magis potest supponere pro re et esse communis et esse illud in quo res intelligitur quam intellectio vel alia qualitas.' See also Ord. I, dist. 2, q. 8, OTh II, p. 283, ad 2. Quodl. V, q. 7, OTh IX, p. 506: 'Quarto dico quod intellectus noster de nulla creatura potest habere aliquem talem conceptum simp1icem proprium sine visione creaturae nee cum visione; et hoc quia quaelibet talis cognitio sive conceptus aequaliter est similitudo et repraesentat omnia individua simillima, et ita non plus est conceptus proprius unius quam alterius' (trans!. Freddoso 1991, pp ). Quodl. I, q. 13, OTh IX, p. 74: 'Tertio dico quod cognitio prima abstractiva primitate generationis et simplex non est cognitio propria singulari sed est cognitio communis aliquando, immo semper... assumptum patet, quia nulla cognitio abstractiva simplex est plus similitudo unius rei singularis quam alterius sibi simillimae, nee causatur a re nec nata est causari; igitur nulla talis est propria singulari sed quaelibet est universalis' (trans!. Freddoso and Kelley 1991, p. 65). Biard 1997, p. 29. Crathom, In Sent. I, q. 2, ed. Hoffmann 1988, p See Ockham, Quodl. VII, q. 13, OTh IX, p. 749: '... cognitio quae est conceptus entis est similitudo infinitorum obiectorum... ' Crathorn, In Sent. I, q. 2, ed. Hoffmann 1988, p Exp. in Praedic. 15, 3, OPh II, p. 292: 'Notandum quod stricte sumendo simile et dissimile, nihil dicitur simile vel dissimile alteri nisi secundum qualitatem... A wider _. and more indeterminatc- sense is mentioned hy Ockham immediately after: 'tamen ' large forte accipiendo simile vel dissimile, potest aliquid secundum a1iquid aliud dlel simile vel dissimi1e'. See for example SL 1II-3, 26, OPh I, p. 690, where the definition Ockham gives for 'simile' is: 'simile est quale, correspondens alteri quali, habenti qualitatem eiusdem speciei specialissimae'. See also SL 1,55, OPh I, p. 182: '... secundum qualitatem aliquid dicitur simile vel dissimile'. See Panaccio 1992a, pp. 260 ss. This is illustrated, for example, in the passages from the Quodlibetal Questions quoted above in nn. 17 and J 8. See, for example, Ord. I, dist. 2, q. 6, OTh II, p. 211: '... Sortes secundum substantiam est simillimus Platoni.' See the extract from Ord. I, dist. 2, q. 8 quoted above in n. 8, as well as the other passages mentioned in the same footnote. These, admittedly, were written from the point of view of the jictum-theory, but in the last one mentioned - Ord. I, dist. 13, OTh III, p Ockham later added a short sentence saying that the same holds within the aclujtheory: '... esset proportionaliter dicendum de ipsa intellectione... ' See John Buridan, In Metaph. VII, 8, Paris, 1588, fo. 46r: '... accipiendo similitudinem large vel improprie non pro eadem qualitate: immo solum per quadam appropriata convenientia agentis ad effectum'. See John Buridan, Quaest. de anima I, 5, from ms. Munich 761, as editedin an appendix of Patar 1995, p. 50l: '... quilibet effectus habet et gerit quodammodo in se similitudinem suae causae... ' See Quodl. IV, q. 17, OTh IX, p. 385: '... conceptus speciei potest abstrahi ab uno individuo'; and Quodl. I, q. 13, OTh IX, p. 77: '... dico quod conceptus generis numquam abstrahitur ab uno individuo'. Ockham admits indeed that the intuitive act is already a similitude of its object. See above chap. 1, n. 46. Exp. in Porph. 2, 15: '... et per ipsa accidentia innotescunt nobis individua substantiae, quia accidentia directe et in se cognoscuntur ab intellectu nostro, non sic autem substantiae'. See also Ord. I, dist. 3, q. 2, OTh II, p. 412: ' dico quod nulla substantia corporea exterior potest a nobis in se naturaliter cognosci '; Quodl. III, q. 6, OTh IX, p. 227: 'Cum igitur de substantiis non habeamus experientiam nisi per accidentia...' See Ord. I, Pro!., q. 1, OTh I, p. 23: 'Sed certum est quod intellectus potest habere notitiam incomplexam tam de Sorte quam de albedine.' Similarly, the cameo description of the cognitive process given in SL 1II-2, 29, OPh I, p. 557, is based on the privileged example of tlie simple intuitive grasping of a man, followed by the formation of a simple abstractive concept of 'man'. A few pages further in the same chapter, Ockham contrastll his own case, as somebody who has never seen a lion, with that of a person 'who sees or has seen intuitively the substance of a lion' (OPh I, p. 560: '... ille qui videt vel vidit intuitive substantiam leonis'). See in particular John Buridan, Quaest. de Anima (tertia lectura) I, 6, ed. Patar 1991, pp ; and Quaest. sup. Phys.I, 4, Paris, On this theory, see Reina 2002, pp John Buridan, Quaest. de Anima (tertia lect.) I, 6, as edited in an appendix to Patar 1991, p. 790: '... directa representatione substantia cum accidentibus confuse representat se sensui, et sensu mediante, intellectui qui illam potest extrahere... ' See for ex. Ord. I, Pro!., q. 2, art. 1, OTh I, p. 87: '... omnia individua eiusdem rationis habent effectus eiusdem rationis in passo aequaliter disposito'. The same principle is also found in Onl. I. Prol.. q. I, OTh I. pp Belonging to the same species, actually, involves. from (kkham's point of view. having equivalent causal powers. See On/. I, Prol.. q. I. OTh I. p. 42:... causae eiusdem rationis habent etfeclus eiusdem rationis', a principle he considers as commonly IICl cptcd.

77 Ockham is explicit, for instance, that our concept of 'man', in so far as it is equally a similitude of all men, is that 'according to which we can judge about any thing whether it is a man or not' ('... secundum quod nos possumus iudicare de quolibet si est homo vel non', Ord. I, dist. 2, q. 8, OTh II, p. 278). This is said within the framework of the fictum-theory, but it must hold equally well in the actus-theory. Quod!. IV, q. 17, OTh IX, p The text was quoted above in chap. 6, n. 15. See Ord. I, Prol., q. 1, OTh I, p. 31: 'Similiter, notitia intuitiva est talis quod quando aliquae res cognoscuntur quarum una inhaeret alteri... statim virtute illius notitiae incomplexae illarum rerum scitur si res inhaeret vel non inhaeret...' See Ord. I, dist. 2, q. 6, OTh II, p. 211: '... ex hoc ipso quod Sortes et Plato se ipsis differunt solo numero, et Sortes secundum substantiam est simillimus Platoni, omni alio circumscripto, potest intellectus abstrahere aliquid commune Sorti et Platoni...' (italics are mine). See Quod!. I, q. 13, OTh IX, p. 77: '... dico quod conceptus generis numquam abstrahitur ab uno individuo'; and SL III-2, 29, OPh I, p. 557, where after having described the formation of the concept 'man' upon a single encounter with a man, Ockham adds: 'Deinde apprehenso alio animali ab homine vel aliis animalibus, elicitur una notitia generalis omni anima1i, et illa notitia generalis omni animali vocatur passio seu intentio animae sive conceptus communis omni animali.' Ockham does not give any precise clue as to which concept exactly that would be in such a case. Presumably, it would not yet be the general concept of a bird, but something less general. Quod!. VII, q. 13, OTh IX, p. 749: 'Sicut cognitio quae est conceptus entis est similitudo infinitorum obiectorum.' See also Quod!. IV, q. 5, OTh IX, p. 319: '... cognitioentis est cognitio generalis infinitorum... ' Cf. Quod!. I, q. 13, OTh IX, p. 78: 'Semper tamen imprimitur conceptus entis, quia quando obiectum est debito modo approximatum, simul causatur a re singu1ari extra conceptus specificus et conceptus entis.' It can even happen that the concept of 'being' be impressed upon the mind without any specific absolute concept whatsoever: if the intuitedobject is far away, I can realize that something is there in the far distance without having the slightest idea of what it is. Says Ockham: 'Manifestum est quod in isto casu cognitio abstractiva quam habeo primo primitate generationis est cognitio entis et nullius inferioris' (ibid., p. 74). See above n. 37. As explained in detail in chap. 3 above. See Klima 2002, Brown See in particular Klima 1988, 1993, 1999,2001. See Klima See SL I, 10, OPh I, p. 35: 'Nomina mere absoluta sunt illa quae non significant aliquid principaliter et a1iud vel idem secundario, sed quidquid significatur per illud nomen, aeque primo significatur... ' Ord. I, Pro1, q. 2, OTh I, p. 92: '... omnia agentia eiusdem speciei specialissimae sunt effectiva effectuum eiusdem rationis'. See ibid. '... dico quod aliquando sufficit unum experimentum ad habendum principium artis et scientiae... Si enim in principio primo artis et scientiae subiciatur species specialissima, sufficit unum experimentum. Sicut ad sciendum istam propositionem "omnis calor est ca1efactivus", quae est principium primum, sufficit evidenter cognoscere quod iste calor calefacit vel calefecit...' Brown 1996, p. 79. Ibid., p. 88. See the passages from Quodl. V, q. 7, and Quodl. I, q. U quoted above in nn See above,-,hap. ~, sect CONCEPTS AS SIMILITUDES 143 Note, however, that the often striking differences between males and females within a given species of bird, for example, or between youngsters and adults, would not constitute good examples of this. Ockham could easily admit that these correspond to different - and non-synonymous - simple absolute concepts. The standard presentation of the Twin Earth thought experiment is found in Putnam 1975a. Quodl. V, q. 9, OTh IX, p. 518: '... multitudini nominum synonymorum non correspondet multitudo conceptuum' (transl. Freddoso 1991, p. 432). See also SL I, 3, OPh I, pp , and Quodl. V, q. 8, OTh IX, pp SL I, 3, OPh I, p. 11: '... non videtur magna necessitas talem pluralitatem ponere in mentalibus terminis'. Ibid., p. 12: 'Unde quaecumque pluralitas et varietas talium accidentium, quae potest competere nominibus synonymis, potest convenienter a mentalibus amoveri '(italics are mine; transl. Loux 1974, p. 53, with a slight amendment, Loux having simply neglected to translate the adverb 'convenienter'!). About common and deponent voices among verbs, for example, Ockham's conclusion is the following: 'cum verba communia aequivaleant activis et passivis et deponentia neutris vel activis, et ideo non oportet talem pluralitatem in verbis mentalibus ponere' (SL I, 3, OPh I, p. 13; italics are mine). See for example Quod!. V, q. 8, OTh IX, p : 'Et ideo multitudinem talium accidentium quae competunt nominibus synonymis non oportet tribuere naturalibus signis cuiusmodi sunt conceptus mentales, sicut nec nominibus synonymis correspondet pluralitas conceptuum' (italics are mine). The same 'sicut' structure is also found in similar arguments in SL 1,3, OPh I, p. 11, and Quod!. V, q. 8, OTh IX, p See for example Brown 1996, p. 79: 'In his writings on semantics and logic, William of Ockham combines two very stronge/aims about mental language', the second one being 'that mental language contains neither synonymous nor equivocal terms' (italics are mine). This was originally suggested by Spade 1975, p. 66. It is very clearly assumed, in particular, by Gaskin 2001, p. 232: 'We can characterize spoken or written synonymy in terms of subordination to a single mental item, but what would synonymy of mental items be?' Gaskin, consequently, takes 'the absence of synonymy as being constitutive of mental language, rather than as being merely a regulative principle' (p. 231). See SL I, 6, OPh I, p. 19: 'Large dicuntur illa synonyma quae simpliciter idem significant omnibus modis, ita quod nihil aliquo modo significatur per unum quin eodem modo significetur per reliquum... ' The same definition is also found in Quod!. V, q. 10, OTh IX, p The seminal paper, here, was that of John Trentman (1970). Paul Spade, of course, has been the pre-eminent defender of this approach in the last decades. Gaskin 2001, p. 232.

78 Chapter 8 Logical Concepts In addition to absolute and connotative concepts, the vocabulary of the language of thought, according to Ockham, also contains syncategorematic terms.! These correspond for the most part to what we call today 'logical constants': quantifiers such as 'all', 'every', 'no', 'some', 'neither', etc.; connectors such as 'if', 'and', 'or', 'because', 'when', 'where', 'unless', and so on; prepositions such as 'except'or 'as'; some adverbs such as 'only', 'per se', etc.; negations ofcourse, and even the copula. 2 Ockham's most famous thesis about this group of terms is that, in the strict sense of 'to signify', they signify nothing. There are no special objects in the world to which 'all', 'and', 'no' or 'is' could be said to refer: None of these expressions has a definite and detenninate signification, nor does any of them signify anything distinct from what is signified by categorematic terms. The number system provides a parallel here. 'Zero', taken by itself, does not signify anything, but when combined with some other numeral it makes that numeral signify something new. Likewise, a syncategorematic tenn does not, properly speaking, signify anything; however, when it is combined with a categorematic expression it makes that categorematic expression signify something or supposit for something in a detenninate manner, or it perfonns some other function with regard to the relevant categorematic tenn. 3 Ockham's position on syncategoremata thus comes down to what Wittgenstein calls his 'fundamental idea' in the Tractatus: 'the "logical constants" are not representatives'. 4 This thesis, however, raises a problem for Ockham's theory of mental language. While simple categorematic terms, whether absolute or connotative, are supposed to be implemented in the mind as a result of an abstraction process rooted in direct encounters with concrete individuals, no such explanation is available for syncategorematic concepts: since they do not signify or represent anything, they cannot be acquired on the basis of particular encounters; there simply is no syncategorematic object in the world for the abstraction process to get started with. Ockham saw the problem very early and gave it a first answer in the Ordinatia from the point of view of the fictum-theory of concepts. Very briefly expounded, this answer has puzzled recent commentators. Yet it is quite interesting philosophically, and I will first reconstruct it in some detail here (section 1). It will then be argued that, even though he is not explicit about it, logical concepts can only be innate in Ockham's later actus-theory (section 2). The status of some apparently non-logical syncategorematic prepositions and verbs such as 'in', 'of', 'before'. 'to have', 'to inhere in', and so on, will finally be briefly discussed (section 3). 14~

79 The earlier theory: logical words internalized The question of the origin of syncategorematic concepts in the mind is raised in the Ordinatio along with that of connotative and negative terms as an objection to the fictum-theory: The fourth problem has to do with syncategorematic, connotative, and negative concepts: where can they be taken or abstracted from? For if it is precisely from things, then it is not seen how they can be distinguished from other concepts. But it is clear that there are such concepts, since to every spoken proposition, there can correspond a similar one in the mind, and therefore to this proposition 'every man is an animal' and this one 'some man is an animal', there correspond distinct propositions in the mind; something therefore corresponds to the quantifier in one proposition which is not found in the other. s The two propositions used as examples 'every man is an animal' and 'some man is an animal' differ only in their syncategorematic components, but obviously they are not equivalent: distinct thoughts correspond to these two sentences. The presumption, then, is that mental language contains distinct syncategorematic concepts corresponding to the quantifiers 'every' and 'some'. Where can they come from? Ockham's reply holds in a single paragraph, which is well worth a detailed examination. Here it is, almost in its entirety (the letters between brackets are adqe d by me to facilitate further references): [a] To the fourth problem, I say that syncategorematic, connotative and negative concepts are not concepts abstracted from things and capable by their very nature to supposit for things or to signify them in some special way with respect to other concepts. [b] And for this reason I say that there are no syncategorematic, or connotative, or negative concepts, except by mere institution... [c] Such concepts, however, can be imposed or they can be abstracted from words, and this is what happens in fact either always or generally. Strictly speaking, to the spoken word 'homo' there applies such grammatical modes as the singular, the nominative case, the masculine gender, and so on, while to the spoken word 'hominis', there applies other grammatical modes. Similarly, the spoken word 'homo' signifies a thing determinately when taken by itself, while this does not apply to the spoken word'omnis', which signifies only when taken with some other term. And the same holds for the spoken word 'non' and for words such as 'per se' and 'in quantum', 'si' and other such syncategorematic terms. [d] From words which thus signify, then, the intellect abstracts common concepts which can be predicated of them, and it imposes these concepts at signifying the same thing as these external spoken words signify. [e] And in the same way, it forms with such concepts propositions which are similar to the spoken propositions and have similar properties. [f] And just as it can institute such concepts to signify in this way, it can institute the concepts that are abstracted from things to signify in the same grammatical modes as spoken words do. [g] In order to avoid equivocities, however, this is more convenient with the concepts that are abstracted from the spoken words, for these concepts are distinct among themselves just as spoken words are, which is not the case for other concepts. [h] And in this way any proposition can be distinguished, for example the proposition that corresponds to 'homo est homines', or to 'homo est hominis', and so on.1> This is a complex passage, undouhtedly. which, I helieve, has not yet heen satisfactorily explained in the literature. Having said a word in a previous chapter LOGICAL CONCEPTS about the treatment of connotative concepts in it,? I will now concentrate on the doctrine it proposes for syncategoremata. What has surprised the commentators about this doctrine is the idea that mental syncategorematic terms could be derived from the corresponding spoken units. Calvin Normore, for one, sees this as an 'astonishing view' for Ockham to have held: 8 'since spoken language is arbitrary (ad piacitum)', he writes, 'it would make hash of the idea that mental language is natural'.9 And Marilyn Adams expresses a similar worry: 'This proposalmakes the syncategorematic force that binds terms into the subject and predicate position of a proposition derivative from that of conventional language.' 10 To put it in a more general way: all logical operators and connectors - including the copula - would, apparently, be derived from external language, without which, consequently, thought would be devoid of all logical or grammatical structure. Logical structure would be essentially conventional. Which is indeed an 'astonishing' consequence for a fourteenth-century Aristotelian thinkeri And Ockham, in addition, would not evenhave felt the need to explain himself about this. There is obviously something wrong here. We must look at the text more closely. The theory presented there comes down to two theses: (1) There are no natural syncategorematic concepts. (2) There are syncategorematic concepts, nevertheless, but by 'institution' only. Let us consider these two points in turn. Thesis (1), to begin with, cannot mean, as was sometimes believed, that the human mind is naturally incapable of logical operations. This would be radically inconsistent with Ockham's acceptance of both the Aristotelian and the Augustinian frameworks. And it would be incompatible with Ockham's earlier philosophy as well as with his later theory of concepts. Human thought, in his view, can certainly combine a subject and a predicate without the help of spoken language; it can form universal propositions as well as particular ones, negative propositions as well as affirmative ones, conjunctions as well as conditionals, and so on. What thesis (I) asserts, instead, is this: such logical operations do not generate special intelligible objects as categorematic concepts are supposed to be in the fictum-theory. How could they, since syncategoremata represent nothing, while concepts, according to the fictum-theory, have existence within the mind in so far only as they represent something? The key to a sound understanding of Ockham's position here is to remember that he is exclusively adopting in this passage the standpoint of the fictumtheory. Saying thatthere are no natural syncategorematic concepts, in this context. amounts to saying that there are no such intelligibleficta. It does not follow from this that the intellect is incapable of logical operations without the help of conventional language, but only that such operations, when they occur, are not intelligible objects before the mind, as concepts are supposed to be. They are the intellectual handling of ideal objects, rather than special objects themselves. They are what the mind does rather than what it contemplates. An important consequence of thesis (I) thus understood is that it ruins the parallelism that is often expected from a theory of mental language hetween mental and spoken or written propositions. While the latter are linear sequences of discrete 147

80 148 LOGICAL CONCEPTS 149 units, some of which are categorematic and some syncategorematic, mental propositions, in the view we are presently considering, are operations accomplished by the mind upon the sole categorematic concepts. The written sentence 'every horse is mortal', for example, has four successive components, while only two concepts would be involved in the corresponding thought: 'horse' and 'mortal', which the mind would be handling in a determinate way, performing on them a quantification operation on the one hand, and a predication on the other hand. Is this a picture Ockham could have self-consciously subscribed to? My contention is that it is. Adam Wodeham, indeed, who was Ockham's pupil at one point and who usually shows exceptional familiarity with his thought, does attribute such a position to him, at least with respect to the copula(which he calls a 'comparative act').ll Doubts about this attribution, however, have recently been raised by Elizabeth Karger, on the basis of two main arguments.j2 One of them rests on a passage from the Reportatio where Ockham clearly treats the natural concept corresponding to the copula as afictum endowed with mere objective being. 13 He cannot, therefore, have identified the copula with an act, Karger concludes. But the passage in question dates from a very early period in Ockham's career, when he was still accepting the existence of special relational entities of inherence; it corresponds to an even earlier stage in his thought than what is illustrated by the (favourable) discussion of tb,e fictum-theory in distinction 2 of the Ordinatio, and one which he soon moved away from. It should not, therefore, be considered as relevant in the present context. Karger's second argument is that the identification of the natural copula with an act within the fic'tum-theory would bring about an unacceptable consequence for Ockham, namely that: 'every act of assenting to (or dissenting from) a mental sentence would be an act which, though "direct" with respect to the terms of the sentence, which are not mental acts, would be "reflexive" with respect to its copula, which is a mental act'.14 Wodeham saw that consequence and accepted it without qualms,15 as Karger acknowledges, but Ockham could not have, she contends, since it would entail that 'every act of knowledge is partly reflexive', while Ockham admits that 'there are unreflective people who are unaware of their mental acts and who are nonetheless capable of knowledge'.16 Karger's reference, here, is to Quodlibeta III, 8, where Ockham indeed distinguishes two varieties of assent: a reflective one and an unreflective one. 17 The latter, however, which is the one that counts for Karger's argument, is not considered by Ockham as knowledge in the strong sense, as I indicated earlier. IS It is.a simple judicative act in the mind, which is equivalent in many respects to a complex propositional sequence, but which does not have in itself any component parts: 19 strictly speaking, it has no mental subject, no predicate, and a fortiori no copula. Only in the case of reflective assent is a mental proposition with a copula directly involved. Contrary to what Karger suggests, then, Ockham would have no difficulty in conceding that whenever a mental proposition does have a copula, the act of assenting to it is a reflective act. It would certainly be, at any rate, a much more unacceptable consequence for him if all logical operations should tum out to be conventional! When he holds in Ordinatio I, distinction 2, question 8, that there are no natural syncategorematicfieta (as stated by what I called thesis (1), he cannot mean to exclude logical acts such as predication, quantification, negation, and so on, from the range of what human beings are n:;tturally capable of. The non-linear structure of mental propositions _ composed of both logical acts and categorematic conceptualficta - is, collsequently, to be accepted in this doctrine. What thesis (2) tells us in this regard is that the mind can nevertheless form, if it so decides, closer mental analogs for spoken and written sentences, and thus institute in the mentalorder a much stricter parallelism with the linguistic order, presumably facilitating.in this way abstract thought and mental reasoning.. How the mind does it is what Ockham mostly insists upon in the relevant passage. His starting point is that the intellect can form certain concepts by abstracting them from spoken words, just as it can from anything else in the world. And it can do so in particular on the basis olthe conventional features of spoken and written words. This is whatthe development on singular, masculine and nominative case in sentence [c] is all about: 'Such concepts', Ockham says, 'can be imposed or they can be abstracted from words, and this is what happens in fact either always or generally.' This is how we acquire grammatical concepts, such as the concept of 'nominative' or that of 'singular', and semantical concepts too, such as the concept of 'signifying something in a certain mode'. Even more precise metalinguistic concepts can be. formed in the same way, such as the concept of 'the genitive of "homo"', for example, or the concept of 'the syncategorematic term "if"'. Those are concepts, indeed, in the sense of the fictum-theory: they are ideal intelligible objects representing something in the mind. They are metalinguistic concepts, more precisely, since what they represent are linguistic signs. Remember sentence [d] of the quoted passage: 'From words which thus signify, then, the intellect abstracts common concepts which can be predicated of them... ' (italics are mine). The concept 'genitive of "homo''', for example; can be abstracted from singular tokens of'hominis', and it can be predicated of them in the sense that of each one of these tokens, the following thought is true: 'this is the genitive of "homo'''. We have thus reached in this case a precise grammatical metalinguistic concept. An even more directly relevant example for our present discussion is the concept of 'the syncategorematic term "if"': I can have the thought that this written word I am looking at is the syncategorematic term 'if'. The mental unit which occurs as predicate in that thought is a concept representing certain things out there in the world, namely the various tokens of the word 'if'. This concept is not itself, at this stage, a mental syncategorematic term, but a categorematic concept, albeit a metalinguistic one. But now comes the crucial move. Let us carefully read the rest of the sentence: '... and it [the intellect] imposes these concepts at signifying the same thing as these external words signify'. Having formed the metalinguistic categorematic concept of 'the syncategorematic term "if''', the mind can, by decision - by 'institution', Ockham says - endow this very concept with the same semantical value as is conventionally possessed by the word it represents. This is a distinct operation. The metalinguistic concept now receives a new function, not that of representing the tokens of 'if' anymore (in the sense that the concept of 'horse', for example, represents horses), but that of occurring as a syncategorematic term itse(l within mental sequences. The mind can then form mental propositions which are strictly parallel with the corresponding spoken or written sentences. This is precisely the point of part lei in our passage: 'And in the same way. it Ithe intellecti forms with

81 150 LOGICAL CONCEPTS 151 such concepts propositions which are similar to the spoken propositions and have similar properties.' What do we find, now, in the curious sentence [f]: 'And just as it can institute such concepts to signify in this way, it can institute the concepts that are abstracted from things to signify in the same grammatical modes as spoken words do'? It is that the mind could, if it so wanted, do the same with non-metalinguistic concepts ('the concepts that are abstracted from things')! I could decide to use my concept of 'horse' as a genitive in some mental sequences, or maybe even as a syncategorematic term, why not? In Ockham's eyes, the advantage of preferring metalinguistic concepts for this kind of stipulative roles is that they 'are distinct among themselves just as spoken words are, which is not the case for other concepts' (sentence [g]). The mind can.thus form a special mental sequence for each spoken proposition, including those that make use of both the nominative and the genitive of the same word, for example, such as 'homo est hominis', or both the singular and the plural of the same word, such as 'homo est homines' (sentence [h]). This, then, is the theory proposed to us in this somewhat surprising page of the Ordinatia. Let us recapitulate, for the sake of clarity, the various stages of the cognitive process it hypothesizes: 1. There is, first, the normal formation ofsimple categorematic concepts through abstraction, as described in the Prologue of the Ordinatio. According to the text we are now discussing, no syncategorematic concepts, or connotative or negative concepts are formed in this way The mind can, from there, accomplish various operations uponthe concepts so formed. It can combine them, in particular, through predication, negation, conjunction, and so on, and it can quantify over them. These operations are intellectual handlings of previously acquired categorematic concepts, but they are not concepts themselves. Since it is not acquired on the basis of experience, the aptitude to accomplish such operations must be seen as innate within the framework of Ockhamism (more on this in the next section). 3. These intellectual operations can be conventionally expressed in spoken or written languages by special words, the function of which is to indicate to the hearer (or the reader) which mental operations are accomplished by the speaker. These special words are the syncategorematic terms of spoken and written languages. Strictly speaking, they must not be said to signify the operations they express, since they are not used to refer to these operations or to connote them, but their presence can modify the signification, supposition or truth-conditions of the other linguistic units they are combined with, according to what logical operations they correspond to. 4. Like any other things in the world, linguistic items - and in particular the syncategorematic words introduced at the previous step - can be intuitively grasped by the intellect and thus bring about the formation of categorematic concepts capable of suppositing for them within mental propositions. Some of these metalinguistic concepts will be quite general, such as the concept of 'genitive' or 'masculine', and some will be much more narrowly circumscribed, such as the concept of 'the genitive of "homo'" or that of 'the feminine of "albus'''. 5. The metalinguistic concepts thus introduced can then be endowed, by the agent's decision, with a new semantical function, which reduplicates within the mind that of the spoken or written words which these concepts naturally represent. A kind of reverse subordination occurs, in virtue of which a naturally implemented mental unit inherits, by a voluntary institution, the. semantical properties of certain spoken or written expressions The mind, finally, can use these reassigned concepts to construct mental sequences which are strictly parallel with spoken or written sentences, and in which logical operations - like predication, negation, or quantification - are objectivated into ad hoc units. Although Ockham does not say it in so many words, it is to be presumed that such mental objectivations facilitate reflection and reasoning by displaying, so to say, the logical processes before the mind; Even independently of the fletum-theory with which Ockham associated it, this approach is quite remarkable. What it draws attention to is that the process of 'thinking with words' -often taken for granted in recent philosophy - requires that some of our naturally acquired mental representations - our representations ofspoken or written items, more precisely -,-- should be semantically reassigned according to linguistic conventions. Ifso much is admitted, then how such a process works exactly, and how it bears upon the nature and scope of human thought, is certainly something that today's cognitive scientists should be interested in investigating. 2. Logical constants in the actus-theory However that may be, Ockham eventually renounced this approach to mental syncategoremata as he moved away from the fletum-theory. It is true, as Beatrice Beretta remarks, that there seems to be a later allusive reappearance of it in the Quodlibeta,22 but one should probably suppose that the relevant passage - in which Ockham introduces a wide sense for 'second intention' - had originally been written at a time when he was still favouring the fictum-theory and that he failed to adjust this particular sentence as he integrated the whole surrounding development into the final version of the Quodlibeta. 23 It is clear at any rate that the actus-theory of concepts does allow, in Ockham's own eyes, for natural mental syncategoremata. This is explicit in the very presentation of the actus-theory that he inserted in the revised version of distinction 2, question 8 of the Ordinatio, just a couple of pages after the discussion of syncategorematic terms that we have been scrutinizing in the previous section. In the new account, he says,... just as among spoken words and conventional signs... there are Some that do not signify things, but merely consignify along with others, such as syncategorematic.terms..., similarly there are some qualities subjectively existing in the mind which proportionally - have by their very nature such features as these spoken words conventionally have. 24 Concepts. in general, in this approach, are no more seen as ideal objects produced by mental acts and liable to logical operations. They are mental acts themselves -

82 152 LOGICAL CONCEPTS. 153 operations, that is, or states of the mind -'- which do not require special intelligible correlates. The parallelism between mental and spoken or written propositions, therefore,is now more easily secured: mental propositions are concatenations of acts just as spoken and written sentences are concatenations of words; it becomes quite natural to postulate a distinct conceptual act in the mind for each semantically relevant word in spoken or written discourse, whether categorematic or syncategorematic.25 The overall picture can thus be significantly simplified. This is not to say that syncategorematic concepts. are now treated on a par with categorematic ones. They are still not granted a determinate signification of their own and their semantical role continues to be considered as subordinate. In the actus- as well as in the fictum-theory a syncategorematic conceptual act has no semantical relevance if it is not accompanied by some categorematic concepts. Take the paradigmatic case of the copula. The corresponding mental term, Ockham says in Quodlibeta VI, question 29 (written from the point of view of the actus-theory) is 'the union of the extremes of a mental proposition'. It is not an intrinsically relational unit for all that, but 'a quality of the mind, viz., an act of understanding'. 'And this concept', Ockham adds, 'is really distinct from the subject and the predicate, which are also diverse acts of understanding.'26 Yet being an act of 'union\ it seems clear that contrary to the categorematic terms, it cannotnormally be posited alone. And the same must be true of all other syncategoremata. One can, of course, think ofa syncategorematic term without forming a proposition; the mind, then, has a second-order categorematic concept representing the syncategorematic term which is what Ockham calls a 'second intention' (intentio secunda);27 but it is not actually producing in this circumstance a truly syncategorematic mental act as such. What is the difference, Paul Spade asks at one point, 'between merely having the three concepts "Socrates", "is", and "mortal", and having the mental proposition or judgment "Socrates is mortal"'?28 Ockham's answer to this, in the framework of the actusctheory, must be that there is no difference there, as long, at least, as the three mental items are concatenated into a single whole. As we just saw, the syncategorematic copulative act is described in Quodlibeta VI as the 'union' of the subject and the predicate. This means that whenever such a copula accompanies a subject-term and a predicate-term within a given mental complex act, a mental proposition is ipso facto constituted. No additional binding act is required, except the mereological concatenation itself. 29 This does not distort the usual Ockhamistic parallelism between mental and spoken or written propositions, as Spade thinks it does,30 since the mental subject term, predicate term and copula are still seen as three distinct 'absolute' units, just as the corresponding spoken or written words are. 31 Spade's argument is that if the copula is what binds the extremes together, then it is not a 'separable ingredient', and, consequently, not a distinct component either. But this is an inference Ockham would reject,since he both holds that the copula is a distinct component and that its addition to the other two suffices to bind them together into a proposition, just likethe addition of the written word 'est' to 'homo' and'animal' within a single sequence suffices for turning the resulting whole, 'homo animal est', into a written sentence. There is an interesting difference between the two cases, admittedly, since the order is relevant in the written sequence while it is not supposed to be in the mental corresponding concatenation. 32 But this does not substantially affect the present point: the mental proposition is composed out of distinct units just as the spoken or written sentence is, and one of them corresponds to the copula. The same thing, moreover, must hold mutatis mutandis for the other syncategorematic logical acts such as negation, conjunction, quantification, and so on. Even though a mental act of negation, for example, is not normally accomplished alone in the natural order, it suffices, when it is added to a bundle of categorematic concepts plus copula, for turning the whole group into a negative mental proposition, of which it is itself a distinct simple component. 33 While clearly realizing that such was Ockham's position -at least with respect to thecopula,34 Marilyn Adams raised interesting objections against it in the guise of two counterexamples, and it turns out to be quite instructive to see just how the doctrine presented here can handle them. The key to it is to take seriously the requirement that the relevant mental units should be concatenated into wholes for propositional acts to occur. Adams's first counterexample is a situation in which 'I think the proposition "an oak is a tree" and think the concepts of man and animal, without thinking the proposition "a man is an animal"..'35 In this case, the categorematic concepts 'man' and 'animal' are both present inthe mind along with a copula (the 'is'of 'an oak is a tree'), but the mental proposition 'man is an animal' is not formed for all that; therefore, Adams correctly concludes, the merecopresence of the copula with two categorematic concepts is not sufficient for the corresponding mental proposition to be formed. What is relevant in the example is that the copula in question is not mentally concatenated with the categorematic concepts of 'man' and 'animal' into a single mental unit. As we saw in chapter 2, the mereological grouping of mental items into distinctive wholes must be a crucial process in Ockham~s a.ccbunt of how complex mental propositions are structured.36 'Omnis homo est animal' and'omnis animal est homo', for example, differ from each other in having different parts: 'omhis homo' occurs in the former and 'omnis animal' in the latter. Adams is right that the mere copresence of certain units within the mind is not sufficient for the corresponding complex item to be there as well. Some sort of binding act is required. But the mental groqping of parts into wholes is enough to do the job in Ockham's view. The principle is that whenever two or more simple acts are thus assembled, they form a new complex act, the semantical properties of which depend only on those of the components, and which can in tum occur as a part in some further grouping. 37 Adams's second counterexample, however, can be seen as a challenge to precisely that idea. 38 'A man is an animal' and 'an ahimalis a man', she points out. have exactly the same components, but 'man' is the subject in the former and the predicate in the latter;mereological grouping, therefore, is not sufficient for determining which of the two propositions we end up with when these components are mentally assembled. One way out, apparently, would be to concede that these two propositions, which are semantically equivalent, are in fact indistinguishable in the mind. 39 But Ockham, I suspect, would not be too happy with such an approach since it entails that in mental particular affirmative propositions (such as 'homo est animal'), it is indeterminate which term is the subject-term and which the predicateterm. A more plausible solution within the Ockhamistic framework would be to postulate a special mental syncategorematic term which turns the categorematic concept with which it is immediately grouped. into the subject-term of the proposition in which they both occur (and maybe another special predicate forming

83 154 LOGICAL CONCEPTS 155 syncategorematic functor as well). Ockham nowhere considers such a possibility, but it is mentioned in one of his anonymous successors as an integral part of the doctrine that mental propositions are complex units. 40 And it does seem to be a rather natural elaboration on Ockham's own general approach. It saves at a low cost both the subject-predicate structure for all atomic propositions, and the mereological principle of compositionality identified above. Ockham certainly could have accepted it without any great drawback. The actus-theory thus provides a rather elegant account oflogical concepts: they are functional mental acts of a special sort, that normally do not occur alone in the mind, but that can be combined with categorematic conceptual acts to form complex units endowed with precise semantical roles, that of being a quantified subject or predicate term, for example, or a predicative proposition, or a negative one, or a complex proposition such as a conjunction, and so on. As to the original problem raised in the Ordinatio - where do such logical syncategorematic concepts come from? - there is only one plausible answer to itin the context of Ockham's later theory as well as in the former: these logical acts correspond to innate capacities ofthe mind. Although he is not explicit on the matter, as Beretta stresses,42 Ockham can be shown to be committed to innatism in this case by the following straightforward considerations. Just as about every medieval philosopher does, he often repeats the Aristotelian motto that the specific difference of human beings with respect to other animals is that they are endowed with rationality. This 'potentia' or aptitude - which he squarely identifies with the intellectual soul43 - certainly has to be innate in his eyes, since it constitutes the main distinctive feature by which a member of the human species essentially differs from a horse or a donkey.44 Now, a rational being is defined by Ockham, unsurprisingly, as a 'being or substance that can reason [ratiocinari]';45 and the capacity for reasoning, in his view, does require the capacity for assembling propositions, making negations, drawing inferences, and so on.46 The consequence follows: the basiclogical aptitudes must be innate in rational beings. But syncategorematic logical concepts are nothing but actualizations of such aptitudes in Ockham's theory. The possession of logical concepts, therefore, has to be seen as an innate feature of human beings. While considering this as a 'plausible' option, Calvin Normore thinks that other possibilities are also open to Ockham for explaining the origin of mental syncategoremata.47 The only one he mentions, however, corresponds to what we have identified in chapter 6 as Ockham's account for the implementation of simple connotative concepts,48 and it could hardly apply to logical constants: 'If the objects of intuition are sufficiently complex', Normore says, 'if; for example, one intuits not justsocrates'spaleness or Socrates himselfbut, say, pale Socrates sitting on a fence, then it may be that the syncategorematic elements of the complex are presented in intuition.'49 But since syncategorematic terms do not signify any special external objects, according to Ockham, there simply is no such syncategorematic element in the external intuited complex in such cases. Relational and other connotative terms, by contrast, do have external significates: those are not relations, of course - which have no independent reality according to Ockham - but n-tuples of singular substances or qualities, as we have repeatedly insisted upon in chapter 6. When complex groups of objects are intuited, then, what is implemented in the mind are connotative concepts, not syncafl'gol'l'mafa Prepositions and non-standard copulas Our focus in this chapter up to now has been on logical constants: quantifiers, connectors, the copula 'est', the negation, and so on. These correspond indeed to Ockham's standard examples of syncategorematic terms. 51 But the category in his view also includes a bunch of terms which cannot easily be classified as logical constants in the modern sense. Think of prepositions such as 'in', 'for', 'from', and the like; or of verbs like 'to have' or 'to inhere in', which Ockham regularly treats as copulas. Those might well be the sort of terms Calvin Normore had in mind in the argument just discussed at the end of the previous section; and it must be conceded that the account offered so far does not smoothly extend to such cases. How are we to deal with them, then? Ockhamdidn't say much on the matter. The important thing fothim was to acknowledge the existence of non-referring auxiliary terms (whether spoken, written, or mental), of which he took logical constants to be the more salient instances. 52 Where exactly the limits of that category are to be drawn is not something he cared to investigate in any detail. This does not mean, however, that his system is without resources for facing the question of what concepts in the mind correspond to prepositions and verbs. Two complementary approaches could be explored. Since Ockham doesn't do it himself, I will be content to simply sketch them here, and to outline one important aspect which they share. The first possibility is to consider some spoken or written prepositions such as 'in', 'with' or 'after', and verbs such as 'to have' and 'to inhere', as expressing in fact categorematic relational concepts rather than genuine mental syncategoremata. It is true that all prepositions are cursorily said by Ockham to lack determinate and finite signification.53 But maybe this can be seen as a loose formulation. What matters for him in such cases is that no special 'prepositional' objects be admitted in the ontology as denotations for 'in', 'from', and the like, and this constraint would be satisfied if the corresponding concepts were analysed as having a primary and a secondary signification each pertaining to nothing but singular substances and qualities. The relational concept corresponding to 'in' in the spatial sense, for instance, couldbe said to primarily signifyevery substanceorquality which is inside something, and to secondarily signify (or connote) every substance or quality which is such that something is inside it, just as 'taller' is supposed to primarily signify every thing which is taller than something else, and to connote every thing which is such that something else is taller than it. This comes even more naturally for a verb such as 'to inhere' ('inhaerere' or 'inesse'), which is used to connect qualities with substances (as in 'albedo inest Sorti': 'a whiteness inheres in Socrates'), and which Spade, for one, has suggested at one point to treat as a special syncategorematic copula within Ockhamism. 54 Ockham himself, in an important passage, puts the cognition corresponding to this particular verb on a par with any other intuitively ascertainable relational concept. 55 Why shouldn't he? It is true, as Adams points out, that he had previously rejected such a connotative analysis of 'to inhere',56 but this was when he still accepted inherences as special relational objects in the world, a position he abandoned very soon in his career. 57 As he reduced ontology to substances and qualities, he came to see relational tenns in general as connotative, and there is no obvious reason why this should not hold good for the concept of inherence. That Ockham sometimes

84 156 LOGICAL CONCEPTS 157 treats 'inest' as a copula 58.is no objection here since he often uses a very general sense of the technical term 'copula', according to which it applies to any relational verb whatsoever. 5?Just as 'to see' can be analysed in mentalese into 'to be seeing', 'to inhere in' in this approach would be analysed into a regular mental copula, 'to be', and a connotative concept'inhering in'. Being~onnotative, this concept 'inhering in' would be liable to a nominal definition, like any other connotative term. Contrary to what Spade and Gaskin believe,60 this would not bring about an infinite regress. 'Inhering in' is an asymmetrical relational term in this' approach, and this means, according to Ockham's view of suc):l terms, that any language where it appears,. including meritalese, contains a converse relational term ~ 'inhered in', let's.say - which should occur in obliquo in its correct complete nominal definition, and in the definition of which,in turn, it should occur itself in obliquo. 61 The complete Ockhamistic nominal definition of 'inhering in' would be something like 'a quality with respect to which a substance is inheredin', and the definition of 'inhered in' would be something like 'a substance with respect to which a quality is inhering in'; just as the definition of 'taller', as we have seen, should be something like 'a body with respect to which a body is shorter', and the definition of 'shorter': 'a body with respect to which a body is taller'.62 As I repeatedly stressed, circularity in such cases is not a problem, given what nominal definitions are intended for in Ockham's system:. What matters is that the two definitions, taken jointly, make it perspicuous that only substances and qualities are needed as primary and secondary significates of these two correlative terms. Circularity, on the other hand, efficiently prevents infinite regress 'in this case. Spade and Gaskin think otherwise because they mistakenly take the Ockhamistic programme to aim at the ultimate elimination of all connotative terms from the language of thought through nominal definitions. A' second fruitful approach for the treatment of such prepositions and verbs within Ockhamism would be to follow an intriguing suggestion, from Paul Spade, and to postulate primitive mental syncategoremata corresponding to certain special terms like 'inhering', 'having', 'doing', and a few others. 63 The idea is found in Buridan, for example, who rightly noticed that the nominalistic theory of connotation requires the acceptance of several distinct modes of connotation. 'Father', presumably, does not connote the children in the same way 'horseman' connotes the horses or 'white' connotes the whitenesses. If it did, the nominal definitions of these connotative terms could not distinctively spell out how their, primarysignificates should beconnected in reality with theirconnotata, as required by what I called condition C3 for a good nominal definition. 64 Buridan spoke of distinct 'modes of adjacence' in this respect (modi adiacentiae);65 he conjectured that to every such mode there corresponds a distinct mode of predication in the language of thought, and a distinct' Aristotelian category of simple accidental categorematic terms. Spade appropriately quotes on the matter the following passage from Buridan's Sophismata: Third, note that in accordance with the different positive ways of joining connoted things lathe things to which the terms refer, there arise the different modes of predkating. For example, how, how much, when, where, how this is related to that, and so on. From these different modes of predicating, the different categories arc taken."" As Spade remarks, 'there is no evidence that Ockham himself would have accepted [this] development ofhis theory ofconnotation'.67 But something like this does seem to be called for, as I said earlier, if a variety of modes of connotation is to be posited in the language of thought, as Ockham is committed to. His treatment of oblique cases (genitive, accusative, and the like) turns ouno be directly relevant here. For one thing, he clearly admits a variety of them in the language of thought,68 and this, as we have seen, simply amounts to accepting different sorts of grammatical complements. 69 It seems natural to suppose -although Ockham does not say so - that each such distinctive sort of complement is associated in mental sentences with a distinctive syncategorematic indicator (corresponding, for instance, to 'in', 'with', 'on', or maybe 'having', and so on). This grammatical variety, on the other hand, is connected with the plurality of possible modes of connotation by the theory ofnominal definitions: what I have called condition C3 for such definitions in chapter 5 requires in effect that the precise grammatical functions of the terms in obliquo within a nominal definition (as revealed by their grammatical cases and/or by the prepositions or the copulas which govern them) should indicate the modes under which the secondary significates of the defined term are connoted.70 Now, ifthere are simple connotative concepts, as Ockham admits, and if-as he also holds ~every connotative term has one, and only one, correct nominal definition,7\ it follows that there must be in the language of thought one primitive and distinctive way of being a grammatical complement for each primitive mode of connotation. And if each sort of grammatical complement is associated with a distinctive syncategorematic indicator, there should be at least one distinctive syncategorematic concept in the mind for each distinctive mode ofconnotation. Since terms in obliquo can occur as predicates or even subjects according to Ockham,72 these should also correspond to different modes of predication. Which brings us very close to the Buridanian view mentioned by Spade. Letus take an example. Ockham refers at one point in the Summa Logicae to a grammatical situation where 'the oblique case is governed by the force of possession',73 as in the Latin sentence 'aliquis asinus est Sortis' (literally: 'some donkey is Socrates's'), where the genitive 'Sortis' indicates a possession connection. Since Ockham speaks there of the oblique case being 'governed' [regitur] by the 'force of possession' [ex vi possessionisj, it is permitted to extrapolate that the term 'Socrates' in the corresponding mental sentence should indeed be grammatically governed by a special copula (such as 'is had by') or a special possession preposition (such as the postposited 's' in English or the preposition 'd' in the French corresponding sentence 'un fine est d Socrate'). Of course, it is not clear in this particular case that the possession connection in question does correspond to a primitive mode of mental connotation for Ockham (although it might), but the point here is that such a distinctive copula or preposition should be innately available for each variety of connotation that the human mind is spontaneously capable of. whatever these varieties are. These two approaches to prepositions and copulas - the relational one and the grammatical one - could indeed be combined within Ockhamism. It is possible to suppose hath that the human mind is equipped with a limited numher of innate capacities for different sorts of complementation and predication (as in Buridan's suggestion), and that. nevertheless. all other spoken or written prepositions and

85 158 LOGICAL CONCEPTS 159 copulas are represented in the human mind as categorematic relational concepts. As to which is which, given the complexity of the problem and its ultimately empirical character, it is not surprising that no solution to it is even sketched in Ockham. Seriously trying to devise a list ofthe basic modes ofmental connotation would most probably have led to an in-depth revision of the Aristotelian theory of categories, a revolutionary move in the context of the early fourteenth century, which even Ockham might not have felt ready to make, had he been tempted to. 74 Whichever way we choose for elaborating it, however, it is clear that the Ockhamistic approach to prepositions and copulas should ultimately involve some further innate mental apparatus in addition to purely logical capacities. Whether they are syncategorernaticor not, not all the required concepts could plausibly be acquired on the basis of perceptual experience without some sort of antecedently available aptitudes. As we have previously seen, the human mind, in Ockham's picture, must be able to judge that certain things are ordered in certain ways with respect to others,75 and while some of these orderings might be perceptually recognizable, this cannot be generalized. The tendency to associate perceptible qualities with underlying substances through inherence-judgements is a salient case in point: in order to intuitively recognize an inherence connection in the first place, one must, apparently, prepossess, in one form or another, the concept of 'inhering in'. And special dispositions must also be supposed for causal judgements, mereological judgements, and maybe a number of others as well. The theory, consequently, requires the postulation, in one way or another, of an innate array of categories, which Ockham - unsurprisingly - never fully explored. This is not to say that he was unconsciously anticipating some sort of Kantian idealism with an a priori structured mind on the one side and an ultimately uncognizable world in itself on the other side. Ockham clearly admits that although orders are not absolute things in themselves, the ordering capacities of the mind adequately correspond in principle to the organization of external reality. This is not as nai've as it may sollnd: how, otherwise, could human beings survive in this world? Whether in. the fictum- or in the actus-theory of concepts, the human mind, in conclusion, is endowed with a rich innate set of aptitudes. Some are called for by its very rationality, in order to account for logical operations such as predication, quantification, negation, conjunction, etc. When actualized in mental discourse, they correspond to syncategorematic conceptual acts. Others, in addition, are required to account for the various sorts of basic relational judgements that the mind is capable of. It is left open in Ockham's writings whether these should be thought of as syncategorematic copulas and prepositions or as special innate connotative concepts, the more promising approach probably being a combination of these two possibilities. Notes 1. See SL I, 4, OPh I, p. 15: 'Adhuc aliter dividitur terminus, tam vocalis quam mentalis, quia terminorum quidam sunt categorematici, quidam syncategorematici' (the italics are mine). 2. Ockham's paradigmatic examples of sy"cale!-iol'/,liiallj in SL I. 4, OF" I, p. 15 arc 'omllis'. ' ", 'alil/llis'. ' ", '!,ral'lel", 'Iallllllll' and 'illl/iilii/iiii11'. A lillie further down the same chapter, he makes it clear that all conjunctions and prepositions are to be treated in the same way (p. 16). Elsewhere, the negative sign 'non' is also given as an example of syncategorematic term (see SL III-3, 7, OPh I, p. 613). As to the copula, see Quodl. VI, 29, OTh IX, p. 695 (the passage is quoted above in chap. 5, n. 28, where other references are also given). The case of adverbs is somewhat special, since some of them are considered as partially categorematical (see SL I, 4, OPh I, p. 16: 'De quibusdam autem adverbiis aliter est, quia quaedam eorum determinate significant illa quae significant nomina categorematica, quamvis alio modo significandi important'); 'courageously' would be a good example of this: it obviously includes some sort of categorematical reference to courage. SL I, 4, OPh I, p. 15: 'Termini autem syncategorematici... non habent finitam significationem et certam, nec significant aliquas res distinctas a rebus significatis per categoremata, immo sicut in algorismo cifra per seposita nihil significat, sed addita alteri figurae facit earn significare, ita syncategorema proprie loquendo nihil significat, sed magis additum alteri facit ipsum aliquid significare sive facit ipsum pro aliquo vel aliquibus modo determinato supponere vel aliud officium circa categorema exercet' (transl. Loux 1974, p. 55). Wittgenstein 1921, (transl. Pears and McGuinness 1961, p. 22). Ord. I, dist. 2, q. 8, OTh II, p.. 282: 'Quartum dubium est de conceptibus syncategorematicis et connotativis et negativis: unde possunt sumi vel abstrahi? Quia si praecise a rebus, non videtur quomodo possunt distingui ab aliis conceptibus.quod autem sint tales conceptus patet, quia omni propositioni in voce potest correspondere consimilis in mente, igitur isti propositioni "omnis homo est animal" et isti "aliquis homo est animal" correspondent distinctae propositiones in mente; igitur aliquid correspondet signo in una propositione quod non correspondet in alia.' Ibid., pp : '[a] Ad quartum dubium dicerent [or rather dico according to ms. A and F] quod conceptus syncategorematici et connotativi et negativi non suntconceptus abstracti a rebus ex sua natura supponentes pro rebus vel ipsas modo distincto ab aliis conceptibus significantes. [b] Et ideo dicerent [or rather dico according to ms. A, B, C and F] quod nullus [est according to ms. C, D, E and F] conceptus syncategorematicus nec connotativus nec negativus, nisi tantum ex institutione... [c] Possunt autem tales conceptus imponi vel conceptus abstrahi a vocibus, et ita fit de facto vel semper vel communiter. Verbi gratia, isti voci "homo" competit talis modus grammaticalis quod est singularis numeri, nominativicasus, masculini generis, et sic de ahis; et isti voci "hominis" competunt alii modi grammaticales. Similiter, isti voci "homo" competit quod significet determinate rem per se; isti voci "omnis" non sic competit, sed quod significet tantum cum alio. Similiter est de ista voce "non" et de istis "per se" et "in quantum", "si" et huiusmodi syncategorematicis. [d] Tunc ab istis vocibus sic significantibus abstrahit intellectus conceptus communes praedicabiles de eis, et imponit istos conceptus ad significandum illa eadem quae significant ipsae voces extra. [e] Et eodem modo et de talibus format propositiones consimiles et habentes consimiles proprietates quales habent propositiones prolatae. [f] Et sicut potest instituere tales conceptus ad sic significandum; ita potest instituere ipsos conceptus abstractos a rebus ad significandum sub eisdem modis grammaticalibus sub quibus significant ipsae voces. [g] Hoc tamen fit convenientius per conceptus abstractos a vocibus propter aequivocationem vitandam, quia illi conceptus sunt distincti sicut ipsae voces, quamvis non omnes sint distinctae; conceptus autem alii non sunt distincti. [hi Et ita quaelibet talis propositio esset distinguenda, puta propositio correspondens tali propositioni "homo est homines", "homo est homini", et sic de aliis.' In sentences [a] and lbl, it seems prohahle that Ockham originally wrote 'dico' at a time when he was still suhsrrihing tolhe.!icllllii-theory, and laler changed to 'dicerell/' as he moved towards lhe lu III.\ -lheory. Thai even then, he uses 'dic(,/,/'ii/' ('Ihey would say') ralher Ihan 'dinll//'

86 ('they say') shows that he did not find this doctrine in other writers, but devised it himself. In sentence [b], on the other hand, the insertion of 'est', as found in some good manuscripts, makes more sense. See above chap. 6, sect. 2. Normore 1990, p. 59. Ibid., p. 60. Adams 1987, p Adams provides a presentation of this particular theory on pp. 298~302. Fot other short discussions of it, see Gelber 1984, pp , Karger 1996, pp , and Maieru 2002, pp. 5ss. See Adam de Wodeham; Lectura Secunda in I Sent., ed. Wood and Gal 1990, vol. I, pp See Karger 1996, pp Rep. II, q. 1, OTh V, pp See,for instance, p.. 20: '... per copulam importatur conceptus absolutus qui habet tantum esse obiective in anima...' Karger 1996, p See A. de Wodeham, Lect. Secunda, ed.wood and Gal 1990, vol. I, p Karger 1996, p. 223 n. 96. See Quodl. III, q. 8, OTh IX, pp ; the relevant passages are quoted above iqchap. 2, nn See above chap. 2, sect See Quodl. III, q. 8, OTh IX, p. 234: '... et iste actus aequivalet quantum ad multa alicui complexo quo aliquid scitur'(italics are mine). Ockham's. use of 'aequivalere' in this context refers us back to his hypothesis that some mental propositions are actually simple acts that are equivalent in some respects with complex sequences. See the passages from Exp. in Perih., Prooemium, and Quaest. in Phys., q. 6 quoted above in chap. 2, nn. 83-6, as well as the discussion I propose of these in sect. 4.2 of the same chapter. As argued earlier, however, Ockham admits, even at this stage, that some relational concepts cart be acquired by abstraction. It must be concluded that he probably uses the word 'connotative' in this passage in its narrow sense, which does not include relational terms (see above chap. 6, nn ). More on reverse subordination in chap. 9, sect. 2 below. See Beretta 1999, p The author refers to Quodl. IV, q. 35, OTh IX, p. 471, where Ockham does indeed speak of 'signa mentalia ad placitum signijicantia... puta syn(:ategoremata mentalia'. This conjecture tends to be supported by the fact that there is no such reference to conventional mental syncategoremata in the otherwise parallel passages of the Summa Logicae, where Ockham defines the wide senses of 'nomen secundae intentionis' (SL I, 11, OPh I, p. 40, lines 57-60) and 'intentio secunda' (SL I, 12, OPh I, p. 44, lines 75-7). Ord. I,'dist. 2, quest. 8, OTh II, pp : '... sicut voces et signa voluntarie instituta.., quaedam sunt quae non significant sed tantum consigrtificant cum aliis, cuiusmodi sunt syncategoremata..., ita sunt quaedam qualitates exsistentes in mente subiective, quibus ex natura competunt talia proportionabiliter - qualia competunt vocibus per voluntariam institutionem' (italics are mine). The addition to the Ordinatio from which these lines are drawn is placed by the editors under the suitably descriptive heading: 'Opinio posterior auctoris' (p. 289). Remember, however, that since mental propositions, for Ockham, are not extended in space or in time, they cannot be thought of as displaying a simple linear sequential order as spoken and written words do. See my remarks on this in chap. 2, sect. 4.2 above. Quodl. VI, q. 29, OTh IX, p. 695: '... dico quod unio extremorum propositionis in mente est conceptus syncategorematicus verhi copulativi sive copulantis suhiectum cum praedicato... quod est coi1l:eptus copulae, quac esl qualitas quaedam mcntis, puta actus LOGICAL CONCEPTS 161 intelligendi. Et iste conceptus distinguitur realiter a subiecto et praedicato, quae sunt etiam diversi actus intelligendi' (transi. Freddoso 1991, pp ). See SL I, 12, OPh I, pp Second intentions are defined there as those concepts that signify first intentions; and. first intentions (in the wide sense) are said to include mental syncategoremata (p. 43, lines 55-7). Spade 1996, p Spade 1996, p. 134, singles out a similar answer as being 'perhaps' Buridan's (his reference is to Buridan's Sophismata I, ed. Scott 1977, p. 32: '... haec dictio "est", prout praecise est copula, nihil significat ad extra ultra significationem terminorum cathegorematicorum, sed significat solum illum conceptum complexivum quo intellectus format propositiones ex. istis terminis... '): but whether it is Buridan's or not, a version of it is clearly found in Ockham. See Quodl. VI, q. 29, OTh IX, p. 696: 'Potest enim aliquis absolute cognitione incomplexa intelligere hominem etanimll1, et tamen nec homo erit subiectum nec animal praedicatum. Et hoc, quia deficit iste conceptus syncategorematicus "est", quo po$ito, sine omni alia respectu statim homo erit subiectum et animal praedicatum, et habetur tota propositio' (italics are mine). The italicized part of this quotation plainly states that the addition of the mental copula to a pair of categorematic concepts suffices to bring about a mental proposition. See Spade 1996, pp See Quodl. VI, q. 29, OT,h IX, p. 695: 'Sicut in ista propositione prolata "homo est animal" unio extremorum ponitur istud verbum ~'est", quod est quaedam qualitas absoluta, ita proportionaliter est de unione istorum extremorum in mente, quod est conceptus' copulae, quae est qualitas quaedam mentis, puta actus intelligendi. Et iste conceptus distinguitur realiter a $ubiecto et praedicato, quae suntetiam diversi actus intelligendi' (italics are mine). Seen. 25 above, as well as chap. 2, sect Note that since syncategorematic acts such as the copula or the negation are 'absolute' qualities, according to Ockham, he ~s committed to accept thatgod at leastcould make these occur alone.if He so wished. But this is no more surprising than the possible supernatural occurrence of a whiteness without a substratum or, for that matter, of a categorematic conceptual act without an underlying thinking mind. See Adams 1987, pp Adams bases her interpretation on a passage from the Reportatio where Ockham firmly states that the copula is a distinct absolute unit rather than a relational one (Rep. II,I, OTh V, pp ), but this is, as we saw, a thesis he still maintains in Quodl. VI, q. 29, while adapting it to the actus-theory. Adams 1987, p See above chap. 2, sect It follows that contrary to what we find in Nelson Goodman's mereology (see, for example, Goodman 1956), the identity of the ultimate parts are not sufficient here for determining that of the resulting whole, since 'omnis homo est animal' and 'omnis animal est homo' do have the same four ultimate components. Mereological grouping, then, has to be granted an arborescent structure in the Ockhamistic approach. Adams 1987, p In modern logic, both would be expressed by the same formula: '(::Ii:) x is a man /\.r is an animal'. And Ockham admits, as a matter of fact, that in 'homo est animal' hoth extremes have the same mode of supposition (see SL I, 70, OPh I, p. 210: '... in ista propositione "homo est animal" utrumque extremum habet suppositionem determinatam '). The text is quoted hy Spade 1996, PI'. 12R-9. His reference is to Vienna, Nationalhihliothek, Pal. Lal. MS 4X5.1, 1'01. Imr. Sec ahovc chap. 2, n. XX. Berctta IW'J, pp,

87 162 LOGICAL CONCEPTS See, for example, Exp. in Porph. 1,5, OPh II, p. 23: 'Unde si "rationale" est differentia hominis, "rationalitas" importabit idem quod "anima intellectiva", et rationalitas erit anima intellectiva...' 44. See Ord. I, dist. 8, q. 4, OTh III, p. 233: 'Haec enim est prius vera "omrris homo per rationalitatem vel per animam intellectivam differt specie ab asino"... ' Or Exp. in Porph. 3, 1, OPh II, p. 55: 'Sed differentia magis proprie dicta est illa qua aliquid differt ab alio differentia specifica, sicut homo differt ab equo per rationale.' 45. Exp. in Praedic. 9, 3, OPh II, p. 187: 'Sicut si definiatur rationale, quod est ens vel substantia quae potest ratiocinari...' 46. Even angels, in Ockham's view, in so far as they are intelligent beings, can form complex propositions and draw inferences; see Rep. II, q. 14, OTh V, pp , esp Normore 1990, pp See above chap. 6, sect Normore 1990, p Normore ultimately closes this development by raising a doubt about Ockham's acceptance of such complex intuitive cognitions (Normore 1990, pp ), but the texts q:uoted above in chap. 6, nn. 14~15, are quite conclusive in this respect. 51. See above n. 2:. 52. Taking logical constants as paradigms for syncategoremata was the standard attitude of medieval logicians. Kretzmann (1982a, p. 213, n. 10) identifies the following logical features as constituting the domain of syncategorematic terms in the thirteenth and fourteenth centuries: 'Distribution (or quantification), e.g., "omnis", "totus"; Nega~ion, e.g., "non", "nihil"; Exclusion, e.g., "solus", "tantum"; Exception, e.g., "praeter", "nisi"; Composition (or predication), e.g., "est", "incipit"; Modality, e.g., "necessario", "contingenter";conditionality, e.g., "si", "quin"; Copulation (or conjunction), e.g., "et"; Disjunction, e.g., "vel", "utrum"; Comparison, e.g., "qua"; Reduplication, e.g., "inquantum", "secundum quod".' 53. See SL I, 4, OPh I, p. 16: 'Et sicut hoc nomen "omnis" nihil determinate et finite significat..., sic est de omnibus syncategorematibus et universaliter de coniunctionibus et praepositionibus.' 54. Spade 1990, p See also Panaccio 1992a, pp for a related idea. The suggestion is critically discussed in Michon 1994, pp , and Gaskin 2001, pp See Ord. I, Prol., q. 1, OTh I, p. 31. The passage is quoted above in chap. 6, n See Adams 1987, pp , n The passage she refers to is found in Rep. II, q. 1, OTh V, pp For Ockham's final position concerning the ontological status of relations in the natural world, see in particular Ord. I, dist. 30, q. 2-4, OTh IV, pp. 32~74, and Quodl. VI, q. 8-30, OTh IX, pp See for example SL II, OPh I, p In SL I, 31, OPh I, p. 94, Ockham provides the following definition: 'Copula autem vocatur verbum copulans praedicatum cum subiecto.' A 'predicate' in this way of speaking can be any grammatical complement, and the copula, consequently can be just about any verb. See for example SL II, 21, OPh I, p. 321, where 'asinum' is treated as the predicate of 'nullus homo videt asinum', and 'videt' as the copula. 60. See Spade 1990,p. 606, and Gaskin 2001, p This is what I called Thesis 1 concerning nominal definitions in chap. 5 above. See, in particular, the passages quoted in chap. 4, nn. 28, 30, and See above chap. 6, sect Spade 1990, pp See above chap. 5, sect See for example John Buridan, Summulae de Suppositiollihu,I' 5, cd. Van der Lecq )l)l)8. pp ; in partie. p. 80: 'Sed omnis lerminus connolans aliud atl co pro quo supponit dicitur appellativus; et appellat illud quod connotat per modum adiacentis ei pro quo supponit' (the italics are mine). 66. John Buridan, Sophismata 4, ed. Scott 1977, p. 62: 'Tertio notandum est quod secundum diversus modos positivos adiacentiae rerum appellatarum ad res pro quibus termini supponunt, proveniunt diversi modi praedicandi, ut in quale, in quantum, in quando, in ubi, in quomodo hoc se habet hoc ad illud, etc. Ex quibus diversis modis praedicandi, sumuntur diversa praedicamenta, prout debet videri supra librum Praedicamentorum.' The passage is quoted by Spade 1990, p. 606; the English translation I use is his. 67. Spade 1990, p See Quodl. V, q. 8, OTh IX, pp : 'Prima [conclusio] est quod accidentia communia nominum vocalium conveniunt nominibus mentalibus... Nam primum accidens commune est casus... Et manifestum est quod sicut istae propositiones in voce "homo est homo", "homo est hominis" habent distincta praedicata variata solum per casus..., sic istae propositiones in mente propositionibus vocalibus correspondentes "Sortes est homo", "Sortes est hominis" habent distincta praedicata variata solum per casus.' See also SL I, 3, OPh I, pp See above chap. 5, sect See chap. 5, sect This is what was labelled Thesis 3 about nominal definitions in chap. 5, sect. 1. See in particular the passage from Quodl. V, q. 19, OTh IX, p. 555 quoted there in n See for example SL II, 8, OPh I, p. 272, about propositions with an oblique subject or predicate. 73. Ibid.:'... quando casus obliquus regitur ex vi possessionis'. 74. Note that Ockham, contrary to many othermedieval authors, never tried to provide a full justification for the Aristotelian list of categories. See Exp. in Praedic. 7, 1, OPh II, p. 161: 'Quinto, circa sufficientiam et numerum praedicamentorum est sciendum quod difficile est probare quod sint tantum decem praedicamenta.' In SL I, 41, OPh I, pp , and Quodl. V, q. 22, OTh IX, pp , he links the diversity of the categories with the number of distinct questions which can be answered with one single term; but he gives no argument to prove the correctness of the Aristotelian list. Buridan, on the other hand, goes so far as to say that no such argument is found in Aristotle himself; see his Summulae in Praedicamenta 3, ed. Bos 1994, p. 18: 'Et sciendum est quod numquam Aristoteles posuit rationem ad ostendendum quod non essent alia praedicamenta praeter ista decem. Nec esset inconveniens ponere alia... ' 75. See above chap. 5, sect. 3.3.

88 Chapter 9 The Meaning of Words Everybody admitted in the Middle Ages that spoken and written words receive their meaning by being associated with concepts somehow. But how exactly is this dependency to be theorized came to be the subject of a lively debate.:.., a. 'magna altercatio', Scotus says -towards the end of the thirteenth century.! Ockhatn in this discussion resolutely sided with those who held that words are not properly said to signify concepts, butthings. Yet he did not believe, any more than anybody else, that words. signify things independently of concepts: words, he would S;lY, are 'subordinated' to concepts. How he thinks of this relation turns out, as we shall see, to be of far-reaching significance for his whole philosophy of mind and language, but still, I ;lm afraid, widely misunderstood. My intent in this final chapter is to elucidate Ockham's notion of subordination. It will be shown, in particular, to require a resolutely externalistic conception of the meanings ofwords, of the sort Hilary Putnam's name has been associated with in contemporary philosophy (section 1). How subordination thus understood fits with Ockham's nominalism will then be considered (section 2). And it will be argued, finally, that Ockham's best theory on this matter opens the way for the acceptance of what I will call 'reverse subordination': a direct semantlcal dependency of certain mental units upon their linguistic counterparts. The case of singular terms, at this point, will come out as especiallyrevealing (section 3). 1. Subordination The basic scheme is the following. First, simple categorematic concepts are acquired as natural signs of external things. And then comes subordination: certain spoken sounds are conventionally associated with certain concepts, in such a way that the spoken sounds in question inherit the signification of the concepts they are associated with. The point is made in the very first chapter of the Summa Lo!?icae: I say that spoken words are signs subordinated to concepts or intentions of the soul not because in the strict sense of 'signify' they always signify the concepts of the soul primarily and properly. The point is rather that spoken words are i!dposed to signify the very things that are signified by concepts of the mind, so that a concept primarily and naturally signifies something and a spoken word signifies the same thing secondarily... The same sort of relation I have claimed to hold between spoken words and impressions or intentions or concepts holds between written words and spoken sounds. 2 The concept of 'horse', for example, naturally signifies horses. When the spoken sound 'horse' - or 'CqlillS' or '('/i('\'al' - is conventionally suoordinated to thai 1M

89 166 THE MEANING OF WORDS 167 concept, it ipso facto starts signifying horses too, albeit conventionally. And when the corresponding written word is conventionally subordinated in turn to that spoken sound, it also ipso facto starts to conventionally signify horses. Ockham's source for this way of speaking is probably John Duns Scotus in the very passage where he refers to the magna altercatio about whether words signify concepts or things: Let me concede, for short, that what is properly signified by a spoken word is a thing. Nevertheless, written words, spoken words, and concepts, are ordered signs [signa ordinata] of the same significate, just as there are many ordered effects of a same cause, none of which being the cause of the other, as is clear in the case of the sun illuminating several parts of the environment} Ockham expresses himself in a closely related way in his own Ordinatio:...a certain spoken word primarily signifies several things equally, because it has been imposed by a single imposition to everything which a determinate concept of the impositor is common to, so that the word and the concept are to each other like ordered signs [signa quasi ordinata]; not because the word primarily signifies the concept, but because it is imposed at primarily and precisely signifying every single thing the concept is true of... 4 What we have here is a transitional phase, with respect to vocabulary, between Scotus's phrase 'signa ordinata' and Ockham's later use of 'subordinata' in the Summa Logicae. But what is the advantage of saying that words are subordinated to concepts, rather than signify them? For Ockham at least, this is not just an innocuous variation in the way of speaking. Having decided to count concepts as signs, Ockham clearly realized that it was preferable for him to use some other verb than 'to signify' for the connection of words to concepts. For this relation, from a semantical point of view, is very different from the relation that concepts have with the things they signify. In Ockham's framework, to say that a concept signifies certain things in the world is to say that this concept will normally be used by the agent to refer to those very things: when taken in personal supposition (which is the standard case), the concept is expected to supposit for its sign~ficata. It is exactly the same sense of 'to signify' that Ockham wants to transpose to spoken and written words: just like concepts, they will be said to signify whatever it is that they are normally usedto refer to when taken in personal supposition. This corresponds, in fact, to the technical notion of 'signum' Ockham endorses in the first chapter of the Summa (and which I lengthily commented upon in chapter 3 above).5 Ockham's distinctive move here is to reserve the technical appellation of 'signification', and the consequent use of 'to signify', for the semantical connection he wants to posit between signs - whether conceptual, spoken or written - and the individual things out there in the world that they normally refer to when they occur within propositions. This allows for fruitful parallelisms between the semantics of mental language and that of spoken and written languages. And it characteristically brings into focus the referential aspects of meaning which Ockham persistently wanted to stress. Yet a hierarchy is still assumed, of course: concepts are taken to he more fundamental, and their signitication, heing natural, is seen as usually prior and determinative with resped to that of words. The term 'subordination' turns out to be judiciously chosen after all. So far so good. This is generally well understood. But there is more to Ockham's idea of subordination, and the really interesting part of it seems to have been missed in much of the recent secondary literature. The picture many commentators give is the following: whenever I speak intelligently, I have some concepts present to my mind and the spoken words I utter to communicate what I mean are subordinated to those concepts. IfI utter, for example, the English sentence 'some horses are black', my token utterance is supposed to be subordinated to a corresponding mental proposition which is actually present to my mind. This picture, however, is not Ockham's. Two preliminary considerations will begin to shake it. First, subordination, as we have seen, holds not only between spoken words and concepts, but also - and in the same sense - between written words and spoken words. How can the received parallelistic picture be extended to this other case? Should we say that whenever I write down something, I should also utter the. corresponding spoken words, or at least have a mental representation of these corresponding spoken words? If so, I would need three simultaneous levels of mental representations for each word I actually write down: a conceptual representation of horses, let's say, a mental representation of the spoken word 'horse', and a mental representation of the written marks I am about to draw. Is this a theory we want to attribute to Ockham? Well, maybe, if we have to. But the fact is he doesn't say anything ofthe sort. And, as we shall see shortly, he is not committed to it either. My second preliminary consideration has to do with equivocity, which is indeed a subject where Ockham saliently resorts to the terminology of 'subordination'. In the Summa Logicae, for example, the very definition of what an equivocal word is, is coined in terms of 'subordination': 'A word is equivocal if, in signifying different things, it is a sign subordinated to several rather than one concept or intention of the soul.'6 This definition should strike us as not easily compatible with the parallelistic picture. If subordination was the relation between a speaker's token utterance and what that speaker has in mind at the moment ofutterance, the strange consequence would ensue that in order to utter an equivocal word, I would need to have the two corresponding concepts in my mind at the moment of utterance. This is normally not what happens. If I say 'I cashed money at the bank yesterday', I might very well have in my mind, in uttering the word 'bank', only the concept of a financial institution, and not that of a shore. Is it to say that 'bank' is not equivocal in such occasions? If so, we could easily avoid equivocity in our external languages simply by mentally sticking to one particular interpretation of some potentially equivocal word. This sounds strange. The very idea of 'equivocity' is typically used for cases where the meanings of the words do differ from what the speaker wants to convey. What these two preliminary considerations jointly lead to, is that the received parallelistic picture of subordination as the relation holding between a given utterance of a speaker and what that speaker has in mind at the moment of utterance, would force us to attrihute to Ockham either strange theses, or imprecisions of speech, or hoth. This is not yet a refutation of the picture, of course, hut it should arouse our suspicion. Let us corne down now 10 the main poin!. In the crucial passage of Summa

90 168 Logicae I, I quoted above, whereockham introduces the technical phrase 'signa subordinata', he explicitly - and thoughtfully -links subordination with the imposition of the term: 'I say that spoken words are signs subordinated to concepts... because... spoken words are imposed [imponuntur] to signify the very things that are signified by concepts of the mind... '7 This is to be taken seriously. '!mponere' is a technical term in medieval philosophy of language. The 'impositio', in Ockham's vocabulary, is the original attribution ofa conventional signification to a given sound or a written mark. The subordination of a word to a concept, then, is what happens at the original moment of the imposition, not at the moment of utterance. Look again at the passage from distinction 2 of the Ordinatio quoted above: a word and a concept, there, are said to be like 'ordered signs' (signa quasi ordinata) because the word is imposed, by means of a single act of imposition, at signifying everything that the concept signifies. 8 Ockham's example is the univocal term 'homo'; the reason he gives why 'homo' is indeed a univocal term overtly has to do with the original intention of whoever imposed that word: 'Such is the word "homo" and this is why it simply is a univocal term; for the impositor [imponens] intended this word "homo" to signify every single thing which a certain determinate mental Goncept is true of... '9 The use of 'imponens' along with the past tense 'intendebat' in these lines clearly refers the reader to some past event of imposition. This is even more explicit in the Summa Logicae as Ockham discusses the same example: Whoever it was that first instituted the use of the term'homo' saw some particular man and cqined the term to signify that man and every substance like him... But even though it signifies indifferently many men, 'man' is not equivocal, for in signifying indifferently many men it is a sign subordinated [signum subordinatum] to just one concept and not many. 10 Subordination, once more,is based upon the original imposition. of the term by whoever it was that first coined it and associated it with a given concepr Ockham's well-considered idea of subordination can thus be captured in the following formula: A given sign S2 is subordinated to a previously signifying term Sl if and only if S2 has been imposed to signify whatever S1 signifies. I I This is a very general notion. And a useful one too. In the light of it, we can readily understand, for example, that a written word can be said to be subordinated to a spoken word in exactly the same sen.se in which a spoken word is subordinated to a concept: in both cases the original impositor of the newly coined sign simply attributed it the signification of some other previously existing sign, whether natural or conventional. And there is no problem either, in this approach, with Ockham's characterization of equivocity in the Summa: 12 a given sign typically gets equivocal by being successively the object of several different impositions. Subordination, again, is seen to hang on to what happened at the original mornent(s) of imposition.. The point seems to have heen missed hy many. Paul Spade. for example. writes THE MEANING OF WORDS the following about Ockham: 'Insofar as a translation is supposed to "express the same thought" as the original, we can say that a statement in one language is correct translation of a statement in another language if the two statements are subordinated to the same mental proposition.'j3 Strictly speaking, this cannot be Ockham'sown notion of subordination. Spoken sentences, in Ockham's theory, are not imposed to signify anything andthey are not, therefore, subordinated to anything in the relevant sense. Ockham; remember, subscribes to some form of semantical atomism. What is attributed a conventional signification at the moment of imposition normally is a simple tetm, and the semantical properties of complex phrases, such as a complete sentenge, are supposed to be systematically derived from those of their simple components, without any new subordination being needed in the process.. We could set out, of course, to introduce a derivative notion of 'subordination' which would be applicable.10 complete sentences. But for one thing, it is not quite obvious how to do it. And we should be aware, moreover, that this would not be subordination in Ockham:s own sense anymore: the relation between a spoken and a mental sentence, in his view, does not normally depend upon a special act of imposition. Another.interesting example of a similar discrepancy with respect toockham's use is found in Calvin Normore's oft-quoted paper on Ockham's mental language: 'Each spoken connotfltive term and its defining expression', Normore writes, 'will be subordinated to the same complex expressionofthe mental.'14 Normore supposes here that every spoken connotative term is strictly synonymous with its own definition for Ockham and that both are represented in the mind ofa competent speaker by a single complex conceptual sequence, which is, of course, the very reading of Ockham we found reasons to reject in chapters 4 to 6 above. What I am concerned with now is.the place attributed to subordination in this interpretative scheme: it does not fit very well, I contend, with Ockham'sown way of characterizing the relation. Even admitting that a simple spoken word could in certain cases be subordinated to a complex conceptual sequence,15 the spoken definition of that word would not in such cases, properly be said to be subordinated, in Ockham~s technical sense, to the same mental expression. Suppose, for the sake of discussion, that 'father' is adequately defined as 'male animal having a child'. And suppose, in addition, as seems probable, that we do have the mental capacity for assembling a conceptual complex sequence of the form: 'male' + 'animal' + 'having' + 'child'. It might be that the person or group who originally struck the word 'father' decided to subordinate it to precisely this complex mental phrase (or some token of it). Butwhatabout the spoken definition? The English phrase 'male animal having a child', for example, will not properly be said to be subordinated, inockham's sense, to the corresponding complex mental phrase. What has been independently subordinated in this case, as in that of complete sentences, is each component of the English expression. The English word 'male' must have been subordinated to the concept 'male', the English word 'animal' to the concept 'animal', and so on. But the complex English phrase as a whole is not subordinated to anything, since it is not (normally) the object of a new imposition. It will turn out, of course, to have some very strong semantical equivalence with the corresponding mental sequence. But, however strong, this relation will not be subordination in Ockharn's sense. This is not mere quihhling. A whole view of language is at stake here. Spade and 169

91 170 THE MEANING OF WORDS 171 Nonnore's use of 'subordination' rests, as far as I can see, upon their common assumption that what a spoken sound is subordinated to is the intellectual item the speaker has in mind when uttering the sound. If this was the case, it would be innocuous, indeed, to say that a complex spoken phrase or sentence is subordinated to the mental sequence the speaker has in mind at the moment of utterance, just as each one of its components is supposed to be subordinated to some conceptual component of this very mental sequence. But this is not Ockham's approach. The important point, let me repeat, is the following: the meanings of words, for Ockham, depend not upon what is going on in the head of the speaker at the moment of utterance, but upon whatthey were originally subordinated to by the impositor. Put in modem tenns: Ockham is a proponent of externalism rather than intemalism with respect to the signification of spoken and written words. A passage that strikingly supports this reading is found in the Third Part of the Summa, when Ockham discusses the use of absolute - that is, non-connotative spoken names such as 'man', 'horse', 'animal', 'flower', 'water', and the like: how does a spoken sound get to be an absolute name? The answer, as one expects, is!bat this happens when the spoken sound in question is conventionally subordinated to a concept which is already an absolute name, albeit a natural one. What I want to draw attention to in this passage is Ockham's explicit acceptance that some of the spoken words I use might depend for their signification upon concepts that other people have:,." since spoken words are conventional, absolute spoken names canbe imposed upon the very same things of which we have- or others have - such absolute concepts'.16 The example he gives is the tenn 'lion'. Someone like myself, says Ockham, who has never met with a lion, cannot have acquired the simple absolute concept of 'lion'. What such a person has in mind when uttering the spoken word 'lion' (in a sentence like 'a lion is an animal', for example) is actually a 'complex mental expression composed of several simple cognitions, none of which is both simple and proper to the lion'. I? Yet this does not prevent the speaker in question from using the tenn 'lion' as an absolute term in her own spoken speech. Ockham's clear suggestion in this passage is that a speaker can legitimately utter a sentence with an absolute spoken name as subject while not having in her own mind the corresponding absolute simple concept. This is possible because, subordination being what it is, the meanings of our own words can depend upon the concepts that other people have. The spoken word 'lion' is an absolute tenn in English, according to Ockham's story, because it was originally subordinated to the absolute concept of 'lion' by people who had that concept. And it does not lose this instituted signification when it is uttered by a speaker like Ockham himself who does not happen to have the relevant simple concept. What detennines the semantical status of a given spoken word within a given linguistic community is the concept which the original impositor (or group of impositors) associated it with, not the concept that any particular speaker happens to have at the moment of utterance. The consequences of this are far-reaching. There will inevitably be, according to this view, frequent discrepancies between the semantical status of the spoken words uttered by a particular speaker, and the concepts this speaker actually has present to his mind at the moment of utterance, even if the speaker is speaking intelligently and literally. Contrast this, for example, with John Locke's well-known account of linguistic meaning in his Essay Concernin[? Human Understandin[?:... words, in their primary or immediate signification, stand for nothing but the ideas in the mind ofhim that uses them... That then which words are the marks of are the ideas of the speaker: nor can anyone apply them as marks, immediately, to anything else but the ideas that he himself hath Locke's radical intemalism in these lines is at complete variance with Ockham's view. on the signification of spoken words as I read it. Let us return again to a passage from Summa Logicae I, 43, quoted earlier: 'Whoever it was that first instituted the use of the term "homo" saw some particular man and coined the tenn to signify that man and every substance like him... '19 To a modern reader, this must inevitably recall Hilary Putnam's canonical presentation of semantic externalism in his famous 1975 paper 'The meaning of "meaning"';20 and it is as remote as can be from Locke's solipsistic view. Ockham goes as far as to say that in some cases even the impositor might not have in his own mind the concept which he chooses to subordinate a given spoken sound to. For subordination to succeed, it is sufficient, Ockham thinks, that the impositor be able to identify the relevant concept in some distinctive way. This does not automatically require that he should possess it himself: Moreover, it is possible for somebody to impose this name a to signify whatever animal will appear to him tomorrow. This being done, the word a distinctly signifies this animal, and it will signify it for all those who are willing to use the word as it was imposed, even though the impositor does not have a distinct intellection of this animal, and maybe will not have one when it appears to him. 21 In such a case, the concept which the impositor intends to subordinate the word a to, is uniquely identified: it is the absolute specific concept that a well-placed observer would spontaneously fonn if he or she was to have a nonnal intuitive cognition of the first animal the impositor will meet with tomorrow. It might happen, as Ockham explicitly acknowledges,.that the impositor, at the moment of coming across the animal in question, should not himselfbe in a position to form that concept (he might not have a good view of the animal, for example, or his attention might be drawn elsewhere, or whatever... ). Yet the relevant absolute concept is entirely detenninate. And so is its signification: it signifies a certain individual animal (the first one that the impositor will come across tomorrow) plus 'every other substance like it';22 and nothing else. The subordinated spoken word, consequently, will inherit this same signification from the very moment of its imposition; this is the signification it will have, from then on, whenever it is uttered by a speaker who accepts to use this particular spoken sound in accordance with this imposition, even if this particular speaker - and everybody else, for that matter - should lack theconcept in question at the moment of utterance. 2. Types and tokens again Our description of the subordination process has been given so far in terms of linguistic and conceptual types rather than tokens. This is easier to do for a start. and

92 172 THE MEANING OF WORDS 173 it does correspond,to Ockham's usual practice. When he writes, for example, that 'a certain spoken word [aliquis vox] signifies several things equally, becauseit has been imposed [etc.]',23 he speaks as ifthe spoken.sound that now signifies something was the very same one that was originally associated with a 'determinate concept' by the impositor. Yet the relevant tokens, of co~rse, can be numerically distinct. Sameness in such cases must be generic or specific identity, not numerical identity.24 As was explained in chapter3, the type idiom is acceptable to Ockham ill semiotical matters in so far as it is reducible in principle to talk about tokens. 25 How such a reduction canbe achieved in the particular case of subordination-theory, is not - unsurprisingly - developed by Ockham,but it turns out to be rather straightforward. Types can enjoy no independent existence in Ockham's ontology. What must have happened, then, ina typical situation of semantical subordination, is that the original impositor associated at the moment of imposition a characteristic token of a certain spoken sound with a certain conceptual token. 26 The latter naturally signified, prior to the impositional act, certain individuals in the world, in virtue of belonging to a mental chain of intellectual acts and habitus causally grounded in the right. way. When the signification of this conceptual token is conventionally transferred by the impositor to the spoken token, the convention put forward is that it should ipso facto be transmitted to an indefinite number of future spoken tokens as well. Which ones exactly? Two conditions, as far as I can see, must be met in the Ockhamistic context for a newly uttered token to inherit the signification attributed to one of its ancestors by the impositor: first, the new token must be phonologically similar to the one that was associated with a certain concept at the time of imposition; and second, the utterer of the new token must agree to use the sign as it was imposed. The latter condition is especially interesting. Ockham clearly points to it in a passage of the Ordinatio I have just.quoted, where he writes that a certain word a will signify a: certain animal 'for all those who are willing to use the word as it was imposed'.27 If a speaker is to use certain spoken sounds as linguistic signs, he or she must accept the authority of some past impositional acts. When and how these.acts occurred, and who the impositors were, is something, of course, that most speakers don't know anything about.ockham himself does not provide any precise description of how such conventions are originally supposed to be implemented, and his general approach is compatible with a wide. variety ofhypotheses on this subject. There is no need, in particular, to suppose that there was only one single impositor in each case, and that the event of imposition always occurred instantaneously. Collective and indirect processes must be at work lilost of the time, obviously, and this is something Ockham would have no problems with; The willingness to use a certain token 'as it was imposed' cannot depend, for a particular speaker, upon her ability to correctly locate or describe the corresponding original act of imposition. What must suffice is an agreement from the speaker to use spoken sounds and written marks in conformity with the impositions that are currently in force within the linguistic community. This accounts for equivocity, in particular. A given utterance of a certain speaker might be equally similar to two past tokens that were independently subordinated to lj,on-equivalent concepts; given the accidents of collective linguistic history, both impositions might simultaneously be in force within a given linguistic community. In such a case, the general willingness of the speaker to use the signs 'as they were imposed' results in her uttering a token with more than one signification, even if she has only one determinate concept in mind at the moment of utterance, and even if she is entirely unaware of the ambiguity of her own speech. Saying, in short, that a certain linguistic type is subordinated to one Or more concepts amounts, ultimately, to say something about a plurality of phonologically similar tokens, uttered under the authority of some determinate past impositions; 3. Reverse. subordination? The instructive case of proper names An interesting feature of Ockham's idea of subordination, as I understand it, is that it is neutral with respect to the sorts of signs it applies to. I have insisted so far on the subordination of spoken sounds to concepts because it is the basic case in Ockham's view, but a newly coined sign could in principle be subordinated, in the Ockhamistic sense, to any well-identified previously existing sign, whethernatural or conventional, and inherit in this way the semanticalfeatures of this previous sign, whether they are or notentirely clear to the impositorherself. One couldlegitimately propose, for example, to create a new general term in English by subordinating an acceptable string of English sounds to a previously existing Latin word, even if the proposer himself has no clear understanding of the Latin term. Ockham's own example of the impositor who subordinates a newly coined word to a concept he doesn't yet possess (and maybe never will) shows that he is committed to go quite far in this direction. The question, of course, is: how far? Should he admit, in particular, the possibility of what I will call 'reverse subordination',.defined as the conventional subordination of a mental concept to a spoken sound or a written mark? Ifso, the signification of many ofourown concepts could turn out to be quite opaque to us. Prima facie, the case ~eems to be excluded by Ockham's insistence that concepts are natural signs. But this means only that they naturally signify certain things in the world. It does not by itself prevent them from stipulatively signifying some other things. Ockham himself had clearly acknowledged this possibility in his first theory of how connotative, negative, and syncategorematic concepts receive their signification. As we saw in some detail in chapter 8, he thought it acceptable at one point to say that a ment(ll sign already endowed with a natural signification can, in addition, be imposed by the intellect 'at signifying the same thing as [some] external spoken words signify'.28 Theconcept, admittedly, would then become ambiguous for the agent, since it would have two different significations: a natural one and a stipulative one. But Ockham, apparently, did not see this as fatal. The case he wanted to stress was when the concepts that arestipulatively reimposed by the agent are those that naturally signify spoken words: my mental concept of the word 'horse', let's say, can be reimposed, by a mere stipulation ofmy intellect, to signify whatever it is that the word 'horse' signifies. Such reimpositions, at the time of the Ordinatio. were considered by Ockham as legitimate and, presumably, fruitful. He must have taken the standard intellectual agent to be able to disentangle somehow the resulting ambiguities when assembling mental propositions. As we have seen, Ockham later abandoned this approach as an account of how all syncategorematic and connotative concepts are acquired. He came to accept some

93 174 THE MEANING OF WORDS 175 natural syncategoremata in mental discourse,29 as well as some naturally implemented simple connotative concepts. 30 Yet, nothing in his later position forces him to renounce the very possibility of those stipulative reimpositions of concepts that he had hypothesized in the Ordinatio. It is true that he never developed the point again in his later works, but his best theory, as I will now argue, readily accommodates it. It will be instructive in this regard to ponder a bit over the paradigmatic case of proper names. As we saw in chapter 1, intuitive acts themselves can serve as singular terms in the language of thought, according to Ockham. 31 My intuitive grasping of this cat in front of me is a simple mental sign, naturally signifying the individual that caused it, and nothing else. 32 It is perfectly possible, then, on the Ockhamistic conception, that a certain spoken sound be subordinated to such an intuitive act by the impositor. This is, presumably, how proper names are supposed to enter spoken language in the first place. I can subordinate the spoken sound a, let's say, to my present intuitive grasping of this ~at in front of me, and the word a from then on will signify this particular cat for all those who are willing to use the word as I imposed it. Whether they ever saw the cat in question or not, the word a in their mouth will normally supposit for this very cat. 33 Now, Ockham also holds, as I insisted in chapter 1, that no abstractive concept can naturally be a simple singular term. While the representational import of an intuitive act depends upon what causes it, that of an abstractive act, or concept, is based on some sort of similitude; and no similitude relation, however we understand it exactly, can discriminate between maximally similar objects. 34 All abstractive acts tum out to be general. If this is so, a speaker who uses a spoken proper name in the absence of the referent will not normally have in her mind at the moment of utterance a simple concept naturally signifying this very individual and nothing else. In uttering the spoken name, she might, instead, have in mind a complex cognition 'composed of several simple ones';35 in other words: a description. A semantical discrepancy transpires once more between what the speaker has in mind at the moment of utterance and the meaning of her words. This is the very same sort of phenomenon we acknowledged earlier with respect to general terms like 'lion'.36 What we now realize is that such discrepancies between thought and language are bound to be widespread in the case of proper names on the Ockhamistic view. This is not innocuous. Different users of the same proper name will inevitably associate it with different mental descriptions, and will be attracted, consequently, to various inferences, nolle of which would be valid in the spoken language. This is a potential source for confusion and misunderstanding. Suppose I associate the English name 'Aristotle' with a given mental description such as 'the author of the Categories'. I might be tempted to infer from a spoken statement of the form 'Aristotle is F' that the author of the Categories is F, which would be both idiosyncratical and logically unwarranted. It is to be supposed, from Ockham's point of view, that an intelligent speaker should normally be capable of avoiding such traps. But how is that to be done? One nice way of achieving it, at least, would be to resort to this reverse subordination process that Ockham had hypothesized himself in the Ordinatio. If a speaker can stipulatively reimpose her conceptual representation of the English word 'Aristotle' so that it signifies for her, from then on, whatever individual the spoken word signifies in virtue of its original imposition, she will be in a position to use proper names in her own mental reasoning with the same semantical and logical features that spoken proper names have. Invalid idiosyncratical inferences would ipso facto lose much of their attractiveness. An ambiguity, admittedly, would thus be introduced in the language of thought. The same mental tokens will naturally signify the English word 'Aristotle' and stipulatively signify at the same time Aristotle himself. But this should be tolerable to Ockham. It is very similar indeed to one specific kind of ambiguity that he does explicitly admit into mental language: ambiguities of supposition namely.3? In a mental sentence such as 'man is a species', the concept of 'man', Ockham contends, can have simple supposition since the predicate is a term of second intention ('species'), and it can have personal supposition as well since personal supposition is always possible according to him. 38 There is a sense, then, in which such amental sentence can be said to be ambiguous: it could mean that a certain real man is a concept (which is false, of course) or that the concept of 'man' is a concept (which is true).39 This is very close to what reverse subordination results in: the mental sentence 'ARISTafLE is a philosopher' (where 'ARISTafLE' is the natural sign for the English word 'Aristotle') might mean, on this view, that a real man, Aristotle himself, is a philosopher (if we favour the stipulative signification) or that a certain English word is a philosopher (if we favour the natural signification). Itis true that the ambiguity now rests on signification rather than supposition as in the case explicitly admitted by Ockham; but considered at the level of complete mental sentences, the parallelism between the two situations seems very strong, and the same disambiguating processes, presumably, could work in both. Ockham is committed to attributing a disambiguating role to some contextual factors or other with respect to suppositional ambiguities. 40 The same factors, whatever they are, should do just as well for a mental sentence like'aristotle is a philosopher'. In short, nothing prevents Ockham's former hypothesis about reverse subordination to be maintained within his mature theory. And there is a job for it there: it would nicely account for our capacity to use mental proper names in our reasoning. Ifthis is so, however, the idea can - and should - be generalized to other categories of terms. If reverse subordination is possible and fruitful in the case of proper names, so will it be for general terms, whether absolute or connotative. An intelligent agent who has never seen a real lion would be able, in this picture, to subordinate her mental representation of the English word 'lion' to a token of that very spoken word, and use it, in her subsequent thought processes, as a simple absolute concept of something she has never seen, lions namely. Even simple connotative concepts could be instituted in the mind in the same way, by being stipulatively subordinated to simple spoken words such as 'father' for example. Let me insist again that this is not explicitly developed in Ockham's later works. But what he says about subordination - for example, in the passages quoted in section 1 of the present chapter - does commit him to a robust form of semantic externalism with respect to the signification of spoken and written words. It is but a small step from there to accepting at least some degree of externalism with respect to the signification of concepts. Ockham might not have thought very deeply about it. but the hypothesis about reverse subordination that he had put forward in the Ordinatio smoothly allows us to take that step without contradicting or threatening

94 176 THE MEANING OF WORDS 177 any of the other main tenets of his later theory. Some of our mental signs would thus come to depend for their (stipulative) signification upon the concepts that other people have naturally acquired. In Representation and Reality, Hilary Putnam lists three assumptions as constituting the basis for the internalistic picture of meaning he wants to break with (and which he traces back to Aristotle): 1. Every word he uses is associated in the mind of the speaker with a certain mental representation. 2. Two words are synonymous (have the same meaning) just in case they are associated with the same mental representation by the speakers who use those words. 3. The mental representation determines what the word refers to, if anything. 4 \ Ockham, we may now conclude, is committed to rejecting this entire budget of theses. In so far, at least, as linguistic meaning is concerned, the picture he favours comes out as strikingly externalistic: l'. Spoken and written words are subordinated to certain previously existing signs (whether natural or conventional), not by each particular speaker, but by the original impositors of the language. 2'. Two words are synonymous if they were originally subordinated to the same concept (or to equivalent conceptual tokens).42 Particular speakers can fail to rec,ognize such synonymies, even ifthey correctly use the words in question in most situations~43 3'. What a word refers to is determined by the signification ofthe sign(s) which it was originally subordinated to. The meaning of our linguistic signs, then, can be opaque to us up to a point. And if, on top of it, reverse subordination is added to the picture, as Ockham had once proposed, externalism extends to thought itself, at least partially: in Ockham's best theory, as it seems, some of our mental representations could be stipulatively endowed with determinate significations, even if we do not happen to possess the corresponding natural concepts. Notes 1. See John Duns Scotus, Ordinatio I, dist. 27, q. 3, in Opera Omnia VI, Vatican edition, 1963, p. 97. A survey of this discussion can be found in Panaccio 1999a, chap SL I, 1 OPh I, pp. 7-8: 'Dico autem voces esse signa subordinata conceptibus seu intentionibus animae, non quia proprie accipiendo hoc vocabulum "signa" ipsae voces semper significant ipsos conceptus animae primo et proprie, sed quia voces imponuntur ad significandum illa eadem quae perconceptus mentis significantur, ita quod conceptus primo naturaliter significat aliquid et secundario vox significat illud idem... Et sicut dictum est de vocibus respectu passionum seu intentionum seu conceptuum, eodem modo proportionaliter, quantum ad hoc, tenendum est de his quae sunt in scripto respectu vocum' (transl. Loux 1974, p. 50, slightly amended). 3. John Duns Scotus, Ord. I, dist. 27, q. 3,in Op. Omnia VI, p. 97: '.., breviter concedo quod iiiud quod signaturper vocem proprie, est res. Sunt tamen signaordinata eiusdem signati.littera, vox et' conceptus, sicut, sunt multi effectus ordinati eiusdem' causae, quorum nullus est causa alterius, ut patet de sole iiiuminante plures partes medii...' (italics are mine). 4. Ord. I, dist. 2, q. 4, OTh II, pp : '... aliquam vocem aeque primo' plura significare, quia una impositione imp0!1itur omnibus quibus conceptus detej;l11inatus 'habitus ab imponente est communis, ita quodsint signa quasi ordinata; non quod vox primo significet ilium conceptum, sed quia imponitur ad significandum primo et praecise omne illud de quo conceptus praedicatur... ' (italics are mine) 5. SLI,l, OPh I, p. 9; text quoted above in chap 3, n SL I, 13 OPh I, p. 45: 'Est autem vox illa. aequivoca quae significans plura non est signum subordinatum uni conceptui, sed est signum unum pluribus conceptibus seu intentionibus animae subordinatum' (trans!. Loux 1994, p. 75). 7. SL I, 1, OPh I, pp. 7-8; text quoted above in n. 2. Loux's translation here unfortunately renders 'imponuntur' by 'are used to' (Loux 1974, p. 50), thus losing the precise technical import of the term. 8. See the text quoted in n Ord. I, dist. 2, q. 4, OTh II, p. 140: 'Talis est haec vox "homo", et ideo est simpliciter univoca, quia imponens hanc vocem intendebat quod significaret omnem rem de qua conceptus mentis determinatus praedicatur...' (italics are mine). 10. SL I, 43, OPh I, p. 124: 'Illeenim qui primo instituit hanc vocem "homo", videns aliquem hominem particularem, instituit hanc vocem ad significandum illum hominem et quamlibet talem substantiam qualis est ille homo... Nontamen esthaec vox "homo" aequivoca, quamvis significet multa aeque primo, quia.est signum, subordinatum uni conceptui et non pluribus in significando illos plures homines aeque primo' (transl. Loux 1974, p. 136, slightly amended; the italics are mine). 11. Literally taken, this formula applies only to categorematic terms, since only these have a determinate signification according to Ockham's way of speaking (SL I, 4, OPh I, p. 15). It can, however, be easily extended to cover syncategoremataas well: a spoken sound can certainly be associated by imposition with a given syncategorematic concept in such a way that it inherits the semantical features of this concept whatever they are (its role in determining the supposition of other terms, for example). 12.' See the text from SL I, 13 quoted above in n Spade 1996, p Normore 1990; pp See above chap. 6, sect SL III-2, 29, OPh I, p. 558: '... ex quo voces sunt ad placitum, voces mere absolutae possunt imponi eisdem de quibus habemus, vel aliihabent, tales conceptus' (italics are mine). 17. Ibid., p. 559 (about the sentence 'leo potest esse animal'): 'Et habeo unam propositionem mentalem cuius subiectum estcompositum ex multis notitiis incomplexis quarum nulla est simplex et propria leoni;. sed propositionem mentalem cuius subiectum sit aliquod mere absolutum proprium leonibus non habeo... ' (italics are mine). 18. John Locke, An Essay Concerning Human Understanding III, 2; SL I, 43, OPh I, p. 124 (text quoted above in n. 10). 20. See for example Putnam 1975b, p. 245: 'We have now seen that the extension of a term is not fixed by a concept that the individual speaker has in his head, and this is true both because extension is, in general, determined socially - there is division of linguistic labor as much as of "real" labor - and because extension is, in part, determined indt xically. The extension of our terms depends upon the actual nature of the particular things that serve as paradigms.'

95 178 THE MEANING OF WORDS Ord. I, dist. 22, q. unica, OTh IV, p. 56: 'Praeterea, potest aliquis imponere hoc nomen a ad significandum quodcumque animal quod occuret sibi cras. Hocfacto, distincte significat iliud animal, et significabit apud omnes vo1entes uti voce sicut imposita est, quantumcumque iliud imponens non distincte intelligat, nec forte distincte intelliget quando sibi occurret.' 22. See the passage from SL I, 43 quoted above in n Ord. I, dist. 2, q. 4, OTh II, pp ; text quoted above in n For Ockham's use of this distinction, see, for example, the passages referred to above in chap. 3, nn. 35-9, and the corresponding discussion in the core of the text. 25. See above chap. 3, sect As we saw, it remains possible for Ockham that the impositor does not himself possess the relevant concept at the moment ofimposition (see the passage from Ord. I, dist. 22 quoted above in n. 21), but he must have, at least, some descriptive way of identifying it, which should not presuppose, of course, the irreducible existence of linguistic or conceptual types. 27. Ord. I, dist. 22, OTh IV, p. 56: 'Hoc facto, [hoc nomen a] distincte significat iliud animal, et significabit apud omnes volentes uti voce sicut imposita est...' (italics are mine; a longer portion of the passage is quoted above in n. 21). 28. Ord. I, dist. 2, q. 8, OTh II, p. 286: '... et [intellectus] imponit istos conceptus ad significandum ilia eadem quae significant ipsae voces extra' (italics are mine; the surrounding passage is quoted above in chap. 8, n. 6 - see sentence d). 29. See above chap. 8, sect See above chap. 4 and See above chap. 1, sect See in particular the passages from Quaest. in Phys., q. 7, OPh VI, p. 411, and Rep. II, q , OTh V, p. 288 quoted above in chap. 1, nn. 41 and We thus arrive at something like the Kripkean picture of how proper names are introduced in the language. See Kripke 1972, p. 302: 'A rough statement of a theory might be the following: An initial baptism takes place. Here the object may be named by ostension, or the reference of the same [sic] may be fixed by a description. When the name is "passed from link to link", the receiver of the name must, I think, intend when he learns it to use it with the same reference as the man from whom he heard it.' 34. See in particular the passages from Quodl. I, q. 13 and V, q. 7 quoted above in chap. 1, nn. 27, 28, and 31, as well as the corresponding discussion in the text. For more on the relevant notion of similitude, see chap. 7, esp. sect. 2 and See Quodl. I, q. 13, OTh IX, p. 77: '... habeo aliquam cognitionem abstractivam propriam, sed ilia non erit simplex sed composita ex simplicibus'. 36. See the passages from SL III-2, 29, OPh I, pp quoted above in nn , and the corresponding discussion in the text. 37. See SL III-4, 4, OPh I, p. 763 (about suppositional equivocations): 'Et est notandum quod iste tertius modus aequivocationis potest reperiri in propositione pure mentali...' That simple and material suppositions are possible in mental language in addition to personal supposition is explicitly acknowledged by Ockham in SL I, 64, OPh I, p. 197: 'Sicut autem talis diversitas suppositionis potest competere termino vocali et scripto, ita etiam potest competere termino mentali, quia intentio potest supponere pro ilio quod significat et pro se ipsa et pro voce et pro scripto' (italics are mine). 38. See SL I, 65, OPh I, pp : 'Notandum est etiam quod semper terminus, in quacumque propositione ponatur, potest habere suppositionem personalem, nisi ex voluntate utentium arctetur ad aliam... Sed terminus non in omni propositione potest habere suppositionem simplicem vel materialem, sed tunc tantum quando terminus talis comparatur alteri extremo quod respicit intentionem animae vel VOCem vel scriptum.' 39. Ibid., p. 198 (about 'homo est species'): 'Et est propositio distinguenda penes tertium modum aequivocationis, eo quod subiectum potest habere suppositionem simplicem vel personalem.' 40. In SL I, 65, in particular, Ockham mentions the 'intention of the users' (voluntas utentium) as a possible disambiguating factor for a sentence like'homo est species' (see the text quoted above in n. 38). Othercontextual elements might also playa role. See on this Panaccio and Perini Santos, forthcoming, esp. sect. 5: 'Le role du contexte.' 41. Putnam 1988, p. 19 (the italics are Putnam's). 42. Note that I do not say 'only if' (or 'just in case', as in Putnam's formulation): being subordinated to the same concept is a sufficient condition for synonymy, in Ockham's view, but not a necessary one. See on this chap. 7, sect. 4.2 above, esp. pp See SL I, 6, OPh I, p. 19: 'Large dicunturilla synonyma quae simpliciter idem significant omnibus modis..., quamvis non omnes utentes credant ipsa idem significare, sed decepti aestiment aliquid significari per unum quod non significatur per reliquum.'

96 Conclusion Concept theory has come to be cruci,al in recent philosophical work on cognitive sciences. Stephen Laurence and Eric Margolis open their well-known 1999 introduction to the state of the discussion,on the topic by squarely asserting: 'Concepts are the most fundamental constructs in theory of the mind.' I The current situation, however, is quite perplexing. After having reviewed and discussed the main, approaches pre~ently debated, the two authors candidly conclude thaf 'no theory stands out as providing the best comprehensive account of concepts'.2 Jerry Fodor, whose recent book Concepts is significantly subtitled Where Cognitive Science Went Wrong; thinks that important confusions about the very idea of a concept are at the root of what he takes to be, a deep theoretical crisis in cognitive sciences: 3 'most of what contemporary cognitive science believes aboutconcepts', he writes, 'is radically,and practically demonstrably, untrue',4 'Who could deny', Eric Margolis adds in a contribution of his own, 'that, we,are in need of some new theoretical options?'5 Well, if 'new' options are needed, maybe old,- and longforgotten -,- ones could help too. The history of philosophy, here as on so many other questions, can be a precious source for intriguing suggestions and potential breakthroughs. And William of Ockham in tpis regard stands out as an exceptionally good prospect for concept theorists: he has an elaborate and sophisticllted network of thes~s and argull:lents to offer about ontology, mind, thought and language,,with the idea of 'conceptus' at the very centre of it. The access to Ockham's doctrine, admittedly, is not immediate fortoday's readers., Not only are the texts in Latin with few translations available, but most importantly the theoreticalbackground is very dissimilar from ours. With most of his colleagues, Ockham accepted the general adequ~cy of the Aristotelian tradition in philosophy, as well'as the predominant authority of the Christian religious beliefs. His writings often responded to, those of the scholastic heroes or the previous decades such as Thoma~ Aquinas orjohn,duns Scotus~ and, of course, he had never read a line of Frege or Russell. The logic he had learnt at school was Aristotle's,. augmented b~l the terministdevelopments of the thirteenth century, the theory of suppositioespecially. Asa consequence of all this, his theoretical vocabulary was different from ours all the way through. Yet there is no fatal incommensurability in this case. Ockham's approach can still be of greatrelevance with respect to certain deep concerns that are at the forefront of today's interest in cognitive sciences and philosophy of mind. Some of the main puzzles we now face in these fields have to do with how to reconcile a thoroughly causal picture of the natural world with the intentionality and intelligence of the human mind, and this, as far as I can see, was Ockham's problem as well. In his small penetrating book, The Elm and the Expert, Jerry Fodor has called attention to the difficulty there is for a modern representationalist view of the mind of harmonizing three basic. and prin/{) facie quite plausible. tenets:" IMI

97 182 CONCLUSION 183 that a sound psychological theory of thought cannot do without intentional terms; that the content of a thought depends upon some external - especially causal - connections it has with the world; that human intellectual processes typically have a syntactical character. It so happens that Ockham's theory saliently incorporates a close analog for each one of these: concepts are taken to be mental signs for things, their signification is made to depend upon causal connections, and they are insistently described as combinable with each other into syntactically structured units comparable to the sentences of a language. How he strives to bring it all together in his mature actus-theory, while consistently subscribing to a nominalistic metaphysics, was the object of the present book. Roughly summarized, what emerges is the following. The basic apparatus Ockham postulates is a complex causal network of mental states. Each encounter with real individuals is supposed to trigger in the mind one or more chains of intellectual acts and habitus, ontologically identified with mental qualities. How semantics enters the picture depends on the roles these intellectual acts are supposed to play in the mental life of human beings. They can be described as signs in the logical sense, Ockham thinks, because they can be combined into more complex units, some of which can legitimately be seen as propositions and attributed a truth-value. Whether a complex propositional act is true or false is made to depend upon its syntactical structure on the one hand, and the signification of (some of) its conceptual components on the other hand. And how a simple concept naturally gets a signification is basically presented as a causal story. The human mind is so built, according to the theory, that when the intelligent agent comes in perceptual contact with one or more individuals of the world, a special intellectual act is immediately produced in him, the 'intuitive cognition', which then launches various series of mental effects, giving rise, among other things, to a new mental act, the 'abstractive' one, which will be, in turn, the starting point of a temporally spread out chain of future similar acts, with the corresponding habitus in between. Simple concepts in this network are those abstractive acts of which no other abstractive act is a part. What each one of them signifies will be all the individuals (or groups of individuals) that relevantly resemble, to some specified degree, the individuals (or groups of individuals) that directly caused the occurrence of the original triggering intuitive act. The 'first abstractive cognition' in particular - the prima abstractiva.:... signifies all the individuals that are essentially equivalent - simillimi, Ockham says - with the original Ones. More and more general abstractive cognitions are then supposed to be progressively induced in the mind as the combined effects of previously acquired ones, with less and less stringently similar individuals as their significates. Why exactly should such concepts signify individuals that are essentially similar to each other rather than any other arbitrarily chosen bunch, is something Ockham does not explain in any detail. What we are left to suppose is that this must depend somehow upon the functions of the conceptual units in the mental life of the agent. As seen in chapter 7, Ockham postulates at least two such functions. First, concepts are thought to play a recognitional role. Without insisting much on the point, Ockham standardly presupposes (hat each simple concept is associated in the mind with a corresponding disposition of the agent to apply this concept to new instances (through judgements oftheform 'this is an F' in particular). Secondly, andockham is much more loquacious about this one, the concept is expected to contribute something to the truth-conditions of the propositional sequences it occurs in as subject or predicate. The former role is fulfilled by the concept being a 'similitude' of some externalthings, while the latter is what we can properly call the semantical role of the concept: this is where its signification comes in. If concepts are to efficiently play both roles at the same time, there had better be a coincidence in most cases between what the agent is disposed to apply the concept to, and what the concept signifies. Yet the two relations must be distinguished in principle in Ockham's theory, for the latter is the norm of the former. The dispositionofahuman agent to apply a previously acquired concept to a new instance must be decisively sensitive, as we have argued, to the perceptual aspects of the newly met individual. Perceptual aspects, however,can sometimes be misleading. In so far as the judgement that this newly met thing here is an F is supposed to guide the agent's future expectations and actions, what makes it true or false ultimately should be, in many cases, the essential causal powers of the thing rather than its superficial appearance. It should not come as a surprise, then, that what a concept is said to naturally signify, in the logical sense, should be those individuals that are essentially - and not only apparently -like the ones that originally triggered its own formation. Be that as it may, Ockham ends up with three basic categories of simple concepts, which he respectively calls 'absolute' terms, 'connotative' terms, and syncategoremata: The syncategorematic concepts, when they occur in actual thought, are the very logical operations of the mind, such as predication, quantification, and so on. The capacity to accomplish such operations must be supposed to be innate, in the. Ockhamistic view, since it essentially belongs to every rational soul. But the actual occurrence of these syncategorematic acts normally requires the accompanying presence of some independently SIgnifying mental categorematic terms, whether absolute or connotative. Absolute concepts are those that equally signify all their significates without obliquely referring to anything else. The sole connection they require among their significates is some degree or other ofessential similarity. They are, in pther words, natural kind concepts (with various degrees of generality available). Simple connotative concepts, by contrast, are endowed with a hierarchized semantical structure: they primarily signify some things, and secondarily - or obliquely - signify some other things. Their significates, in other words, are arranged in ordered n-tuples. These concepts are normally produced within the mind by encounters with groups of individuals, rather than with isolated instances, and are thus made to signify, directly or obliquely, all the individual participants to similarly ordered n-tuples. Despite their special hierarchized structure, they can legitimately be said to be simple acts of conceptual cognition in so far as no part of them is itself an independently signifying concept. In the end, all human reasoning must amount, in Ockham's theory, to combinations of such simple conceptual units, in accordance with the rigid constraints of an

98 184 CONCLUSION 185 innately implemented grammar..the representational value of thought is thus ultimately made to rest on the natural signification of our simple categorematic concepts, whether absolute or connotative. With respect to the. current discussions in philosophy ofmind, this theory displays several interesting features. Let me brieflyrecall some of them: (i) (ii) (iii) (iv) (v) (vi) (vii) (viii) (ix) (x) Epistemological realism. Although universals are not countenanced as real things in the world, cognition by means of general concepts is considered as adequate in principle to reality. The theory admits, in particular, that things are really ordered out there in the world in various ways (temporal, spatial, causal, etc.), and that they can, in many cases, be known to be so ordered. Representationalism in philosophy ofmind. It is assumed that human thought is composed, in a fundamental way, of units or processes which are semantically evaluable (as referring to something else, for example) and syntactically structured. A causal approach to cognition. The semantical values of the mental representational units ultimately depend upon their (often indirect) causal connections with the world. These units, in turn, play a causal role in the psychological life ofthe agent. The idea that concept$ are real mental states, or acts, rather than purely ideal objects grasped, or produced, by the mind. This allows for an appreciable simplification of the ontology of the mental, by avoiding the position of a special mode of being for intentional objects. The postulation ofan innate mechanism ofabstraction, in virtue of which, in particular, a single encounter with a given individual suffices to produce in the human mind the formation of a general concept of the' most basic level. This makes the postulation of innate categorematic concepts superfluous. Semantical atomism and the compositionality of thought. The basic units of human thought are simple concepts, considered as signs. They naturally acquire.their signification before occurring.as the components of propositional or inferential sequences. Once significant, they are combinable with each other in determinate ways, and the semantical properties of the complex units thus produced are usually derivable from the signification of their components, given the structure of the whole. The distinction between the signification of concepts and the referential function (or suppositio) they happen to have in the context of a given propositional sequence. This is'a crucial distinction for Ockham, and a very useful one indeed, which is unfortunately neglected in contemporary semantics (it does not correspond, in particular, to the famous Fregean couple of sense and reference). The idea ofconnotation and its application to simple concepts. Some of the basic conceptual units in our mental equipment are attributed a hierarchized semantical structure. Among them, in particular, are some relational terms. An innatist position with respect to syncategorematic concepts. The human mind is assumed to be innately capable of logical operations. An externalistic conception of the meaning ()f' words. The conventional signification of a spoken sound depends not upon what the speaker has in (xi) mind at the moment of utterance, but upon what happened at the original moment imposition. Some of Ockham's developments even suggest, as we saw in chapter 9, that a related form of externalism might extend to certain conceptual units, in so far as they could receive a stipulative signification in addition to their natural one. The adoption of a strong nominalistic constraint. All the relata. of the cognitional process - agents, concepts, significates, and referents - are held to be singular entities. ' The final picture we arrive at is quite remote, on the whole, from what has been the standard account of Ockham's approach to concepts in the last decades; much ofthis book has been devoted to a critical discussion of the predominant 'reductivistic' view, chiefly put forward by the American scholar Paul Vincent Spade. Ockham's language of thought, in this interpretation, was supposed to be a logically ideal language,and should, consequently, be devoid of ambiguities and redundancies; it should not simultaneously' contain, in particular, a simple term and its complex nominal definition: the only simple terms it can have in the end (in addition to syncategoremata) are the indefinable - and non-relational - 'absolute' concepts. What we are now led to conclude is that this approach cannot legitimately be attributed to Ockham himself. Some ambiguities are indeed accepted by him in the language of thought,7 and Some redundancies as well. Some connotative terms relational ones, in particular - are admitted among naturally acquired simple concepts, and they are authorized to coexist in human minds with their complex definitions. Even the rejection of strict synonymy among simple concepts, which is explicit enough in some of Ockham's texts, cannot be counted as a central tenet in his doctrine. 8 Our whole understanding of Ockham's nominalistic programme has turned out to be at stake in these technical points. Spade reasoned that since Ockham's ontology admitted of only two categories of individuals in the world - substances and qualities - only absolute substance concepts and absolute quality concepts could be accepted by him as simple representational units in the mind; everything else in intellectual thought should be accounted for on this narrow basis. The success of Ockham's nominalism was thought to crucially rest on the reduction of the connotative terms, including all the relational ones, by means of nominal definitions. Which, in the end, was unfeasible! The result of our inquiry, fortunately, is that Ockham's real aim was elsewhere. What the programme he puts forward is supposed to accomplish is to make it clear that none of the concepts and propositions that are needed for human knowledge carries with it an ontological commitment to any entity but singular substances and qualities. This is done not by reducing other concepts to absolute substance and quality concepts, but by showing that all the relata of the relevant semantical connections - including connotation - are singular substances and qualities. In the last analysis, what stood behind the reductivistic interpretation of Ockham, it seems to me, was the misleading assumption that the language of thought, as he conceives of it, should be entirely transparent to the thinking agent. It is this presupposition. mainly, that made it difficult for the proponents of this interpretation to understand how a connotative concept could bl' kept distinct from its definition in

99 186 Notes Ockham's approach. Yet the Venerabilis Inceptor, as we have found, did not take the conceptual content of our mental propositions to be thoroughly manifest to us in all cases. Not only did he accept ambiguities as possible in mentalese, but he is also committed to admit (as we haye seen in chapter 5) that one can have a concept without knowing what its definition is, and without being clear as to what entities exactly this concept signifies, and what their ontological status is. The upshot is that the externalistic trend in Ockham's theory of concepts should not be minimized: the signification of simple concepts is made to depend upon external relations that the agent might not be entirely aware of. And this ipso facto introduces some degree of opacity right at the heart of the mental language! This is not a problem for the Venerabilis Inceptor's philosophical programme, because his goal was not to construe human thought as a logically ideal and semantically translucent system. It was to clean up the ontology. Lac des Erables December Laurence and Margolis 1999, p Ibid., p See Fodor 1998, p Ibid., p. viii (with Fodor's italics). 5. Margolis 1998, p See Fodor 1995, esp. pp Spade 1974 had already noticed the point, but he suggested that this was an inconsistency on Ockham's part. 8. As argued above in chapter 7, section Ockham's works Bibliography Ockham's philosophical and theological writings have been edited by the Franciscan Institute of St. Bonaventure University under the direction of Fr. Gedeon Gal in two series: Guillelmi de Ockham Opera Philosophica [abbr. : OPh], 7 vols., St. Bonaventure, NY: The Franciscan Institute, Guillelmi de Ockham Opera Theologica [abbr. : OTh], 10 vols., St. Bonaventure, NY : The Franciscan Institute, Quoted in the present book are the following : Expositio in Libros Physicorum Aristotelis [abbr. : Exp. in Phys.], V. Richter et al. eds., OPh IV"V, Expositio in Librum Perihermenias Aristotelis [abbr. : Exp. in Perih.], A. Gambatese and S. Brown eds., in OPh II, 1978, pp Expositio in Librum Porphyrii de Praedicabilibus [abbr. : Exp. in Porph.], E. A. Moody ed., in OPh II, 1978, pp (Eng!. trans!. in Kluge ). Expositio in Librum Praedicamentorum Aristotelis [abbr. : Exp. in Praedic.], G. Gal ed., in OPh II, 1978, pp ; Expositio super Libros Elenchorum [abbr. : Exp. sup. Elench.], F. Del Punta ed., OPh III, Ordinatio. Scriptum in Librum Primum Sententiarum [abbr. : Ord.], G. Gal et al. eds., OTh 1 IV, 1967-, Quaestiones in Libros Physicorum Aristotelis [abbr. : Quaest. in Phys.], S. Brown ed., in OPh VI, 1984,pp Quaestiones Variae [abbr. : Quaest. Var.], G. I. Etzkorn et al. eds., OTh VIII, Quodlibeta Septem [abbr. : Quodl.], J. C. Wey ed., OTh IX, 1980 (Engl. trans!. in Freddo80 and Kelley 1991, and Freddoso 1991). Reportatio. Quaestiones in Libros II, Ill, IV Sententiarum [abbr. : Rep.], G. Gal et al. eds., OTh V-VII, Summa Logicae [abbr. : SL], P. Boehner, G. Gal and S. Brown eds, OPh I, 1974 (partial Engl. trans!. in Loux 1974, and Freddoso and Schuurman 1980). Summula Philosophiae Naturalis [abbr. : Summ. Phil. Nat.], S. Brown ed., OPh VI, 1984, pp Tractatus de Corpore Christi [abbr. : Tract. de Corp. Chr.], C. A. Grassi ed., in OTh X, 1986, pp Tractatus de Quantitate [abbr. : Tract. de Quant.], C. A. Grassi ed., in OTh X, 1986, pp Other references N.B.- It has not seemed useful to include such classical authors as Aristotle, Thomas Aquinas or John Locke in the present bibliography. References to their works when they occurred were given in the standard ways in the relevant footnotes. Adam de Wodeham. See Wood and Gal Adams. Marilyn (1970), '(n!uitive cognition. certainty. and scepticism in William Ockham', '1im/ifill. 20..lH9--9H. 1M?

100 188 BIBLIOGRAPHY 189 ~- (1987), William Ockham, 2 vols., Notre Dame, IN: University of Notre Dame Press. Alferi, Pierre (1989), Guillaume d'ockham. Le singulier, Paris: Minuit. Almog, Joseph, John Perry, and Howard Wettstein, eds (1989), Themesfrom Kaplan, Oxford: Oxford University Press. Ashworth, Jennifer (1973), 'The doctrine of Exponibilia in the fifteenth and sixteenth centuries', Vivarium, 11, ~- (1981), 'Mental language and the unity of propositions', Franciscan Studies, 41, (reprinted in Ashworth 1985). ~- (1985), Studies in Post-Medieval Semantics, London: Variorum Reprints. Augustine. See Martin Bakker, Paul J. J. M., ed. (2002), Chemins de la pensee medievale. Etudes offertes azenon Kaluza, Tumhout: Brepols. Beretta, Beatrice (1999), Ad aliquid. La relation chez Guillaume d' Occam, Fribourg, Suisse: Editions Universitaires. Berube, Camille (1964), La connaissance de l' individuel au Moyen Age, Montreal: Presses de l'universite. de Montreal, and Paris: P.U.F. Biard, Joel (1981), 'La redefinition ockhamiste de la signification', Miscellanea Mediaevalia, 13, 451~8. ~- (1989), Logique et tmorie du signe au XlVe siecle, Paris: Vrin. ~- (1997), Guillaume d' Ockham. Logique et philosophie, Paris: P.U.F. Boehner, Philotheus (1943), 'The notitia intuitiva of non-existents according to William of Ockhain', Traditio, 1, (also in Boehner 1958, ). ~- (1945), 'In propria causa', Franciscan Studies, 5, (also in Boehner 1958, ). ~- (1951), 'The relative date of Ockham's Commentary on the Sentences', Franciscan Studies, 11, (also in Boehner 1958, ). ~- (1958), Collected Articles on Ockham, St. Bonaventure, NY: The Franciscan Institute. Boler, John (1985), 'Connotative terms in Ockham', History of Philosophy Quarterly, 2, Bos, Egbert P. (1994), Johannes Buridanus. Summulae in Praedicameilta, Nijmegen: Ingenium. Braakhuis, H. A. G and C. H. Kneepkens, eds (2003), Aristotles'sPeri Hermeneias in the Latin Middle Ages. Essays onthe Commentary Tradition, Groningen: Ingenium. Brown, Deborah (1996), 'The puzzle of names in Ockham's theory of mental language', The Review ofmetaphysics, 50, Brown, Stephen (1985), 'Walter Chatton's Lectura and William of Ockham's Quaestiones in Libros Physicorum Aristotelis', in Frank artd Etzkorn 1985, Cesalli, Laprent (2001), 'Le realisme propositionnel de Walter Burley', Archives d' Histoire Doctrinale et Litterairedu Moyen Age, 68, Chalmers, David (1999), 'Is there synonymy in Ockham's mental language?', in Spade 1999, Chatelet, Fran\(ois, ed. (1972), La philosophie medievdle, Paris: Hachette. Conti, Alessandro (2000), 'Significato e verita in Walter Burley', Documenti e studi sulla tradizionefilosofica medievale, 11, Crathorn. See Hoffmann Davidson, Donald and Gilbert Harman, eds (1972), Semantics of Natural Language, Dordrecht: Reidel. Day, Sebastian (1947), Intuitive Cognition. A Key to the Significance (!f'later Scholastics, St. Bonaventure, NY: The Franciscan Institute. De Andres, Teodoro (1969), EI nominalismo de Guillermo de Ockham como tilosolla del lenguaje. Madrid: Clredos. Eco, Umberto and Costantino Marmo,eds. (1989) On the Medieval Theory of Signs, Amsterdam: John Benjamins. Fodor, Jerry A. (1995), The Elm and the Expert, Cambridge, MA: MIT Press. ~- (1998), Concepts. Where Cognitive Science Went Wrong, Oxford: Clarendon Press. Frank, W.and Girard J. Etzkorn, eds (1985), Essays Honoring Allan B. Wolter, St. Bonaventure, NY: The Franciscan Institute. Fredborg, K. M., Lauge Nielsen, and Jan Pinborg (1978), 'An unedited part of Roger Bacon's Opus Maius: De Signis', Traditio, 34, ; Freddoso, Alfred J. (1991), William of Ockham. Quodlibetal Questions, vol. II: Quodlibets 5-7, New Haven, CT: Yale University Press. Freddoso, Alfred J. and Francis E. Kelley (1991), William ofockham. Quodlibetal Questions, vol. I: Quodlibets 1-4, New Haven, CT: Yale University Press. Freddoso, Alfred J. and Henry Schuurman (1980), Ockham's Theory ofpropositions. Part II ofthe Summa Logicae, Notre Dame, IN: University of Notre Dame Press: Gal, Gedeon (1967), 'Gualteri de Chatton et Guillelmi de Ockham controversiade natura concept4s universalis', Franciscan Studies, 27, 191~212. Gaskin, Richard (2001), 'Ockham's mental language, connotation, and the inherence regress', in Perler 2001, Gelber, Hester (1984), "'I cannot tell a lie": Hugh Lawton's critique of Ockham on mental language', Franciscan Studies, 44, Goddu, Andre (1984), The Physics ofwilliam ofockham, Leiden: Brill. ~- (1993), 'Connotative concepts and mathematics in Ockham's natural philosophy', Vivarium, 31, Goodman, Nelson (1956), 'A world of individuals', in The Problem of Universals (with Alonzo Church and I. M. Bochenski), Notre Dame, IN: University of Notre Dame Press (also in Goodman 1972, ). ~- (1972), Problems and Projects, Indianapolis, IN: Bobbs-Merrill. Gregory of Rimini. See Trapp et al Hamesse, Jacqueline (1974), Les Auctoritates Aristotelis. Un jlorilege medieval, Louvain: Publications Universitaires, and Paris: Beatrice-Nauwelaerts. Hamlyn, D. W. (1993), Aristotle. De Anima. Books II and Ill, Oxford:Clarelldon Press. Henninger, Mark G. (1989), Relations. Medieval Theories , Oxford: Clarendon Press. Rochart, Patrick (1972), 'G!Iillaume d'occam : Le signe et sa duplicite', in Chatelet 1972, Hoffmann, Fritz (1988), Crathorn. Quiistionen zum ersten Sentenzenbuch, MUnsler: Aschendorff. Hornsby, Jennifer (1998), 'Action', in Routledge Encyclopedia of Philosophy, I, London: Routledge, John Buridan. See Bos 1994, King 1985, Patar 1991, Scott 1977, and Van der Lecq Kaplan, David (1978), 'On the logic of demonstratives', Journal of Philosophical Logic, ~- (1989), 'Demonstratives', inalinog et al. 1989, Karger, Elizabeth (1978), 'Consequences et inconsequences de la supposition vide dans la logique d'ockham', Vivarium, 16, ~- (1994), 'Theories de la. pensee, de ses objets et de son discours chez Guillaume d'occam', Dialogue, 33, ~- (1996), 'Mental sentences according to Burley and to the early Ockham', Vivarium ~- (1999a), 'Ockham's misunderstood theory of intuitive and ahstractive cognition', in Spade _.- (191.)90), 'Waller Burley's ~ealism', Vi\'(/rillm,.n, 24-4(),

101 190 BIBLIOGRAPHY 191 Kaufmann, Matthias (2003), 'The discussion of the nature of the concept in Ockham's Perihermeneias commentary', in Braakhuis and Kneepkens 2003, Kelley, Francis E. (1981), 'Walter Chatton vs. Aureoli and Ockham regarding the universal concept', Franciscan Studies, 41, King, Peter (1985), Jean Buridan's Logic. The Treatise on Supposition. The Treatise on Consequences, Dordrecht: Reidel. -_. (2002), 'The failure of Ockham's nominalism', klima/pking.htm Klima, Gyula (1988), Ars Artium. Essays in Philosophical Semantics, Mediaeval and Modern, Budapest: Hungarian Academy of Sciences. -_. (1993), 'The changing role of entia rationis in mediaeval semantics and ontology: a comparative study with a reconstruction', Synthese, 96, (1999),'Ockham's semantics and ontology of the categories', in Spade 1999, (2001), 'Buridah's theory ofdefinitions in his scientific practice', in Thijssen and Zupko 2001, (2002), 'Comments on Peter King "The failure of Ockham's nominalism''', Kluge, Eike-Henner W. ( ), 'William of Ockham's Commentary on Porphyry. Introduction and English translation', Franciscan Studies, 33, , and 34, Kneale, William and Martha Kneale (1962), The Development oflogic, Oxford: Clarendon Press. Kretzmann, Norman (1982a), 'Syncategoremata, exponibilia, sophismata', in Kretzmann, et al. 1982, , ed. (1982b), Infinity and Continuity in Ancient and Medieval Thought, Ithaca, NY: Cornell University Press. Kretzmann, Norman, Anthony Kenny and JanPinborg, eds. (1982), TheCambridge History oflater Medieval Philosophy, Cambridge: Cambridge University Press. Kripke, Saul (1972), 'Naming and necessity', in Davidson and Harman 1972, Laurence, Stephen and Eric Margolis (1999), 'Concepts and cognitive science', in Margolis and Laurence 1999,3-81. Leff, Gordon (1975), William of Ockham. The Metamorphosis of Scholastic Discourse, Manchester: Manchester University Press. Lindbergh, David C (1976), Theories ofvision from. Al-Kindi to Kepler, Chicago, IL: The University of Chicago Press. Loux, Michael (1974), Ockham's Theory ofterms. Part I ofthe Summa Logicae, Notre Dame, IN: University of Notre Dame Press. -- (1979), 'Significatio and suppositio. Reflections on Ockham's semantics', The New Scholasticism, 53, Maieril, Alfonso (2002), 'Linguaggio mentale e sincategoremi nel secolo XIV', in Bakker 2002,3-25. Margolis, Eric (1998), 'How to acquire a concept', Mind and Language, 13 (also in Margolis and Laurence 1999, ). Margolis, Eric and StephenLaurence, eds (1999), Concepts. Core Readings, Cambridge, MA: MIT Press. Martin, I. (1962), Augustine. De Doctrina Christiana, Tumhout: Brepols. Maurer, Armand (1978), 'Method in Ockham's nominalism', The Monist, 61, (also in Maurer 1990,403-21). -- (1981), 'William of Ockham on language and reality', Miscellanea Mediaevalia, 13, 795-R02 (also in Maurer 1990, ). -~ (19l\4), 'Ockham's Razor and Chalton's Anti-Razor', Medilll'\'al Studies, 46, (also in Maurer 1990, ). -- (1990), Being and Knowing. Studi~, ill Thomas Aquinas and Later Medieval Philosophers, Toronto: Pontifical Institute of MedileYal Studies. -- (1999), The Philosophy ofwilliam ofo~ III,,.. Light ofits Principles, Toronto: Pontifical Institute of Mediaeval Studies, McGrade, Stephen (1986), 'Seeing things: Ockhun IlIclr.preaentationalism', in Wenin 1986, Meier-Oeser, Stephan (1997), Die Spur des UiCMIII." Z.kMn und seine Funktion in der Philosophie des Mittelalters und der fruhen Neuz,',. Berlin: Walter de Gruyter. Mertz, D. W. (1997), Moderate Realism and Its Logic, NIw Haven, CI': Yale University Press. Michon, Cyrille (1994), Nominalisme. La tmorie de ItI,i,IIUfctllion d'occam, Paris: Vrin. -- (2000), 'Le nominalisme et les relations', Le~"",., ltwoi", I, Minio-Paluello, Laurentius (1966), Aristoteles LatirllU I, ~7: Porphyrii Isagoge et Anonymi Fragmentum vulgo locatum 'Liber Sex PrincipiarlUll', Bruge.: Descl6e de Brouwer. Normore, Calvin (1990), 'Ockham on mental languaae', In Smith 1990, O'Callaghan, Jeremiah (1955), 'The second question of the Prologue to Walter Catton's Commentary on the Sentences on intuitive and abstractive knowledge', in O'Donnell 1955, O'Donnell, J. Reginald, ed. (1955), Nine Mediaeval Thinkers. A Collection of Hitherto Unedited Texts, Toronto: Pontifical Institute of Mediaeval Studies. Panaccio, Claude (1980), 'Occam et les d6monstratifs', Historiographia Linguistica, 7, '--(1984), 'Propositionalism and atomism in Ockham's semantics', Franciscan Studies, 44, (1990a), 'Connotative terms in Ockham's mental language', Cahiers d'epistemologie no. 9016, Montreal: Universite du Quebec 11 Montreal. -- (1990b), 'Reference et representation' (about Alferi 1989), L'Age de la science, 3, (1992a), Les mots, les concepts et les choses. La semantique de Guillaume d' Occam et Ie nominalisme d' aujourd' hui, Montreal: Bellarmin, and Paris: Vrin. --(1992b), 'From mental word to mental language', Philosophical Topics, 20, (1999a), Le discours interieur. De Platon a Guillaume d'ockham, Paris: Seuil. --(1999b), 'Semantics and mental language', in Spade 1999, (2001), 'Aquinas on intellectual representation', in Perler 2001, (forthcoming), 'Ockham and Buridan on epistemic sentences: Appellation of the form and appellation of reason', in the acts of the conference 'Logica Modernorum' held in Amsterdam in 1998, Leiden: Brill. Panaccio, Claude and Ernesto Perini-Santos (forthcoming), 'Guillaume d'ockham et la suppositio materialis', Vivarium, 42. Parrett, Hermann, ed. (1976), History oflinguistic Thought and Contemporary Linguistics, Berlin: Walter de Gruyter. Pasnau, Robert (1993), 'Petri Iohannis Olivi Tractatus de Verbo', Franciscan Studies, 53, (1997), Theories of Cognition in the Later Middle Ages, Cambridge: Cambridge University Press. Patar, Benoit (1991), Le Traite de l'ame de Jean Buridan, Louvain-Ia-Neuve: Editions de l'institut Superieur de Philosophie. -- (1995), Nicolai Oresme Expositio et Quaestiones in Aristotelis De Anima, Louvain-Ia Neuve: Editions de l'institut Superieur de Philosophie, and Louvain: Peeters. Perler, Dominik (I 996a), 'Things in the mind. Fourteenth-century controversies over "intelligible species'''. Vivarium, 34, (1996b), Review of Spruit , Vivarium, 34, 2l\O-83. -~~, cd. (200 I), Ancient ami M, ti//'i'al 1'11,'01'/".\' ofilitl'litiollality, Leiden: Brill.

102 192 BIBLIOGRAPHY 193 Peter John Olivi. See Pasnau Pinborg, Jan (1976), 'Some problems of semantic representations in medieval logic', in Parrett 1976, (reprinted in Pinborg 1984). --(1984), Medieval Semantics, London: Variorum Reprints. Porphyry. See Minio-Palluelo Putnam, Hilary (1975a), 'The meaning of "meaning''', in Putnam 1975b, (1975b), Mind, Language.and Reality [=Philosophical Papers, vol. 2], Cambridge: Cambridge University Press. -.- (1988), Representation and Reality, Cambridge, MA: MIT Press. Reina, Maria Elena (2002), HocRic etnunc. Buridano, Marsilio di Inghen e la conoscenza del singolare, Florence: Leo S. Olschki. Richard Campsall (Pseudo). See Synan Rijk, Lambert Marie de (1967), Logica Modernorum. A Contribution to the History ofearly Terminist Logic, vol. II: The Origin and Early Development of the Theory of Supposition, Assen: Van Gorcum. Roger Bacon. See Fredborg et al Rosier, Irene (1994), La parole comme acte. Sur la grammaire et la semantique au XlIle siecle, Paris: Vrin. Russell, Bertrand (1918), 'The philosophy of logical atomism', in Russell1956, (1956), Logic and Knowledge. Essays , ed. by R. C. Marsh, London: George Allen and Unwin. --(1959), My Philosophical Development, London: George Allen and Unwin. Scott, Theodore K. (1977), Johannes Buridanus. Sophismata, Stuttgart: Frommann Holzboog. Smith, 1. C., ed. (1990), Historical Foundations ofcognitive Science, Dordrecht: Kluwer. Spade, Paul Vincent (1974), 'Ockham's rule ofsupposition: Two conflicts in his theory',. Vivarium, 12,63-7 (reprinted in Spade 1988). -- (1975), 'Ockham's distinctions between absolute and connotative terms', Vivarium, 13, (reprinted in Spade 1988) (1980), 'Synonymy and equivocation in Ockham's mental language', Journal of the History ofphilosophy, 18,9-22 (reprinted in Spade 1988). -.- (1988), Lies, Language and Logic in the Late Middle Ages, London: VariorumReprints. --(1990), 'Ockham, Adams and connotation. A critical notice of Marilyn Adams, William Ockham', The Philosophical Review, 99, (1996), Thoughts, Words and Things: An Introduction to Late Mediaeval Logic and Semantic Theory, -- (1998), 'Three versions of Ockham's reductionist program', Franciscan Studies, 56, , ed. (1999), The Cambridge Companion to Ockham, Cambridge: Cambridge University Press. Spruit, Leen ( ), 'Species Intelligibilis'. From Perception to Knowledge, 2 vols, Leiden: Brill. Streveler, Paul (1975), 'Ockham and his critics on intuitive 'cognition', Franciscan Studies, 35, Stump, Eleonore (1982), 'Theology and physics in De sacramento altaris: Ockham's theory of indivisibles', in Kretzmann 1982b, (1999), 'The mechanisms of cognition: Ockham on mediating species', in Spade 1999, Synan, Edward A. (1982), The Works o.t"richard (~t" Campsall, Toronto: Pontifical Institute of Mediaeval Studies. Tabarroni, Andrea (1989), 'Mental signs and the theory of representation in (kkham', in Ero and Marmo 1989, Tachau, Katherine H. (1988), Vision and Certitude in the Age of Ockham, Leiden: Brill. Thijssen, J. M. M. H. and Jack Zubko, eds (2001), The Metaphysics and Natural Philosophy ofjohn Buridan, Leiden: Brill. Trapp, Damasius et al. (1978- ), Gregorii Ariminensis Lectura super Primum et Secundum Sententiarum, Berlin: Walter de Gruyter. Trentman, John (1970), 'Ockham on mental', Mind, 79, Tweedale, Martin (1990), 'Mental representations in later medieval scholasticism', in Smith 1990, (1992), 'Ockham's supposed elimination of connotative terms and his ontological parsimony', Dialogue, 31, Van der Lecq, Ria (1998), Johannes Buridanus. Summulae de Suppositionibus, Nijmegen: Ingenium (Engl. transl. in King 1985). Walter Chatton. See G::il1967,Kelley 1981, O'Callaghan 1955, Wey 1989, Weyand Etzkorn Wenin, Christian, ed. (1986), L'homme et son univers au Moyen Age, Louvain-la-Neuve: Editions de l'institut Superieur de Philosophie. Wey, Joseph C. (1989), Walter Chatton. Reportatio et Lectura super Sententias: Collatio ad Librum Primum et Prologus, Toronto: Pontifical Institute of Mediaeval Studies. Wey, Joseph and Girard 1. Etzkorn (2002), Walter Chatton. Reportatio super Sententias Liber I, Distinctions 1-9, Toronto: Pontifical Institute of Mediaeval Studies. Wittgenstein, Ludwig (1921), Tractatus Logico-Philosophicus, transl. by D. F. Pears and B. F. McGuinness, London: Routledge and Kegan Paul, Wood, Rega (1997), Ockham on the Virtues, West Lafayette, IN: Purdue University Press. Wood, Rega and Gedeon Gal (1990), Adam de Wodeham. Lectura Secunda in LibrumPrimum Sententiarum, St. Bonaventure, NY: St. Bonaventure University, 3vols. Yrjonsuuri, Mikko (1997), 'Supposition and truth in Ockham's mental language', Topoi, 16, Zavalloni, Roberto (1951), Richard de Mediavilla et la controverse sur la pluralite des formes, Louvain: Publications Universitaires. Zheng, Yiwei (1998), 'Metaphysical simplicity and semantical complexity of connotative terms in Ockham's mental language', The Modern Schoolman, 75,

103 Index abbreviations 93--4, absolute terms (or concepts) 63-6, 69, 75, 85, 89-90, 92, 103, 105, 108-9, 112, 126-9,131-7,170-71,175,183 abstractive cognition 6-7, 9, 113, 121-2, 124,127,174,182 actus-theory (of concepts) 8-9, 23-7, 47, , , 158 Adam de Wodeham 38,148,160 Adams, Marilyn 1, 11-13, 16-20,25,37-8, 69,74,77,98,102,116,119,139,147, 153, 155, Alferi, Pierre 17,20,119,139 Alhazen 27 Aquinas see Thomas Aquinas Aristotle 8, 21-2,28, 30, 34, 42, 51, 55, 60,. 67,70,78-9,81, 121-2, 139, 163, 181 Ashworth, Jennifer 74, 82 assent see judgement Augustine 45, 58 Bacon see Roger Bacon Beretta, Beatrice 77, 97, 102, 151, Berube, Camille 17 Biard, Joel 39, 60, 119, 122-3, Boehner, Philotheus Boethius 8, 139 Boler, John 71 Brown, Deborah 101, 133, 136-8, Brown, Stephen 37, 98 Buridan see John Buridan. Campsall see Pseudo-Richard Campsall categories (Aristotelian theory of) 23,67, 75~6, Cesalli, Laurent 37 Chalmers, David 101 Chatton see Walter Chatton Classical View (of concepts) see Definitional View connotative terms (or concepts) 63-77, 85-'97, , , 154-8, 169, 173-5, Conti, Alessandro 37 copula XI), 1)1, 96, X, 152-3, ISS-X correlative terms 6X I), X6, I I I cospecificity see maximal similarity Crathom see William Crathorn Day, Sebastian 17 De Andres, Teodoro 60 definition see nominal definitions Definitional View (of concepts) 85, Donatus 8 equivocity 167-8, Etzkorn, Jerry 38-9 exponib1e propositions externalism 165,170-71,175-6,184-6 jictum-theory (of concepts) 8, 23-7, 35, 106, , , 158 Fodor, Jerry 3, 60, 97, 181-2, 186 Frege, Gottlob 2, 65, 89, 181 Gal, Gedeon 37-9, 98 Gaskin, Richard 74-7,79,82-3,97, 139, 143, 156, 162 Gelber, Hester 160 Goddu, Andre 77 Goodman, Nelson 161 Gregory of Rimini 32-3, 41 habitus 7,21-3,28-30,36,56-8,137,172, 182 Henninger, Mark G. 102 Hochart, Patrick 60 Hornsby, Jennifer 36 imposition 166, 168, ' in recto I in obliquo 71-2,87-93, , innatism 96, 150, 154, 157-8, intellectual acts see mental acts intelligible species intentions see second intentions intuitive cognition 6-7, 11-14,21-2,29, 103, 107, 110, , 124, 126-7, 171, 174, 182 John Buridan 72,82, 125, 127, 141, 156-7, IlJ~

104 196 INDEX 197 John Duns Scotus 5, 25, 28, 31, 38,105, 166, 176-7, 181 John Peckham 27 Judgements 31, 35-6,148 Kaplan, David 13, 20 Karger, ElizabethJ7-20, 23, 37,42,120, 140, 148, 160 Kelley, Francis E. 37 King, Peter 82 Klima, Gyula 133--,6, 142 Kneale, William and Martha 19 Kretzmann, Norman 59, 162 Kripke, Saul 3, 178 language of thought see menta11anguage Laurence, Stephen 97, 181,186 Leff, Gordon 18 Lindbergh, David C. 39 Locke, John , 177 Loux, Michael 59-60, 143, 177 Maieru, Alfonso 160 Margolis, Eric 2-3, 97, 181, 186 Maurer, Armand 18, 39 maximal similarity 10, 123, 126, 129, 135 McGrade, Stephen 20 Meier-Oeser, Stephan 58 mental acts 5-7, 15-16,21-36,55-7, 122-4, 151-4, mental language 8-15, 51, 55, 63, 66-9, 73-5,86-7,91-2,136-9,145-7,156-7, 166, 174-5, mental popositions 11-12,31-4,53-5, 136-7, , 169, 182~ mental signs 8, 11-12,26'-7,45-58, 166, 173-4, 182; see also signification Mertz, D. W. 77 Michon, Cyrille 17,45-50,58-9,76,78, 81-2,97,102,119-21,140,162 modes of signification 70-72, 91 nominal definitions 65-77,85-97, 103, , 156-7, 169, nominalism 1,3, 10,45, 55-8, 63-5, 74, 90, 96-7, , 134, 156,182, 185 Nonnore, Calvin 18,69,74,77, 116, 147, 154-5, 160, 162, , 177 Ockham's Razor 9,20,29-30, 120 Panaccio, Claude 18-20, 37-8,41, 58-61, 77-8, , 162, 176, 179 Pasnau, Robert 29, Perini-Santos, Ernesto 179 Perler, Dominik 30, 39 Peter Aureoli 38 PeterJohn Olivi 38 Peter of Ailly 74 Peter of Spain 8 Pinborg, Jan 82 Plato 28 Porphyry 97 Priscian 8 proper names see singular terms Pseudo-Richard Campsall 117 Putnam, Hilary 3, 138,143, 165, 171, 176-7, 179 Quine, Willard Van Orman 65 Reina, Maria Elena 141 relational tenns (or concepts) 63-9, 93-6, 106-7,109,111-12,115,130-31,155-6, 158, representationalism 16, 29-30, 184 Rijk, Lambert Marie de 18 Roger Bacon 27, 45, 58-9 Roger Marston 40 Rosier, Irene 59 Russell, Bertrand 2, 13-14,20,63,69,89, 181 Scotus see John Duns Scotus second intentions 151-2, 175 sign see mental sign, signification signification (of concepts) 8, 12,26,48, 51-4, 132-4, 166, 171-5, 182-4; see also mental signs similitude (or similarity) 10-11,25,29, , 174, 183; see also maximal similarity simple terms 31-2, 67-8,75, 103-5, 109, 134, Simplicius 96, 102 singular terms 11-15, ]36, Spade, Paul Vincent 4]-2,63-70,72,74--8, 80-82,85,89,91,97,99-100,103-6, 109,115-10,118,143,152, , ,177,185-6 Spruit, Lccll Streve1er, Paul 17 Stump, Eleonor 17, 30, 39, 101 subordination 114, 151, substance concepts 126-9, 131-4, 137, 139 supposition (suppositio) 5, 8-9, 11-12, ,26,49-54,63,94,120,150,166, , 181 syncategoremata 32, 34, 49-50, 52, 54, 66, 69, 89, 91, 112, , 173-4, synonymy 66, 69-74, 76-7, 85-7, 89,92, 114, 136-9, 176, 185 Tabarroni, Andrea 18 Tachau, Katherine 17,20,29,39-40 terms see absolute tenns, connotative tenns, correlative terms, in recto/in obliquo, mental signs, relational tenns, simple signs, type/token Thomas Aquinas 21, 23, 28~9, 31, 37-40, 51,59, 181 Trentman, John 143 Tweedale, Martin 39, 77, 79-80, 83 type / token 55--'-8, 136, Walter Burley 12,23,37-8 Walter Chatton 23-7, 37-8, 41 Wey, Joseph 38-9 William Crathorn 122, 131, 140 William of Sherwood 8 William of Ware 38 Witelo 27 Wittgenstein, Ludwig 145,159 Yrjonsuuri, Mikko 82 Zavalloni, Roberto 17 Zheng, Yiwei , 118