Forms and Concepts: Concept Formation in the Platonic Tradition 3110266318, 9783110266313

Forms and Concepts is the first comprehensive study of the central role of concepts and concept acquisition in the Plato

246 14 2MB

English Pages 407 Year 2012

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Forms and Concepts: Concept Formation in the Platonic Tradition
 3110266318, 9783110266313

Table of contents :
A word of thanks
Introduction
1. ‘How comes the mind to be furnished?’
2. Survey of recent literature
3. Structure and contents of this study
I. Concepts – (ancient) problems and solutions
1. What is a concept?
2. The relevance of concepts in ancient epistemological debates
3. Different models of concept acquisition in antiquity
4. Forms and concepts & problematic concepts
II. Plato on learning as recollection
1. Forms and concepts
1.1. The role of concepts in Plato
1.2. Forms, concepts, language
2. The Parmenides and the archaeology of conceptualism
2.1. Concepts as ‘one over many’
2.2. Refuting conceptualism?
2.3. Conclusion
3. The Meno on the different stages of recollection
3.1. The transition from opinion (doxa) to knowledge
3.2. Conclusion
4. The Phaedo on the necessity of innate knowledge
4.1. The deficiency argument (Phaedo 72e–77a)
4.2. The continuity between Meno and Phaedo
5. The Phaedrus on acquiring universal concepts
5.1. Recollection and concept attainment (Phaedrus 249b–c)
5.2. Forms, concepts, language again
5.3. Conclusion
6. Concept formation and concepts in the Timaeus, Theaetetus, and Sophist
6.1. Recollection in Plato’s later works
6.2. Innateness and the structure of the human soul
7. The limits of recollection
7.1. Some problematic concepts
7.2. Recollection and error
8. Forms, concepts, and recollection
III. Aristotle’s reaction to Plato
1. Aristotle and his teacher Plato
1.1. A strange couple
1.2. Aristotle’s arguments against innate knowledge
2. The origin and nature of mathematical concepts
2.1. Concepts and the division of sciences
2.2. A troublesome emendation
2.3. Abstraction and the qua-operator
2.4. Aristotelian and Platonic separation
2.5. Mathematical objects and concepts
2.6. Linking abstractionM and induction?
3. Universal concepts – induction (epagoge) and its different domains
3.1. A general definition of induction
3.2. Induction and its different domains
3.3. The language of induction
3.4. Different kinds of induction in Aristotle
3.4.1. Induction in dialectical and rhetorical practice
3.4.2. Digression: likeness and the charge of circularity
3.4.3. Induction in ethics and natural science
3.4.4. The troublesome case of ‘complete’ or ‘perfect induction’
3.5. Induction and the starting points of syllogism
4. Induction of first principles (Posterior Analytics II 19)
4.1. Introduction
4.2. What is the object of Analytica Posteriora II 19?
4.3. Articulation and summary of the argument
4.4. The relation of sense perception and intellect
IV. Three case studies: Alcinous, Alexander & Porphyry, and Plotinus
1. Alcinous between empiricism and recollection
1.1. The doctrine of the doxastic logos
1.2. Alcinous’ psychology
1.3. Empiricism vs. innate knowledge
2. Alexander of Aphrodisias & Porphyry on abstraction and universals
2.1. Alexander – elaborating Aristotle’s notion of abstraction
2.2. Neoplatonic readings of Alexander
2.3. Immanent forms, definitional natures, and universal concepts
2.4. A unitary theory of intellect?
2.5. Porphyry – an abstractionist malgré lui?
2.6. The ‘short commentary’ on Aristotle’s Categories
2.7. The epistemological digression in the commentary on Ptolemy
3. Plotinus – ‘Wegbereiter’ of Syrianus and Proclus
3.1. The doctrine of the twofold nature of the logoi (I): logoi as criteria in perceptual judgements
3.2. The doctrine of the twofold nature of the logoi (II): logoi as causes in matter
3.3. Plotinus on innate knowledge and recollection
V. Syrianus’ and Proclus’ attitude towards Aristotle
1. Amicus Aristoteles, sed
2. Syrianus’ and Proclus’ criticism of induction and abstraction
2.1. Criticizing abstracted universals
2.2. Who is the target of Syrianus’ and Proclus’ criticism?
VI. The crucial role of doxastic concepts in Proclus’ epistemology
1. Proclus on sense perception and phantasia
1.1. Place and character of sense perception
1.2. Olympiodorus on the ambiguity of sense perception
1.3. Proclus on phantasia
1.4. Phantasia and geometry
1.5. Resume: sense perception, phantasia, and concept formation
2. An innovation by Proclus: his theory of doxa
2.1. Unfair to Proclus? – Alleged inconsistencies in his theory of doxa
2.2. More recent studies on Proclus’ theory of doxa
2.3. Digression: doxa in Plato and Aristotle
2.4. Proclus on doxa and innate knowledge
2.5. Proclus on doxastic concepts (logoi doxastikoi)
2.6. Doxa correcting sense perception
VII. Proclus’ Platonic theory of concept attainment
1. The soul and its innate knowledge
1.1. The discursive nature of soul: Proclus on dianoia
1.2. Dianoia and the logoi of the soul
1.3. Sources of Proclus’ doctrine of the psychic logoi
1.4. Common notions and psychic logoi
2. The triad of recollection: forgetting – articulation – probole
2.1. Recollection after Aristotle
2.2. Forgetting
2.3. Articulation (diarthrōsis)
2.3.1. The sources: Plato, Aristotle, the Stoics, and Middle Platonism
2.3.2. The Anonymous in Theaetetum (AT)
2.3.3. Articulation: talent and error
2.3.4. Proclus on articulation
2.4. Probole
2.4.1. Meaning and translation
2.4.2. Probolē and geometry
2.4.3. Résumé
3. Proclus on learning and the acquisition of concepts
3.1. Concept formation and the Platonic dialogue
3.2. Recollection as an intentional act of the soul
3.3. Different stages of recollection
3.4. Different kinds of concepts
3.5. A difficult passage
3.6. Recollection and error
3.7. Problematic concepts
VIII. Plato and Aristotle in harmony? – Some conclusions
Bibliography
1. Editions and translations
2. Secondary literature
Indices
1. Index nominum
2. Index locorum
3. Index rerum

Citation preview

Christoph Helmig Forms and Concepts

Commentaria in Aristotelem Graeca et Byzantina Quellen und Studien Herausgegeben von Dieter Harlfinger · Christof Rapp · Marwan Rashed Diether R. Reinsch

Band 5

De Gruyter

Christoph Helmig

Forms and Concepts Concept Formation in the Platonic Tradition

De Gruyter

IV

ISBN 978-3-11-026631-3 e-ISBN 978-3-11-026724-2 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress.

Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.dnb.de abrufbar.

© 2012 Walter de Gruyter GmbH, Berlin/Boston Satz: Dörlemann Satz GmbH & Co. KG, Lemförde Druck und buchbinderische Verarbeitung: Hubert & Co. GmbH & Co. KG, Göttingen ÜGedruckt auf säurefreiem Papier Printed in Germany www.degruyter.com

V

Für Marialucrezia, Enrico und Elena

VI

VII

Table of contents

Table of contents A word of thanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . 1. ‘How comes the mind to be furnished?’ 2. Survey of recent literature . . . . . . . 3. Structure and contents of this study . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

XI

. . . .

. . . .

. . . .

. . . .

. . . .

1 1 5 9

I.

Concepts – (ancient) problems and solutions . . . . . . . . . . 1. What is a concept? . . . . . . . . . . . . . . . . . . . . . 2. The relevance of concepts in ancient epistemological debates 3. Different models of concept acquisition in antiquity . . . . 4. Forms and concepts & problematic concepts . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

13 13 24 29 35

II.

Plato on learning as recollection . . . . . . . . . . . . . . . . 1. Forms and concepts . . . . . . . . . . . . . . . . . . . . . 1.1. The role of concepts in Plato . . . . . . . . . . . . . 1.2. Forms, concepts, language . . . . . . . . . . . . . . . 2. The Parmenides and the archaeology of conceptualism . . . 2.1. Concepts as ‘one over many’ . . . . . . . . . . . . . . 2.2. Refuting conceptualism? . . . . . . . . . . . . . . . . 2.3. Conclusion . . . . . . . . . . . . . . . . . . . . . . 3. The Meno on the different stages of recollection . . . . . . . 3.1. The transition from opinion (doxa) to knowledge . . . 3.2. Conclusion . . . . . . . . . . . . . . . . . . . . . . 4. The Phaedo on the necessity of innate knowledge . . . . . . 4.1. The deficiency argument (Phaedo 72e–77a) . . . . . . 4.2. The continuity between Meno and Phaedo . . . . . . . 5. The Phaedrus on acquiring universal concepts . . . . . . . . 5.1. Recollection and concept attainment (Phaedrus 249b–c) . . . . . . . . . . . . . . . . . . . 5.2. Forms, concepts, language again . . . . . . . . . . . . 5.3. Conclusion . . . . . . . . . . . . . . . . . . . . . . 6. Concept formation and concepts in the Timaeus, Theaetetus, and Sophist . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

39 39 39 41 45 45 48 51 52 52 56 57 57 64 65

. . . . . . . . . . . .

65 70 71

. . . .

71

VIII

Table of contents

6.1. Recollection in Plato’s later works . . . . . . . 6.2. Innateness and the structure of the human soul 7. The limits of recollection . . . . . . . . . . . . . . 7.1. Some problematic concepts . . . . . . . . . . 7.2. Recollection and error . . . . . . . . . . . . . 8. Forms, concepts, and recollection . . . . . . . . . . III.

IV.

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

71 73 78 78 83 84

Aristotle’s reaction to Plato . . . . . . . . . . . . . . . . . . . . . . 1. Aristotle and his teacher Plato . . . . . . . . . . . . . . . . . . . 1.1. A strange couple . . . . . . . . . . . . . . . . . . . . . . . 1.2. Aristotle’s arguments against innate knowledge . . . . . . . . 2. The origin and nature of mathematical concepts . . . . . . . . . 2.1. Concepts and the division of sciences . . . . . . . . . . . . 2.2. A troublesome emendation . . . . . . . . . . . . . . . . . . 2.3. Abstraction and the qua-operator . . . . . . . . . . . . . . 2.4. Aristotelian and Platonic separation . . . . . . . . . . . . . 2.5. Mathematical objects and concepts . . . . . . . . . . . . . 2.6. Linking abstractionM and induction? . . . . . . . . . . . . . 3. Universal concepts – induction (epag¯og¯e) and its different domains 3.1. A general definition of induction . . . . . . . . . . . . . . . 3.2. Induction and its different domains . . . . . . . . . . . . . 3.3. The language of induction . . . . . . . . . . . . . . . . . . 3.4. Different kinds of induction in Aristotle . . . . . . . . . . . 3.4.1. Induction in dialectical and rhetorical practice . . . . 3.4.2. Digression: likeness and the charge of circularity . . . 3.4.3. Induction in ethics and natural science . . . . . . . . 3.4.4. The troublesome case of ‘complete’ or ‘perfect induction’ . . . . . . . . . . . . . . . . . 3.5. Induction and the starting points of syllogism . . . . . . . . 4. Induction of first principles (Posterior Analytics II 19) . . . . . . . 4.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. What is the object of Analytica Posteriora II 19? . . . . . . . 4.3. Articulation and summary of the argument . . . . . . . . . 4.4. The relation of sense perception and intellect . . . . . . . .

. . . . . . . . . . . . . . . . . . .

87 87 87 88 90 90 92 96 98 100 108 111 112 113 114 116 116 119 121

. . . . . . .

122 125 128 128 129 132 134

Three case studies: Alcinous, Alexander & Porphyry, and Plotinus . . 1. Alcinous between empiricism and recollection . . . . . . . . . . . 1.1. The doctrine of the doxastic logos . . . . . . . . . . . . . . 1.2. Alcinous’ psychology . . . . . . . . . . . . . . . . . . . . . 1.3. Empiricism vs. innate knowledge . . . . . . . . . . . . . . 2. Alexander of Aphrodisias & Porphyry on abstraction and universals 2.1. Alexander – elaborating Aristotle’s notion of abstraction . . .

. . . . . . .

141 141 142 144 147 154 155

IX

Table of contents

2.2. Neoplatonic readings of Alexander . . . . . . . . . . . . . . . 2.3. Immanent forms, definitional natures, and universal concepts . 2.4. A unitary theory of intellect? . . . . . . . . . . . . . . . . . . 2.5. Porphyry – an abstractionist malgré lui? . . . . . . . . . . . . 2.6. The ‘short commentary’ on Aristotle’s Categories . . . . . . . . 2.7. The epistemological digression in the commentary on Ptolemy 3. Plotinus – ‘Wegbereiter’ of Syrianus and Proclus . . . . . . . . . . 3.1. The doctrine of the twofold nature of the logoi (I): logoi as criteria in perceptual judgements . . . . . . . . . . . 3.2. The doctrine of the twofold nature of the logoi (II): logoi as causes in matter . . . . . . . . . . . . . . . . . . . . 3.3. Plotinus on innate knowledge and recollection . . . . . . . .

157 161 164 171 172 175 184

V.

Syrianus’ and Proclus’ attitude towards Aristotle . . . . . . . . 1. Amicus Aristoteles, sed … . . . . . . . . . . . . . . . . . . . 2. Syrianus’ and Proclus’ criticism of induction and abstraction 2.1. Criticizing abstracted universals . . . . . . . . . . . . 2.2. Who is the target of Syrianus’ and Proclus’ criticism? .

. . . . .

. . . . .

205 205 208 209 219

VI.

The crucial role of doxastic concepts in Proclus’ epistemology . . . . 1. Proclus on sense perception and phantasia . . . . . . . . . . . . . 1.1. Place and character of sense perception . . . . . . . . . . . 1.2. Olympiodorus on the ambiguity of sense perception . . . . 1.3. Proclus on phantasia . . . . . . . . . . . . . . . . . . . . . 1.4. Phantasia and geometry . . . . . . . . . . . . . . . . . . . 1.5. Résumé: sense perception, phantasia, and concept formation 2. An innovation by Proclus: his theory of doxa . . . . . . . . . . . 2.1. Unfair to Proclus? – Alleged inconsistencies in his theory of doxa . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. More recent studies on Proclus’ theory of doxa . . . . . . . . 2.3. Digression: doxa in Plato and Aristotle . . . . . . . . . . . . 2.4. Proclus on doxa and innate knowledge . . . . . . . . . . . . 2.5. Proclus on doxastic concepts (logoi doxastikoi) . . . . . . . . 2.6. Doxa correcting sense perception . . . . . . . . . . . . . .

. . . . . . . .

223 225 225 227 228 230 231 232

. . . . . .

232 233 240 243 254 260

. . . . . .

263 264 264 265 268 270

VII. Proclus’ Platonic theory of concept attainment . . . . . . 1. The soul and its innate knowledge . . . . . . . . . . 1.1. The discursive nature of soul: Proclus on dianoia 1.2. Dianoia and the logoi of the soul . . . . . . . . . 1.3. Sources of Proclus’ doctrine of the psychic logoi . 1.4. Common notions and psychic logoi . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . . .

186 191 195

X

Table of contents

2. The triad of recollection: forgetting – articulation – probol¯e . . . 2.1. Recollection after Aristotle . . . . . . . . . . . . . . . . . 2.2. Forgetting . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Articulation (diarthr¯osis) . . . . . . . . . . . . . . . . . . 2.3.1. The sources: Plato, Aristotle, the Stoics, and Middle Platonism . . . . . . . . . . . . . . . . . . . . . 2.3.2. The Anonymous in Theaetetum (AT) . . . . . . . . 2.3.3. Articulation: talent and error . . . . . . . . . . . . 2.3.4. Proclus on articulation . . . . . . . . . . . . . . . 2.4. Probol¯e . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.1. Meaning and translation . . . . . . . . . . . . . . 2.4.2. Probol¯e and geometry . . . . . . . . . . . . . . . . 2.4.3. Résumé . . . . . . . . . . . . . . . . . . . . . . . 3. Proclus on learning and the acquisition of concepts . . . . . . . 3.1. Concept formation and the Platonic dialogue . . . . . . . 3.2. Recollection as an intentional act of the soul . . . . . . . . 3.3. Different stages of recollection . . . . . . . . . . . . . . . 3.4. Different kinds of concepts . . . . . . . . . . . . . . . . 3.5. A difficult passage . . . . . . . . . . . . . . . . . . . . . 3.6. Recollection and error . . . . . . . . . . . . . . . . . . . 3.7. Problematic concepts . . . . . . . . . . . . . . . . . . . .

. . . .

. . . .

272 272 274 278

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

278 282 284 286 289 290 295 299 299 300 304 305 309 313 317 325

VIII. Plato and Aristotle in harmony? – Some conclusions . . . . . . . . . . 335 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343 1. Editions and translations . . . . . . . . . . . . . . . . . . . . . . 343 2. Secondary literature . . . . . . . . . . . . . . . . . . . . . . . . . 355 Indices . 1. 2. 3.

. . . . . . . . . Index nominum Index locorum . Index rerum . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

381 381 383 390

Table of contents

XI

A word of thanks This study took its departure a long time ago at the De Wulf-Mansion Centre (DWMC) of the Catholic University of Leuven, where the generous financial support from the Flemish Fund for Scientific Research (FWO – Vlaanderen) enabled me to work on a PhD-thesis about Proclus’ epistemology under the supervision of Carlos Steel. I would especially like to thank Carlos Steel, but also all the members and associate members of the Leuven Proclus group (esp. Pieter d’Hoine, Caroline Macé, Leen van Campe, and Gerd van Riel) for their help and inspiration. In this regard, special thanks are due to Guy Guldentops (now Thomas-Institut, Cologne). Initially, I started working on the fourth book of Proclus’ Commentary on the Parmenides, but eventually the focus of the project widened and came to encompass a broad range of ancient authors. One of my main sources of interest and motivation was the desire to better understand the views of Neoplatonists on the relation between Plato and Aristotle. The dissertation was defended in Leuven on the 26th of May 2006. I would like to express my gratitude to the members of the jury (Carlos Steel, Theo Kobusch, Frans de Haas, Russ Friedman, Gerd van Riel) for the fair and most helpful discussion at the defense and for their many helpful suggestions and thoughtful remarks, most of which I have tried to incorporate into this book. As such, the monograph is a thoroughly revised and greatly expanded version of the original thesis. I have added chapter one, chapter four, and parts of chapter five, and have reworked substantial sections of the chapters on Plato, Aristotle, and Proclus. It goes without saying that the thesis and the subsequent book would never have been completed without the help of many people, of which I can only name a few here. Let me thank my friends, colleagues and teachers Matthias Baltes, David Butorac, Radek Chlup, Klaus Corcilius, Amos Edelheit, Philip van der Eijk, John Dillon, Grit Galle, Relinde Geys, Pieter Sjoerd Hasper, Marc Hill, Niall Keane, Cal Ledsham, Christel Meier-Staubach, Jörn Müller, Christopher Noble, Jan Opsomer, Pasquale Porro, Christian Pfeiffer, Christof Rapp, Burkhard Reis, Arnis Redovics, Bernd Roling, Andreas Speer, Steven Spileers, and Matthias Vorwerk. Special thanks are due to my parents, Heinrich Josef and Maria Helmig, and to my sister, Riccarda. The final version of the manuscript profited from many fruitful and most enlightening discussions with Antonio L.C. Vargas. In addition, chapter four was scrutinized by the acute eye of George Karamanolis when he was a Humboldt Fellow

XII

A word of thanks

in Berlin. Charlotte Adams and Annika von Lüpke (the latter also contributed to the indices) did their best to correct my English, prepare the book for print and helped to improve the style considerably. Sabrina Lange has been working with me in Berlin for several years now and I would like to thank her for her indispensable help. Finally, I should like to express my sincerest thanks to the esteemed editors of the ‘Commentaria in Aristotelem Graeca et Byzantina’, especially Dieter Harlfinger, for having willingly accepted my study into their young, but already so splendid series. The book is dedicated to my children, Enrico and Elena, and especially to my wife Marialucrezia Leone. She always was and is a source of help and guidance and it was very good indeed that we met each other one particular day, almost ten years ago, in Leuven at the De Wulf-Mansion Centre.

‘How comes the mind to be furnished?’

1

Introduction Du kannst nur von innen dazulernen; das Äußere ist bloßer Anstoß (Peter Handke, Gestern unterwegs, 434)

1. ‘How comes the mind to be furnished?’ ‘Concepts are the most fundamental constructs in the theories of the mind.’1 Given their importance to all aspects of cognition, there has been a lively and continued interest among philosophers in the investigation and determination of how concepts originate or how we attain them.2 A prominent example is John Locke: ‘How comes [the mind] to be furnished? Whence comes it by that vast store […]?’3 The aim of the present study is to pursue the question as to how knowledge was seen to be acquired in the so called Platonic tradition. I shall suggest that knowledge attainment can best be described by appealing to the language of concepts. Special attention is devoted to the period of Athenian Neoplatonism as represented by Syrianus († 434 ad) and, to a greater extent by his pupil Proclus (412–485 ad). The undoubtedly significant epistemological contribution of these two philosophers has suffered from a notable degree of neglect in a time where Neoplatonic studies in general are experiencing a remarkable renaissance. I propose a reconstruction of the Platonic tradition in epistemology as a conscious effort to rethink the difficult interplay between sense perception and innate knowl-

1 2

3

Margolis / Laurence (1999b), 3. In my study, I will use ‘concept’ to signify a (fairly) stable mental content or entity (see below ch. I.1). Margolis / Laurence (1999b), 5–8, argue that the fact that such concepts are admittedly subjective (that is, they depend upon a particular mind) does not exclude that they are shareable. One necessary condition that a concept has to fulfil is, however, that it allows for the naming or recognition of something. Generally speaking, concepts can have different degrees of complexity. It is obvious, for instance, that my concept of ‘sun’ differs considerably from that of an astronomer. In ancient Greek philosophy there is no single term for ‘concept’, but rather several words, for instance, ennoia, katholou, or logos (see ch. I.1). John Locke, An Essay Concerning Human Understanding II.1.2 (ed. P. Nidditch).

2

Introduction

edge (recollection). The inquiry will take its start from the much debated relation between Plato and Aristotle. According to the standard narrative found in encyclopaedias and handbooks Plato and Aristotle are the two great antipodes prefiguring, as it were, more recent debates between empiricists and rationalists on the acquisition of knowledge. Plato’s doctrine of learning as recollection and his appeal to transcendent Forms as the ultimate and sole object of true knowledge do not leave much room, it would seem, for fostering the mind’s interest in epistemology’s worldly constraints. Reciprocally, Aristotle’s comparison of the soul with an empty writing tablet and his criticism of Plato’s theory of Forms and innate knowledge do not appear to betray much sympathy for the speculations of his teacher. Rather, he is viewed as the prototype of the modern empirical scientist who observes, collects, and dissects. Yet, as with every cliché, this one about Plato and Aristotle needs to be somewhat qualified. Both philosophers agree that true knowledge is only about unchanging objects (universals or Forms). Labelling Aristotle an empiricist does not entail that he defends a theory of knowledge comparable to the Greek medical empiricists or British empiricists like Locke and Hume.4 Aristotle’s empiricism is of a peculiar kind. A wholehearted ancient empiricist, such as the medical empiricist, would hold that experience can only provide facts, not reveal causes. Moreover medical art is entirely based on memory and experience without being in need of reason.5 While Aristotle would agree with the medical empiricists that sense perception or experience cannot reveal the causes of things, he firmly believes that we can grasp them by means of reasoning and intellect.6 Aristotle’s empiricism ought not to be confused with the kind of empiricism championed by John Locke or David Hume either. The latter were both antirationalistic and anti-metaphysical. Most importantly, they no longer shared the Aristotelian creed that a thing consists of matter and form. For Locke, as for Hume, the real essence of a thing (in Aristotle, eidos or ousia) is unknowable. Given these qualifications, Aristotle’s theory of knowledge is certainly more empirical in nature than Plato’s. This emerges, inter alia, from his criticism of Plato’s doctrine of innate knowledge. However, since Aristotle, as does Plato, holds that knowledge proper is not of the particular but of the universal, the crucial question in 4

5 6

That it is not an anachronism to speak about rationalism and empiricism in Antiquity is shown by Frede, M. (1990), esp. 225, and Allen (2001), 87–146. Both authors discuss the doctrines of the Greek medical empiricists. Regarding the latter see also Hankinson (2004). On Aristotle’s alleged empiricism, see now Herzberg (2010). See Frede, M. (1990) and Allen (2001), esp. 236. Consider his well-known distinction between things that are ‘more familiar to us’ and things that are ‘more familiar by nature’ (e.g., Physics I 1, 184a16–18). The former are identical with sensible particulars and their properties, while the latter are the true causes of things. Aristotle does not only assume that these causes exist, according to him we actually have the capability of discovering them. On this see Wieland (31992), 69–85.

‘How comes the mind to be furnished?’

3

Aristotle is how we can actually grasp the latter. Whereas for Plato the universal is recollected, for Aristotle it is acquired by induction (epag¯og¯e), which is, in turn, based on sense perception, memory, and experience.7 Induction is described in detail in the last chapter of the Posterior Analytics (II 19). It sets out from the common elements in particulars and ultimately arrives at a universal concept. Rather surprisingly, such an inductive process is already alluded to in Plato’s Phaedrus (249b–c). Should we infer from this that the accounts of Plato and Aristotle are similar in nature? It is here that Proclus comes into play. Scholars have suggested that the Neoplatonist would harmonize Plato and Aristotle in his theory of concept formation.8 Harmonism, as one might call it, is not only a modern phenomenon. Already in antiquity there is rich evidence for the existence of harmonizers.9 Diogenes Laertius, for instance, wrote that Aristotle was Plato’s most genuine or authentic pupil.10 We know that Iamblichus argued that Aristotle’s comparison of the soul with an empty writing tablet can be considered in agreement with Plato’s theory of innate knowledge.11 Another striking example is Asclepius, who writes in his Commentary on the Metaphysics that Aristotle’s criticism of the theory of Forms is not directed against Plato himself, but against those who have misunderstood Plato’s doctrine.12 In his recent monograph entitled Aristotle and other Platonists, L. Gerson maintains: [T]here is […] a baseline agreement among the Neoplatonists as to the lineaments of harmony. Disagreement about details does not change this.13

The harmony thesis, which Gerson himself advocates, seems to get further support from Proclus’ theory of concept formation. Scholars commonly distinguish between two types of concepts in Proclus, namely those concepts that are empirically acquired (sometimes called Aristotelian concepts) and recollected ones (sometimes called Platonic concepts).14 Accordingly, it is widely held that the first stage in the process of 7

8

9

10

11 12 13 14

For instance, Barnes (21993), 167, holds that Aristotle in An. Post. II 19 is putting forth an ‘empiricist thesis’. See now the balanced discussion in Herzberg (2010). Gersh (1978), 109 with note 147 (quoted below). Refer also de Libera (1996), 105–108, and (1999), 233–234 with note 80; Sorabji (2004a), 180–181, and (2004c), 138–139, who consider Hermias, In Phaedr. 171.8–25 an example of a harmonizer (but see below ch. II.5 and ch. VII.3.4). Gerson’s (2005) recent monograph tries to argue comprehensively for the harmony thesis. See also ch. V.1 below. On harmony between Aristotle and Plato in general, see also Karamanolis (2006) and the literature listed below ch. V.1. Diogenes Laertius V 1.6: (gn¯esi¯otatos t¯on Plat¯onos math¯et¯on). See also Ps-Plutarch, Placita 1.9.10; Stobaeus, Anthol. 1.12, and Hippolytus, Elenchos 1.20. I owe these latter references to G.E. Karamanolis (Crete). On this see the remarks in de Haas (2000), 169–170. Asclepius, In Met. 166.35–36. Gerson (2005), 16. On Gerson’s general approach, see Helmig (2009). For this distinction between two kinds of concepts, see Sorabji (2004a), 180–181; (2004c), 138–139, and (2006).

4

Introduction

concept acquisition is purely empirical in nature, while recollection commences only at the second stage. The distinction between two kinds of concepts is, for instance, reflected in the following statement by S. Gersh: Proclus never rejects the view that certain Forms are derived from sensation by a process of mental abstraction but simply the theory that these are the only ones.15

As I hope to show in what follows, this ‘received interpretation’ has to be modified considerably. The main shortcomings of the commonly held view on Proclus’ theory of concept acquisition arise because scholars have as yet not fully explored the evidence on (1) Proclus’ relation to Aristotle and (2) the status and role of concepts in Proclus. An investigation into the latter will entail a study of Proclus’ theory of knowledge insofar as it is of importance for his theory of concept acquisition. Having said this, let us come back to the distinction between two kinds of concepts (Aristotelian and Platonic) in Proclus. It is certainly correct that we can distinguish between different kinds of concepts in Proclus’ philosophy. However, as will turn out, in Proclus’ own theory of concept acquisition there do not figure concepts that are acquired empirically (i.e., derived from sense perception). Moreover, contrary to what Gersh suggests, Proclus does not endorse an abstractionism of any kind. In recent years, scholars have become more and more aware of the fact that Syrianus and Proclus are rather critical regarding Aristotle.16 Hence, it seems less plausible to argue that both Neoplatonists harmonize Plato and Aristotle. However, as far as Proclus is concerned, one problem has never been addressed satisfactorily: How is it that Proclus also makes use of ‘universals of later origin’ (husterogenes – the term usually refers to abstracted universals or universals ‘after the many’ [epi tois pollois, post rem / post res]) in his own philosophy?17 Does this not entail that he after all has adopted Aristotelian concepts? The problem can be solved, I argue below, if one distinguishes between two senses of husterogenes in Proclus, analogous to the two kinds of concepts discussed above.18 At first sight, the distinction between empirical and recollected concepts seems to be a problem that is only of interest in the scholarly debate on Proclus’ theory of concept acquisition. Yet, it ought to be emphasized that it touches a sensitive issue regarding the whole of the Platonic tradition. The issue at stake is: Where does perception 15

16

17

18

Gersh (1978), 109; italics are mine. Gersh, ibid., 109 note 147, identifies the abstracted concept with ‘the Peripatetic notion of form’ and remarks that ‘it is not immediately clear at what level these forms are held to subsist’. See, for instance, Steel (1984), (2003a), (2003b), and Helmig (2008), (2010). As far as Syrianus is concerned, see O’Meara (1989), Saffrey (1990b), Longo (2005), Dillon (2006), and Helmig (2009). Proclus, In Parm. II 767.10 (Steel), IV 894.6 (Steel), and IV 896.20 (Steel). See also Hermias, In Phaedr. 171.20. For such a distinction, see already Helmig (2008) and (2010).

‘How comes the mind to be furnished?’

5

end and thinking begin? Are we in need of recollection for basic forms of conceptual thought? In recent debates on Plato’s epistemology scholars seem to favour a reading according to which recollection is not in charge of rudimentary forms of concept acquisition. A prominent champion of such an interpretation is D. Scott, who defended it at length in his book Recollection and Experience. Plato’s theory of learning and its successors (1995).19 Regarding the Middle Platonist Alcinous a similar interpretation has been suggested.20 The question is crucial for every interpretation of the role of concept formation in Plato and the Platonists after him. We ought to address the issue, if we want to determine the role and significance of recollection for the process of learning and concept acquisition. Ultimately, I shall argue that even Platonists such as Alcinous, Porphyry or Hermias of Alexandria, who by unanimous consent are seen as philosophers who entertained Platonic alongside Aristotelian concepts, do not combine (and harmonize) Plato and Aristotle in their epistemology.

2. Survey of recent literature21 The following survey focuses mainly on the Neoplatonist Proclus and, to a lesser extent, on more general contributions to ancient theories of concepts and the attainment of knowledge. As S. Gersh noted some thirty years ago, ‘the growth of interest in Neoplatonism is largely a phenomenon of the period following the Second World War’.22 However, at the time when Gersh was writing the interest in Neoplatonism was by no means comparable to the developments we are witnessing today, where Neoplatonic studies have established themselves well in the field of ancient philosophy.23 Despite this fortunate development, Neoplatonic theories of knowledge and concept formation are still far from being treated exhaustively.24 On the contrary, one cannot

19 20

21

22 23

24

Scott (1995). See Sedley (1996a), 306ff., Sorabji (2004a), 88, and (2004d), 104, as well as Boys-Stones (2005), 209 note 8. For more recent literature on Proclus, see Steel et al. (2005), especially ch. V (‘Knowledge and Language’) and ch. V.2.c (‘Concepts and Recollection’). For a philosophical introduction into his thought, see now Helmig / Steel (2011) and Chlup (2012). Gersh (1978), 7. More recent research on Neoplatonism is summarized in Steel / Helmig (2004a), (2004b), and Steel / d’Hoine (2009); see also the ‘book notes’ on Neoplatonism in the journal Phronesis (in the fourth number of every year) as well as the International Journal of the Platonic Tradition (2007ff.). A notable exception is the interpretation of the Greek commentators on Aristotle’s distinction in De anima III 5 between active and passive intellect, which has been the object of intensive study.

6

Introduction

help but notice a certain negligence when it comes to epistemological questions. As far as Proclus is concerned, earlier studies on his philosophy were almost exclusively interested in his metaphysical system.25 The modern debate surrounding Proclus’ theory of knowledge and concept formation was instigated by two articles of W. Beierwaltes and H. Blumenthal (1975) respectively.26 The authors seem to share the view that, because of his ‘soteriological interests’, Proclus did not devote much attention to the lower faculties of the soul.27 Some twelve years after the contributions by Beierwaltes and Blumenthal, Th. Kobusch published a study entitled Sein und Sprache. Historische Grundlegung einer Ontologie der Sprache (1987).28 The author discusses several important Proclian passages on the status and origin of intentional objects (‘Gedankendinge’), in the course of which he also addresses Proclus’ criticism of Aristotle’s theory of abstraction in the Commentary on Euclid.29 It is the merit of Kobusch to have pointed out that Proclus’ discussion of concepts in the human mind is deeply anti-Aristotelian in its core. For Proclus all universal concepts (‘Begriffe’), including those derived from sense perception, ultimately depend on the innate knowledge of the soul.30

25

26 27

28

29

30

See Berger (1840); Simon (1845); Vacherot (1846). This also holds true for Zeller (51923), 834–890; Rosán (1949, 22009); Beutler (1957), and Reale (1989). Similarly, Beierwaltes’ (21979) monograph Proklos. Grundzüge seiner Metaphysik is mainly devoted to metaphysical issues, but it also contains an interesting section on the role of dialectic (pp. 240ff.). Siorvanes (1996) contains a long section on ‘Knowledge and the Levels of Being’ (114–206), but is ultimately interested, it seems, in Proclus’ theory of science (physics and astronomy). The many articles of J. Trouillard, listed in Scotti Muth (1993), are also characterized by a strong emphasis on Proclus’ ontology and metaphysics. Influenced by Trouillard is the (as yet unpublished) dissertation of MacIsaac (2005). Beierwaltes (1975) and Blumenthal (1975). The quote is a remark by Blumenthal, in the discussion after Beierwaltes’ paper, see Beierwaltes (1975), 190. Apart from this early contribution, Blumenthal has written other articles on Proclus’ epistemology, see Blumenthal (1982), (1989), and (1999). Blumenthal’s (1996) monograph too contains some interesting material. Kobusch (1987). This book has frequently been overlooked by scholars working on Proclus. It is not mentioned in Scotti Muth’s (1993) bibliography. The book treats Proclus only briefly, but it contains nevertheless important observations on central passages. There is of course a significant number of contributions on ‘mental entities’ in Ancient Thought (see, for instance, Silitti [1980]; Ebbesen [1981], [1986], [1990]; de Libera in de Libera / Segonds [1998], xlv–lxviii, as well as de Libera [1996] and [1999]; Perler [2001]; Caston [1999], [2001], [2002]), but none really focus on Syrianus or Proclus. See Kobusch (1987), 68–74, where he mainly discusses In Parm. IV 890.30ff. (Steel), but also the interesting passage In Remp. I 260.22–25. See, moreover, ibid., 406–408, where Proclus’ criticism of Aristotle’s theory of abstraction in the Commentary on Euclid is addressed. See finally MacIsaac (2005), ch. II and, more specifically, ch. II.ii. Kobusch (1987), 71: ‘Dies macht deutlich, daß nach Proklos alle Allgemeinbegriffe, auch die abstrahierten, auf ein apriorisches Wissen zurückzuführen sind, insofern wir durch die Sinnlichkeit allein nie zur Bildung allgemeiner Begriffe geführt werden.’

‘How comes the mind to be furnished?’

7

Two important articles on Proclus’ criticism of Aristotle’s doctrine of mathematical abstraction and of his general views on the status and origin of mathematical or geometrical objects were published by I. Mueller (1987 and 1990). These articles give a concise account of Proclus’ take on mathematical concepts, but do not discuss universal concepts in general. Moreover, Mueller focuses exclusively on the Commentary on Euclid.31 As regards Proclus’ theory of the formation of universal concepts, an important step forward was taken in the article by C. Steel entitled ‘Breathing thought. Proclus on the innate knowledge of the soul’ (1997).32 The author analyzes central passages from Proclus’ commentaries on the Alcibiades I and on the Parmenides emphasizing the importance of the soul’s innate knowledge (logoi or logoi ousi¯odeis) for concept formation.33 Steel succinctly outlines the character of the soul’s knowledge and points to the importance of ‘putting forth’ (probol¯e)34 and ‘articulation’ (diarthr¯osis) in the process of concept acquisition. As will become clear in the following pages, his article has been one of the main points of departure for my study and I am in many respects indebted to Steel’s findings. Investigations of concepts and theories of concept formation have their place in the philosophical debates on abstraction, induction, and the nature and status of universals. In recent years, important studies have been published that considerably deepen our knowledge on the history of these discussions.35 In 1999 L. Gerson pointed to the fundamental role of concepts in the Platonic tradition (in Plato, Aristotle, Alcinous, Plotinus, Proclus, and Damascius).36 He raises 31

32 33

34

35

36

See Mueller (1987) and (1990). On Proclus’ Commentary on Euclid see also the study by Schmitz (1997) and the review of the latter by Bechtle (1998) as well as the collection of essays edited by Lernould (2010). I myself have discussed Proclus’ criticism of Aristotle’s theory of concept formation in more detail in Helmig (2009) and (2010). Steel (1997). On these logoi see also Dodds (21963), 299–300; O’Meara (1989), 131–138, 156–160, 166–169, and 185–188, as well as Taormina (1993), 227–228 and 240ff., and Dörrie/ Baltes (2002b), 133 with notes. O’Meara (2001) interprets the logoi as ‘intentional objects’. See, finally, Kotzia-Panteli (2000) and Chlup (2011). This word is mostly rendered ‘projection’, but I shall argue below (ch. VII.2.4) that this is not appropriate in most cases. On the status of universals in Antiquity and the Middle Ages, see de Libera (1996). For the complex discussion on abstraction in Antiquity and the Middle Ages and for the respective terminology, refer to de Libera (1999). On abstraction in Aristotle, see the study by Happ (1971), 584–630, and Cherniss (1944), 73–80. On abstraction in the history of philosophy, see the useful summary by Schneider (1970). See, moreover, Aubenque (1971); Kobusch (1971), and Mueller (1990). Important material on abstraction, recollection, and universals can be found in Sorabji (2004a), (2004c), (2006), and (2010). Finally, there are some more recent studies that focus on the last chapter (II 19) of Aristotle’s Posterior Analytics: Biondi (2004) and Herzberg (2010), the latter being concerned with Aristotle’s theory of knowledge as a whole. Gerson (1999). On abstracted concepts in Porphyry, see Chiaradonna (2012). On the ‘starting points of knowledge’, see the monograph by Tuominen (2007).

8

Introduction

important questions about the relation between concepts and Forms. In 2002, P. Lautner’s article on doxa and phantasia in Proclus’ Commentary on the Timaeus was published.37 The author is primarily concerned with distinguishing doxa and phantasia from one another, yet his contribution also contains an interesting analysis of Proclus’ theory of doxa and its relation to innate knowledge (logoi), which, in turn, allows doxa to judge and correct sense perception. The main interest in the article by R. Sorabji entitled ‘Aristotle’s perceptual functions permeated by Platonist reason’ (2004)38 is that the author underlines the contrast between Plato / later Platonists and Aristotle concerning the role of sense perception. He conclusively shows that several later Platonists would hardly agree with Aristotle’s view on what sense perception can achieve. Rather, they emphasized the role of reason (or recollection) in the soul’s processing of sense data. In addition, Sorabji has contributed (as editor) the three volume sourcebook The Philosophy of the Commentators 200–600 ad which contains pertinent texts (with concise introductions) on universals, concepts, abstraction, induction and on Neoplatonic psychology in general along with two extensive articles on concepts and concept formation in ancient philosophy.39 Likewise seminal are the contributions by D. Sedley on Platonist interpretations of the Theaetetus that culminated in his important monograph on the dialogue: The Midwife of Platonism (2004).40 His analyses are fundamental for a correct understanding of the history of Platonic epistemology from Plato onwards and of the central role played by the Theaetetus in this respect. Apart from this, Sedley published an important and rather influential article on the epistemology of Alcinous, where the Middle Platonist is portrayed as combining an empirical approach towards knowledge acquisition with innatism.41 I conclude my bibliographical survey with the book Recollecting the Meno by H. Tarrant (2005) and R. van den Berg’s (2008) recent study on Proclus’ commentary on the Cratylus.42 Tarrant sets out his own reading of the dialogue (first part) and investigates the ‘Wirkungsgeschichte’ (influence) of the Meno in Aristotle, Middle and Neoplatonism (second part). In doing so, he reveals special interest in the Middle and Neoplatonic interpretation of the phrase logismos aitias (Meno 98a).43 37

38

39 40 41 42 43

Lautner (2002). Lautner’s mainly reacts against Blumenthal’s (1975) claim that Proclus would not clearly distinguish between the two faculties of the soul. On the whole debate, see also Watson (1982a). Sorabji (2004d). See also his rich papers on concept formation in the ancient commentators on Aristotle, which contain much material stemming from Syrianus and Proclus (Sorabji [2006] and [2010]). Sorabji (2004a–c), (2006), and (2010). See Sedley (1993), (1996a) and (2004). Sedley (1996b). Tarrant (2005) and van den Berg (2008). See also Marler (1992–1993), who discusses this issue in Proclus.

Structure and contents of this study

9

On the whole, Tarrant provides much material that is of interest, especially from the period of Late Hellenism (Cicero) and Middle Platonism (e.g. the Platonist adaptation and interpretation of the Stoic ‘common notions’, etc.). However, his treatment of Proclus and Neoplatonism in general remains rather sketchy, because it is clearly not his primary aim to reconstruct Proclus’ theory of recollection.44 Van den Berg’s study is relevant in that it contains several pertinent observations on the relation between words, thoughts (concepts), and Forms as well as on the principles of Proclus’ epistemology and his criticism of Aristotle. The author points out that, according to Proclus, Aristotle held the view that names are only by convention. In his commentary on Plato’s Cratylus, Proclus criticizes Aristotle’s view insofar as it is based on a wrong theory of how we acquire knowledge. On the whole, van den Berg’s study provides further evidence for the opposition maintained by Proclus against Aristotle’s views in epistemology.45

3. Structure and contents of this study This study falls into seven chapters and brings together issues connected to universal concepts, the acquisition of knowledge, and the relation between Plato and Aristotle in ancient theories of knowledge, as well as Platonist criticisms of Aristotle and Alexander of Aphrodisias. The whole study leads to and, in a certain sense, takes its inspiration from Proclus. In part, it can be read as an attempt to accurately describe his (and his teacher’s Syrianus) views on Plato and Aristotle and to unearth possible sources for Proclus’ theory of knowledge in the long period between the Stoics and Plotinus. Throughout this study, it has not been my aim to demonstrate the superiority of any one of the theories discussed. Rather, I have done my best to consider each theory in its own right, to point out advantages and disadvantages of the different approaches and to ascertain their respective presuppositions. This has entailed rethinking and reevaluating Neoplatonic criticisms of Aristotle and Alexander. The continuities and discontinuities in the debate from Plato to Proclus are of special interest in order to understand better how and why the latter’s complex epistemology could evolve out of Plato’s dialogues, as it were. 44 45

See Tarrant (2005), 175–178. Apart from van den Berg’s monograph, I shall briefly mention two more recent publications that exclusively focus on Proclus: E. Gritti’s work on dialectic in Proclus (Gritti [2008]), which contains useful material on epistemological issues, as well as M. Martijn’s (Martijn [2010]) book on the notion of physis and philosophy of nature. The latter highlights the continuity of theology and the science of nature in Proclus and lays out the methodological foundations (clearly inspired by Aristotle’s theory of science) of Proclus’ Commentary on the Timaeus (especially the prooemium).

10

Introduction

Chapter I takes its departure from the question ‘What is a concept?’ and proposes altogether eight criteria to mark off and characterize concepts in ancient philosophical debates. I substantiate the claim that concepts are highly relevant in ancient epistemology from Plato to Proclus. Crucial for my investigation is the distinction of different models of concept acquisition (abstraction, induction, collection, anamn¯esis) and different kinds of concepts. The chapter closes by discussing the relation between Forms and concepts and by addressing the issue of problematic concepts (that is, concepts whose acquisition is difficult to account for). Chapter II is devoted to Plato. For him learning is recollection (Meno 81d–e and Phaedo 74a). But how does recollection contribute to concept attainment? Does it make only ‘higher learning’ (D. Scott) possible or also much more rudimentary forms of knowledge acquisition? I start off by looking for traces of the existence and role of concepts in Plato and investigate the interrelation of concepts, Forms, and language, before turning to the dialogue Parmenides, a work that is generally neglected in the study of Plato’s theory of learning, although it contains, I will argue, a criticism of abstraction / induction and an interesting refutation of conceptualism (i.e., the view that Forms or universals are mere thoughts of the human mind). Then, an examination of the dialogues where Plato mentions recollection explicitly follows, namely the Meno, the Phaedo, and the Phaedrus. The last part deals with concepts and recollection in Plato’s later dialogues and addresses the issues of error and of the limits of recollection (problematic concepts, that is, concepts difficult or impossible to recollect). Chapter III examines Aristotle’s views on concepts and concept formation as a conscious reaction to Plato. The investigation takes its start from Aristotle’s criticism of Plato and an important, but frequently misunderstood text on the division of sciences (Met. E 1). The latter text, blurred by an unnecessary conjecture by Schwegler, is crucial for marking off physical and mathematical objects and their respective concepts. The remainder of the chapter falls into three parts: (1) Aristotle’s notion of abstraction (aphairesis), (2) the origin and status of mathematical concepts, (3) his theory of induction (epag¯og¯e). Abstraction is intimately connected with mathematical and geometrical concepts, while the process of induction is how we attain universal concepts (An. Post. II 19). I shall address the problematic points of both abstraction and induction with an eye to post-Aristotelian criticisms of the two methods. Regarding induction I propose to distinguish different domains in which the method is applied. The chapter ends with an attempt at reconsidering An. Post. II 19 (‘induction of first principles’) and a discussion of the interplay between sense perception, universal concepts, and intellect. Chapter IV casts some new light on the time after Plato and Aristotle (Alexander of Aphrodisias, Alcinous, Plotinus, Porphyry). The focus is threefold. Contrary to the received interpretation (Sedley, Sorabji, Boys-Stones), Alcinous will not be presented as a Middle Platonist who combined empirically derived concepts with recollection. It is shown that the overall evidence is far from clear and that much points to the con-

Structure and contents of this study

11

clusion that his epistemology is Platonic in character. The tension between empiricism and innatism, which can already be observed in Plato himself (see ch. II and Scott [1995]), is characteristic for the whole of the Platonic tradition and it raises fundamental systematic questions about the interplay between empiricism and innate knowledge. Alexander and Porphyry are usually considered abstractionists (Mueller [1990]). While this is true as far as Alexander is concerned, who elaborates Aristotle’s notion of abstraction and his theory of intellect, it is very different, I am going to argue, in the case of the Neoplatonist Porphyry. My argument involves a new assessment of the latter’s epistemology as it is to be found in the Commentary on Ptolemy. As with Alcinous, Porphyry’s doctrine should not be considered a combination of Aristotelian and Platonic theories of concept formation. Although Porphyry uses abstraction vocabulary, his approach is entirely Platonic in its core. Most importantly, the text is not concerned with concept formation, but displays a Neoplatonic reading of Ptolemy’s doctrine of the criterion. Despite the frequent use of the label ‘abstractionist’ for both Porphyry and Alexander, the difference between the two is also evident from Neoplatonic criticism of Alexander’s theory of concept attainment. In the last section, Plotinus is introduced as an important ‘Wegbereiter’ (predecessor or trailblazer) of Syrianus and Proclus in two respects: (a) he shares with the two Neoplatonists a theory of innate psychic logoi and (b) he already makes explicit use of them in his account of knowledge acquisition. I elaborate on E.K. Emilsson’s reading of Plotinus’ theory of sense perception, by arguing that it can be made more plausible if we connect it with Platonic recollection and the doctrine of the undescended soul. However, it is still somewhat unclear how the latter relates to the assumption that our soul contains innate logoi. Despite the importance of the Enneads for Proclus, telling differences remain, mainly due to a different psychological theory. Chapters V–VII concentrate on Syrianus’ and Proclus’ theory of knowledge and concept attainment. After summarizing their openly critical attitude towards Aristotle (in many areas of philosophy), and, more specifically, their criticism of theories of abstraction and abstracted concepts, together with explaining the twofold nature of the universal of later origin (husterogenes) in Proclus (chapter V), I turn to the importance of doxastic concepts (logoi doxastikoi) in the latter’s philosophy. This entails a brief discussion of sense perception, phantasia, and doxa and a survey on more recent views on doxa in Proclus. As will turn out, doxa can recognize sensible objects and judge sense perceptions because it has access to logoi (innate knowledge). Closely connected to this is a discussion on the nature and status of the so-called logoi doxastikoi and on the interplay between sense perception and doxa (chapter VI). Chapter VII is entirely devoted to Proclus’ theory of concept formation. It falls into three parts. Beginning with the relation between the soul and its innate knowledge, I subsequently address his systematization of Platonic recollection by distinguishing three elements in what one may call his ‘triad of recollection’: forgetting, articulation (diarthr¯osis) of preconceptions, and the ‘putting forth’ (probol¯e) of innate

12

Introduction

knowledge. While articulation is characteristically Stoic and had already in the period of Middle Platonism been combined with Platonic midwifery (in the Anonymous Commentary on the Theaetetus), probol¯e, which I argue must not be translated ‘projection’, seems to be a Neoplatonic innovation. In a last part, I employ various criteria to organize the elements of Proclus’ doctrine of concept formation: the role of concepts in the exegesis of Plato, recollection as an intentional act of the soul, different kinds of concepts and different stages of recollection, error, and problematic concepts. By means of such a systematic approach I shall, for instance, be able to explain a rather obscure passage in the Commentary on the Parmenides (IV 895.24–896.5 [Steel]). Structural similarities between my analyses of Plato and Proclus (especially as far as recollection, error, and problematic concepts are concerned) reveal the extent to which the Neoplatonist tried to systematize what he found in Plato. The study concludes by reconsidering the results against the background of the relation between Plato and Aristotle (chapter VIII).

What is a concept?

13

I. Concepts – (Ancient) Problems and Solutions This first chapter is a preliminary for the study of concepts in the epistemological theories of different ancient and late ancient philosophers. As I have already stated in the introduction, a main thesis of this book is that many phenomena that were discussed in ancient and late ancient debates around the acquisition and nature of knowledge can be described more adequately if we have recourse to the language of concepts. This claim will be substantiated in what follows. The current chapter falls into four parts. To begin with, I shall provide a general introduction to the nature of concepts (What are concepts?), before outlining the relevance of concepts in ancient epistemological debates. Then follows a short history of concept attainment in antiquity accompanied by an inventory of the relevant technical terms. Finally, I shall say something on the much debated relation between concepts, universals, and Platonic Forms (Ideas). In the first and third part, I aim to provide the prolegomena to an, as of yet, unwritten comprehensive lexicon of the language of concepts and concept attainment in antiquity. As will turn out, especially in the Neoplatonic tradition, there is a stock of expressions, taken directly from Plato, Aristotle, from the Stoics, or the Middle Platonic (anonymous) Commentary on the Theaetetus that recur time and again in the context of knowledge acquisition and concept attainment.

1. What is a concept? Against the background of modern philosophical debates on concepts,1 I propose several criteria to characterize concepts in ancient philosophical texts and more specifically in the late ancient Platonic tradition in order to mark off the object of this study.2

1

2

See, for instance, Margolis / Laurence (1999a), (1999b), (2008), and the useful survey by Swoyer (2006). More recently, there have been several articles devoted to the role and importance of concepts in ancient philosophy. See the literature quoted in the introduction (pp. 5–9 above).

14

Concepts – (Ancient) Problems and Solutions

In Antiquity, several words can stand for a concept: axi¯oma,3 arch¯e,4 eidos / genos,5 ennoia,6 koinai ennoiai,7 epinoia,8 no¯ema,9 enno¯ema,10 logos,11 katholou,12 katholou logoi,13 koinon,14 lekton,15 prol¯epsis.16 Some of these expressions can be qualified by means of adjectives such as doxastos / doxastikos (pertaining to doxa), husterogen¯es

3

4

5

6

7

8 9

10

11

12

13

14

15 16

Themistius, In de anima 104.1–2 talks about pr¯ota axi¯omata. See, moreover, O’Meara (2001), 118, and Syrianus, In Met. 89.30ff. These axioms are mentioned in Aristotle’s Metaphysics  3–7 (law of contradiction, law of the excluded middle). Syrianus, as with Proclus, holds that true axioms of science are innate (In Met. 91.3: $μ  ξ ¹ 4     $   ). See Longo (2005), 141–223, on such axioms in Syrianus. See Aristotle, An. Post. II 19, where he speaks about the pr¯otai archai of syllogism, and below ch. III.4. The terminology recurs in the ancient commentaries on An. Post. II 19 and in Syrianus’ and Proclus’ criticism of Aristotle, see below ch. V.1–2. In Neoplatonic authors, the innate knowledge of the soul (logoi or logoi ousi¯odeis, see below ch. VI.2.4–6 and ch. VII.1) can also be called ‘Forms in the soul’ (eid¯e en psuch¯ei), see Helmig (2008), 46 with note 69. According to the Neoplatonic axiom that ‘all things are in all things’ (panta en pasin, Proclus, Elem. Theol. § 103 with Dodds’ (21963) commentary ad loc.), Forms exist on all levels of reality. For the use of genos in this respect see Ammonius, In Isag. 68.25–69.3. On the term ennoia, which already occurs in Plato (e.g., Phaedo 73c and Philebus 59d), in Middle and Neoplatonism see Dörrie / Baltes (2002b), 128–130. See also the next note. On the Stoic common notions see Cherniss (1976); Gourinat (2000), and Dyson (2009). Chiaradonna (2007) studies the expression in Alcinous, Galen, Plotinus, and Porphyry. See, moreover, Themistius, In de anima 103.38:   ¹  λ  ; On the role and function of common notions in Proclus, see below ch. VII.1.4. Syrianus’ use of the term is investigated by Longo (2005), 182–201. See Barnes (2003), 40 note 74, and on ‘bare notions’ (psillai epinoiai) pp. 41–42. See already Parm. 132b–c (on whether Forms can be thoughts) and below ch. II.2.2. See also Proclus, In Parm. IV 890.30–906.2 (Steel) [= Proclus’ commentary on this passage]. However, usually the word no¯ema refers rather to an episode of thought (thought process) and not to a concept. At In Parm. IV 896.13 (Steel), Proclus explains enno¯emata as  $μ  φ   $ . The word enno¯ema was common in Stoic circles (see von Arnim’s index s.v.). Neoplatonists frequently use the adjective enno¯ematikos (together with eidos or genos) to refer to an abstracted universal or, more generally, a universal in the mind (Ammonius, In Isag. 68.25–69.3; Philoponus, In Cat. 58.20). See below ch. VII.7. On logos / logoi signifying the innate knowledge of the soul (from Plotinus onwards), see below ch. IV.3.2; VI.2.4–6, and VII.1. In Proclus, these logoi are called ‘essential reason-principles’ (logoi ousi¯odeis), because they constitute the essence of the soul (see Steel [1997]). As with koinon (see note 14), katholou can be used for the form in matter or the universal concept in the soul (see already Aristotle, An. Post. II 19, 100a16 und 100b2). The katholou logoi signify universal concepts in Syrianus (In met. 25.21–22; 53.5, and 88.26) and Proclus (In Parm. IV 896.23 [Steel] and In Tim. I 32.6). The word frequently refers to the common element in things or the form in matter. It can also refer to the abstracted universal in the soul, see Simpl., In Cat. 83.8–16. On lekta as concepts, see Caston (1999). On the Stoic notion of pre-conceptions (prol¯epsis) see Sandbach (1971b) and Gourinat (2000) and, most recently, Dyson (2009). Outside Stoicism the word does not seem to play an important role.

What is a concept?

15

(later-born or of later origin, that is, abstracted), ousi¯od¯es (essential), or enno¯ematikos (having the nature of a concept).17 These latter adjectives can, in turn, be used as nouns. Amongst this diversity of designations, the most important terms to pick out a concept are undoubtedly ennoia (koinai ennoiai), logos, and katholou. The latter signifies a universal in the broad sense of the word and especially the universal (concept) in the soul (Aristotle, An. Post. II 19). From Plotinus onwards, Neoplatonists refer to innate knowledge as logoi (reason-principles). The term ennoia goes back to Plato, was frequently used in Stoicism, and gained currency in Middle Platonism as referring to the innate content of the soul (lat. notio). It is only in the Middle Ages (e.g. in Ockham) that we find the term conceptum (‘something conceived’ [sc. by the mind], already Boethius uses conceptio in this sense) in the sense of concept (‘Begriff ’). In order to bridge the gap between the manifold vocabulary of the Greeks and the single term ‘concept’, it is necessary, first of all, to be clear about what is meant by the term. Roughly speaking, I take ‘concept’ to refer to an inner-mental or inner-psychic entity that is, to a certain extent, stable, permanent (persistent across time and transformations), objective, shareable and universal. Hence, a single sense impression, a single episode of thought, a judgement, or a dream does not qualify as a concept. However, an episode of thought, a judgement, or a dream may originate because of already existing concepts or preconceptions (imperfect concepts).18 It is important to note that especially the stability and universality of concepts lets us distinguish them from other psychic phenomena such as sense perception, imagination, and memory. Evidence for such a distinction can already be found in Aristotle who asserts both in Met. A 1 and An. Post. II 19 that universal concepts (katholou) originate on the basis of many similar sense perceptions, memories, and also experience. In this study, I shall focus on universal concepts in the Aristotelian sense of the word, namely ‘universal’ meaning ‘that which is such that it can be predicated of several things’ (De Int. 17a39–4). Concepts of individuals do not seem to be relevant in ancient epistemological debates, since knowledge is equated with universal knowledge.19 One difficulty in defining ‘concept’ is demarcating it from belief. On the one hand, concepts constitute beliefs.20 On the other hand, when we change our beliefs 17

18 19

20

On ‘concepts of later origin’, see below ch. V.2, on the ‘essential reason principles’ ch. VII.1. On the adjective enno¯ematikos see note 10. On the distinction between concepts and preconceptions, see below [A-6]. Of course, the fact that concepts are usually universal does not exclude that I have conceptual knowledge of individuals such as Socrates (the teacher of Plato [469–399 BC]) or Eric Robertson Dodds (an Irish classicist and former Regius Professor of Greek at Oxford [1893–1979]). Following the ancient terminology, I cannot have a katholou or a logos (in the sense of universal concept) of Socrates, but rather a notion (ennoia) or a memory image (phantasma). Further thoughts on this can be found in ch. VII.7. Swoyer (2006), 133: ‘Concepts are constituents of beliefs, desires, intentions and many other thoughts.’

16

Concepts – (Ancient) Problems and Solutions

about something, the concepts that figure in the beliefs usually change as well. Nevertheless, it seems that concepts are more fundamental, more basic, and more stable than beliefs.21 In what follows, I shall suggest different criteria to illustrate my use of ‘concept’ in the rest of this study. [A-1] Concepts are (fairly) stable mental entities A concept (e.g. of tree, dog, justice) is a fairly stable mental entity that (a) allows us to recognize a range of individual entities / instances / actions that are to be categorized under this concept and that (b) determines the conditions under which we use a certain word corresponding to this concept. This is a minimal requirement every concept is to fulfil. Several things are worth noting here: First, if we speak about concepts, we usually mean universal concepts or ‘universals’.22 Second, concepts have to be stable in order to differentiate them from sense impressions, memories, beliefs, experiences, or episodes of thought.23 Sometimes, a distinction is made between concepts and pre-conceptions (or conceptions). This distinction implies that pre-conceptions are only a preliminary phase in the process of concept attainment. Pre-conceptions are less developed or less articulated (less perfect) than concepts (see below [A-5]) and have not been well reflected upon. Calling concepts ‘mental entities’ is common practice with scholars.24 It avoids some difficulties that arise, for instance, from talking about abstract entities, but it also creates others. Saying that a concept is a mental entity entails that it inheres in individual souls or minds. Thus, it is confined to an individual person; it is, as it were, a ‘private’ possession. However, in order to qualify as a (general) concept it must be shareable [A-2] and link up with reality [A-3]. It has been argued, most notably by Frege,25 that concepts cannot be identified with mental entities; rather, they must be ‘senses’ of words (in the ancient debate, this comes close to the Stoic view, who posited lekta which likewise were identified with the meanings of words).26 It is immediately clear what the advantage of Frege’s claim

21

22

23

24 25 26

On the difficulty of distinguishing concepts and beliefs see again Swoyer (2006), 136: ‘[…] [T]here is often no precise point in our concepts where one more change in belief would rub out the old concept and give rise to a new one. On this view there is no useful distinction between change in concept and change in belief.’ On universals and universal concepts in ancient philosophy, see the excellent survey by Sorabji (2006). In Plato and Aristotle, this stability is referred to as ‘coming to a halt’ or ‘coming to a stand’ (Meno 97e–98a, Phaedo 96b, and An. Post. II 19, 100a6 and 100b2). Swoyer (2006), 133. I take the reference to Frege from Margolis / Laurence (1999b), 6–8. On lekta as meanings of words see Rist (1969), 152–154. Caston (1999) interprets lekta as concepts.

What is a concept?

17

is. Since mental entities are subjective, that is, situated in different human minds their objectivity and shareability can by no means be taken for granted. Mental entities are ‘subjective’, whereas senses are supposed to be objective. Hence, in Frege’s reading concepts would correspond to objective contents of concepts. However, against such a view, the following argument can be advanced.27 The fact that mental entities are ‘subjective’, in the above mentioned sense, does not preclude that they are shareable. Concepts are mental entities insofar as they are situated in different minds, but they can be objective insofar as they are shareable, that is, insofar as two or more people share the same or at least a sufficiently similar concept. The subjective character of concepts does not, therefore, result in conceptual idiosyncrasy. What has to be explained, of course, is why concepts are shareable. [A-2] Concepts must be objective and shareable (universal) Concepts are (fairly) stable mental entities. Being situated in different minds, they have to fulfil the requirement of objectivity and shareability. In other words, my concept of cat should agree to such an extent with the same concept of another person that communication and even scientific discourse about cats is possible. In ancient and late ancient debates on the status and acquisition of concepts several strategies can be distinguished to explain the shareability of concepts. The easiest way out is to argue that concept acquisition depends upon innate knowledge and that this knowledge is shared by all human beings. This seems to be a standard Neoplatonic answer that can, for instance, be found in Syrianus or Proclus. The second strategy, found according to some interpreters already in Aristotle, is to assume that the process of concept attainment works in a similar way for everyone. Both arguments for the shareability and objectivity of concepts have to deal with the problem of error [A-6]. There are, of course, many examples for less objective and less shareable concepts (conceptions or pre-conceptions). In the early Platonic dialogues (‘Tugenddialoge’), we find examples of people who possess certain conceptions or pre-conceptions of, say, virtue, justice, or piety, because of their biography, personal background, or profession. One aim of Socrates’ questioning is to uncover and eradicate the more subjective elements in these concepts / conceptions and to lead the interlocutor to an insight into the objective nature of certain values or virtues. Another example for more subjective concepts / conceptions can be found in small children who in the process of concept attainment go through certain stages. For instance, Aristotle mentions small children who at first call all men ‘daddy’ and all women ‘mommy’.28 Less objective and less shareable concepts can more appropriately be described as incomplete or imperfect concepts (see below [A-4] and [A-5]).

27 28

Margolis / Laurence (1999b), 7. Aristotle, Physics I 1.

18

Concepts – (Ancient) Problems and Solutions

To conclude, in spite of the objectivity / shareability requirement, there can be no doubt that concepts insofar as they are mental entities belonging to different individuals also contain idiosyncratic elements. If these idiosyncratic elements are not relevant to the content of the concepts, they do not conflict with the shareability criterion and they can, in most cases of human communication, be neglected. There are cases, however, especially during the process of concept attainment, where concepts are still incomplete or imperfect due to limited experience and / or insight (see [A-4] below). The process of concept attainment and learning should, ideally, provide the possibility to further develop an imperfect concept. [A-3] Concepts must link up with reality In order for concepts to be useful they must correspond to this world and to its elements. If concepts were not linked up with reality they would not teach us anything about the world around us. This criterion is the most fundamental requirement for any theory of concept attainment. For if concepts were purely mental and did not inform us about the world around us, they would be of very limited value. Therefore, concepts cannot be ‘purely mental’.29 Every theory of concept attainment should be able to explain why our concepts tell us something about reality. Hence, such a theory must also accompany a theory about the world. In Aristotle, concept acquisition is based on the fact that we are somehow able to extract or isolate the enmattered forms of things. And these enmattered forms, later called enhula eid¯e, constitute the essence of sensible particulars (see [A-5]).30 Accordingly, it is a crucial assumption of Aristotle’s theory that all human beings or all horses or all dogs share the same enmattered form (An. Post. II 19, 100a3–9). With the Neoplatonists, the link between our concepts and the world around us is established by the fact that (most of ) our concepts are copies or images of the transcendent Forms. And since the latter are at the same time the ontological principles of the sensible world, knowing the world by means of our (innate) concepts is knowing the world according to its causes.31 [A-4] Concept formation and learning In concept formation we encounter the phenomenon of incomplete concepts or concepts with more idiosyncratic than objective features. Through learning we can develop or articulate our concepts and make them more perfect, universal, and objective.

29 30 31

Swoyer (2006), 133. For Aristotle’s theory of concept attainment, see ch. III.4 below. This is the Aristotelian requirement for knowledge, see An. Post. I 2.

What is a concept?

19

‘Mastery of a concept can come in degrees.’32 To decide whether a concept is complete / perfect or not, requires a standard against which this can be judged (see [A-5]). This standard, however, might differ from context to context. We might distinguish between everyday concepts, that is, concepts that allow us to speak about certain phenomena in a more loose way, and developed scientific concepts that fulfil the objectivity / shareability requirement to the greatest extent possible (cf. [A-2]). The scope of the present investigation is not confined to scientific concepts in the strict sense, but identifies several kinds of concepts [see [A-7]). [A-5] Degrees in the mastery of concepts (complete and incomplete concepts) and the importance of ‘essences’ Already the term ‘concept formation’ implies that concepts can be formed and developed and thus perfected. Such a development, however, presupposes the existence of complete and incomplete concepts, that is, concepts that are perfect mental representations of certain things, and concepts that represent such things only to a certain extent. In other words, concepts can be measured according to their capacity to grasp ‘essences’. Modern research on concepts and concept formation is closely linked to work done in cognitive or developmental psychology. The question of what it is for someone to possess a concept is discussed against the background of empirical research. In Antiquity and the Middle Ages this was rather different. Plato, Aristotle, the Middle and Neoplatonists believed that certain kinds of things (substances) have an essence (ousia) and that this essence can be known. ‘Essence’ was identified with ‘what something (really) is’ or with a thing’s true nature. Plato had assumed that these essences exist independently from their sensible instantiations and are not mind-dependent entities. These essences (ousiai) were also called Forms (eid¯e, ideai). With Aristotle the ontological status of essences becomes problematic. On the one hand, he criticizes Plato’s assumption of the existence of Forms and his ‘separation’ (ch¯orismos) of forms from sensible particulars. On the other hand, he frequently talks about forms (eid¯e) in matter. On a certain view, these forms are for him the objects of science and philosophy (as abstracted universals). Consequently, they have to be eternal and unchanging. Philosophy aims at cognizing these essences of things. Later philosophers like Descartes still retained this notion of essence (substantia) and it is especially with the British empiricists that attempts were made to abandon ‘essences’. Against this background, we can understand why it is crucial to speak of complete and incomplete concepts in ancient philosophy. As soon as someone has discovered or perceived the essence of something, he can be said to possess a complete concept.

32

Swoyer (2006), 135.

20

Concepts – (Ancient) Problems and Solutions

We possess a complete concept of x if we have fully grasped the essence or nature of x. Once I assume that essences exist and that they can be known, I have a suitable criterion for the completeness / incompleteness of concepts. On the whole, the ancient view on concept formation is teleological, as it were, in that it aims at a definite end-point. Problems such as scientific progress do not play a role in the ancient debate. This is a crucial difference compared to modern discussions on the completeness or incompleteness of concepts. If I am a natural scientist believing in scientific progress I shall be rather reluctant to admit that there is something like complete knowledge of something (i.e., a complete concept).33 [A-6] Concept formation and error Concept formation, that is, the attainment of an adequate concept of something, is naturally bound to involve error. The same holds true for the application of already acquired concepts. This can have several reasons such as defective sense perception, wrong or deficient concepts or false judgements. Problems concerned with ignorance and error are challenging for every theory of concepts / concept formation.34 For our purposes it is sufficient to outline three main causes of how error may come about in concept attainment.35 First, we might possess a false or insufficiently developed concept of something. Hence the error occurs because of the nature of the concept. Second, the data provided by sense perception might not allow us to identify a given object correctly, although we have a correct concept of the object in question. The reason might be that the object is too far away or that there is a defect within a sense organ and so on. And thirdly, although we possess a correct concept of something and our sense perception is not impeded, our judgement that connects the sense perception to a concept that we have already acquired is mistaken. This might be the case, if we are not sober or if our faculty of judgment is otherwise distorted (because of an illness etc.). The latter two kinds of error do not concern concept attainment as such, but rather judgements about sense impressions that, in turn, depend upon concepts. It is obvious that every ancient theory of knowledge acquisition tries to account for how error can be avoided. We need a criterion of truth, a paradigm / normative case of knowledge that allows us to judge whether something is true or false.36 In Aristotle, for instance, this role is given to intellect that grasps its objects without error (An. Post. II 19, 100b5–14). In Stoicism, it is the cataleptic phantasia that is responsible for cor-

33

34 35 36

A good example of how scientific progress changes our views / concepts about something are theories of space and time before and after Einstein. See Margolis / Laurence (1999b), especially 21–23, 34–35, 47–48, and 55–56. For such a threefold distinction, see Delcomminette (2003). On the criterion of truth see Striker (1974) and (1990) and the volume edited by Huby / Neal (1989).

What is a concept?

21

rectly apprehending something.37 For Epicurus allegedly ‘all sensations are true’.38 In the Platonic tradition, it seems that in grasping the transcendent Forms we cannot err. For either we grasp them or we do not, but there is no room left for grasping them erroneously, as it were. A difficulty that is only insufficiently addressed among Platonists lies in the question of whether the theory of recollection allows for error. It goes without saying, however, that the problem is crucial for an evaluation of Platonist theories of concept formation.39 Finally, the fact that error is possible in the process of concept attainment seems to be a further argument for distinguishing between ‘complete’ and ‘incomplete’ concepts (A-5): Arguments from ignorance and error present compelling reasons to suppose that it’s possible to possess a concept without representing necessary or sufficient conditions for its application.40

[A-7] Different kinds of concepts Concepts can be classified according to their origin or according to their content and function. The former classification is more fundamental. It arises from the important distinction between concepts gained empirically and concepts that are recollected (or innate). The latter classification separates concepts that allow us to recognize objects (factual concepts) from concepts that provide knowledge of the essence or nature of the thing [cf. A-5]. The distinction between empirically acquired concepts and concepts that are recollected, sometimes referred to as Platonic and Aristotelian concepts, can be found in the scholarly literature.41 It is not as anachronistic as it may seem, since it can be traced back to passages in Neoplatonic authors where Plato’s and Aristotle’s (sometimes Peripatetic) theories of knowledge acquisition are contrasted.42 What is important to stress, however, is the fact that the Neoplatonists did not consider Aristotle’s theory of knowledge acquisition as proceeding merely empirically.43 For that would contradict Aristotle’s own statement that sense-perception in itself does not yield knowledge (An. Post. I 31). Rather, the Neoplatonists emphasized that in Aristotle’s account of the acquisition of knowledge sense perception plays a much more fundamental role than in Plato. Consider, for instance, the following text from Olympiodorus’ Commentary on the Phaedo: 37

38 39 40 41 42 43

On the cataleptic phantasia as a criterion of truth see the brief remarks below ch. VII.2.3 (i) as well as Frede, M. (1999). On this see Striker (1977); Taylor (1980); Asmis (1984) as well as Sedley (1989). On recollection and error, see below ch. II.7.2 and ch. VII.6. Margolis / Laurence (1999b), 22. See, for instance, Sorabji (2006) and below ch. V.2. See, for instance, Helmig (2007c), (2009), and (2010). On Aristotle’s empiricism, see also above (pp. 1–5) as well as most recently Herzberg (2010).

22

Concepts – (Ancient) Problems and Solutions We shall not agree with the Peripatos, which considers sense perception the beginning (or cause, arch¯e) of knowledge, for never can inferior and secondary things be the principles or causes of superior things. If we must adhere to what was said in the general course, that sense perception is the beginning of knowledge, we will say that it is the beginning not in the sense of an efficient cause, but in so far as it awakes in the mind the recollection of universals and plays the part of a messenger or herald, provoking the development of latent knowledge in the soul.44

The text contrasts the Peripatetic view that sense perception is the cause of knowledge with the Platonic view that it figures as a mere trigger for recollection. The contrast in question can also be described as between abstractionists (Aristotle, Alexander of Aphrodisias) and those who criticize theories of abstraction (Syrianus, Proclus). We possess several Neoplatonic texts that refer to Aristotle as being an abstractionist.45 In Syrianus and Proclus it can be observed that they criticize abstracted universals (called husterogen¯e) as being deficient and of less value than sensibles.46 A likewise fundamental distinction can be found in Proclus who, in his Commentary on Plato’s Timaeus, explicitly differentiates between factual concepts that tell us something about the ‘that’ (hoti) of a thing and concepts that reveal something about its nature (dihoti – ‘because’). The former he connects to the faculty of opinion (doxa), while the latter concepts belong to discursive reasoning (dianoia).47 In other words, what we have here are two types of concepts: first, concepts that allow us to identify or recognize certain objects and second, concepts that enable us to give scientific (or causal) explanations of certain phenomena. Such distinction goes back to Plato and was developed further by his pupil Aristotle, especially in his Posterior Analytics. According to Aristotle, a type-1 concept allows us to say that something exists (factual concept). Concepts of the second type, on the other hand, allow us to give an account of the phenomenon in question (its cause). The latter is, according to Aristotle and Plato, the requirement for scientific knowledge. Concept formation may describe the attainment of a type-1 concepts, but may also describe the development from type-1 concepts to type-2 concepts. This distinction is fundamental for our whole investigation. Let me give an example. Someone may be able to identify a horse, because he has learnt from others that this particular animal is called horse. However, this person does not necessarily know the essence or nature of a horse. Acquiring the word ‘horse’ 44

Olympiodorus, In Phaedonem § 4, 8, translated by L.G. Westerink: Pμ« ξ μ   φ  Ρ  "   )  P )  #  $%κ  '« κ ( ' · "#    %*  λ   $% λ ν (  ,   

. , ξ -  λ -« . *« '   *    λ $%κ ,- κ ( '  /«  '«, # " κ $%κ "% ³« ' , $

# ³« *1.  κ π #  3.%κ ,« $'     .  λ  $# .  λ .«  ,   κ π #  3.%κ ,« 5 κ   '.

45

This is most prominent in Syrianus’ commentary on the two last books of Aristotle’s Metaphysics. On the term husterogenes and on the two kinds of husterogenes-concepts (abstracted and recollected) in Proclus, see ch. V.2. Proclus, In Tim. I 248.11–13, 18–20; I 251.6–7. On the distinction between the two kinds of concepts, see below ch. VI.2.4 and ch. VII.3.6.

46

47

What is a concept?

23

does not entail being able to give a definition of horse or knowing its essence. Analogously, we may also consider the difference between the knowledge of an expert and a non-expert. After all, it is a truism to say that ‘at any rate a picture is observed differently by an expert and an inexpert person.’48 Appendix: Concepts and the harmony of Plato and Aristotle Given the scholarly distinction between Platonic and Aristotelian concepts [A-7], the question as to whether Platonists combined the two approaches of Plato and Aristotle in their theories of the acquisition of knowledge remains open to debate. The alleged fact that they do so has often been taken as another proof for harmonizing tendencies in (Neo)Platonism. It has often been claimed that Neoplatonists harmonized Plato’s and Aristotle’s theories of concept formation. As I have made clear in my introduction, this is a view that I consider misguided (even in the case of Alcinous or Porphyry). Scholars who endorse the harmonizing view usually argue that knowledge acquisition takes its start from Aristotelian concepts which are later transformed into Platonic concepts. A still widespread reading assumes that by means of the theory of recollection we can somehow make empirical concepts more perfect. Against such an interpretation it has to be said that we possess little if any textual evidence that a Neoplatonist held such a view.49 To the contrary, Proclus and his teacher Syrianus, for instance, who criticize theories of abstraction, argue that empirical concepts cannot be made more perfect. At In Parm. IV 893.15–18 (Steel), Proclus explicitly says that the objects of phantasia, that is the sense-impressions stored in phantasia, cannot be perfected: The concept which is taken from sense-perception is an object of phantasia […] and it has to remain in the soul such as it was originally received, so that it does not become false or something that is not; but it cannot become more perfect or more worthy of respect.50

This passage clearly excludes the possibility that an object of phantasia can be made more perfect than a perception.51 Moreover, Syrianus draws a strong contrast between 48

49

50

51

Diogenes Laertius 7.51 (= SVF 2.61). See Annas (1992), 81 (on the Stoics): ‘Different people, then, will have different perceptual beliefs when faced by the same objects, because they have different thoughts which reflect their different degrees of understanding of what is given them in the appearances.’ Pace Lautner (2009), 382–383 (on Hermias and Syrianus). Note that Porphyry, In Ptol. Harm. 14.2 (where he employs akribo¯o) cannot be used as evidence for such a view, because there phantasia is said to make a memory image resemble its sensible counterpart; see below ch. IV.2.7. In Parm. IV 893.15–18 (Steel): μ    , '  5  φ     λ

"  ,  λ   7φ*  # , 8  φ' κ $%, 9 κ 3« ; 9  λ μ κ