The Thirteenth-Century Notion of Signification: The Discussions and Their Origin and Development 9004298673, 9789004298675

In "The Thirteenth-Century Notion of Signification," Ana María Mora-Márquez presents an exhaustive study of th

787 203 792KB

English Pages VI+184 [192] Year 2015

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

The Thirteenth-Century Notion of Signification: The Discussions and Their Origin and Development
 9004298673,  9789004298675

Table of contents :
Introduction 1
Part 1. Signification of Concepts and Signification of Things
1. Ancient Sources 11
1.1. Aristotle's "Perihermeneias" 11
1.2. Boethius' Second Commentary on the "Perihermeneias" 21
2. Medieval Discussions about Signification of Concepts and Signification of Things 36
2.1. Whether Words ('Qua' Names) Signify Concepts or Things 36
2.2. Whether Words Lose Their Signification with the Destruction of Their Significate 76
Part 2. Signification in Logic and in Grammar
3. Names and Verbs in Priscian and in Aristotle 109
3.1 Priscian on the Constitution of Parts of Speech and Sentences 109
The Role of the Notion of Signification in the Division and Order of the Parts of Speech 111
Names and Verbs and Their Construction according to Priscian 115
3.2. Names and Verbs in Aristotle's "Perihermeneias" 118
4. The Role of the Significate ('significatum') in Grammar and in Logic 123
4.1. The Pre-Modist Tradition 123
4.2. The Modist Tradition 139
Conclusion 157
Bibliography 163
Index of Subjects 176
Index of Modern Authors 180
Index of Ancient and Medieval Authors 182

Citation preview

The Thirteenth-Century Notion of Signification

Investigating Medieval Philosophy Managing Editor John Marenbon

Editorial Board Margaret Cameron Simo Knuuttila Martin Lenz Christopher J. Martin

VOLUME 10

The titles published in this series are listed at brill.com/imp

The Thirteenth-Century Notion of Signification The Discussions and Their Origin and Development By

Ana María Mora-Márquez

LEIDEN | BOSTON

 Library of Congress Cataloging-in-Publication Data Mora-Márquez, Ana María.  The thirteenth-century notion of signification : the discussions and their origin and development / by Ana María Mora-Márquez.   pages cm. -- (Investigating medieval philosophy, ISSN 1879-9787 ; VOLUME 10)  Includes bibliographical references and index.  ISBN 978-90-04-29867-5 (hardback : alk. paper) -- ISBN 978-90-04-30013-2 (e-book) 1. Reference (Linguistics) 2. Signification (Logic) I. Title.  P325.5.R44M66 2015  401’.43--dc23 2015020657

This publication has been typeset in the multilingual “Brill” typeface. With over 5,100 characters covering Latin, ipa, Greek, and Cyrillic, this typeface is especially suitable for use in the humanities. For more information, please see www.brill.com/brill-typeface. issn 1879-9787 isbn 978-90-04-29867-5 (hardback) isbn 978-90-04-30013-2 (e-book) Copyright 2015 by Koninklijke Brill nv, Leiden, The Netherlands. Koninklijke Brill nv incorporates the imprints Brill, Brill Hes & De Graaf, Brill Nijhoff, Brill Rodopi and Hotei Publishing. All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher. Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill nv provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, ma 01923, usa. Fees are subject to change. This book is printed on acid-free paper.

Contents Introduction 1

Part 1 Signification of Concepts and Signification of Things 1 Ancient Sources 11 1.1 Aristotle’s Perihermeneias 11 Perihermeneias 1.16a3–8: Utterances as Symbols and Signs 13 Perihermeneias 1.16a9–18: Simple and Compound Linguistic Items 18 Contradiction and Equivocation 20 1.2 Boethius’ Second Commentary on the Perihermeneias 21 The Subject Matter of the Perihermeneias according to Boethius 22 Boethius on Perihermeneias 1.16a3–8 23 Concepts and Likenesses in Boethius 31 2 Medieval Discussions about Signification of Concepts and Signification of Things 36 2.1 Whether Words (Qua Names) Signify Concepts or Things 36 The Semiotic Angle 41 The Immediate Signification of Concepts 41 The Modist Rejection of the Immediate Signification of Concepts 52 Roger Bacon and Peter John Olivi’s Rejection of the Immediate Signification of Concepts 61 The Verificational Angle 70 2.2 Whether Words Lose Their Signification with the Destruction of Their Significate 76 Anonymus Alani’s ‘Omnis Homo De Necessitate Est Animal’ (Paris bnf Lat. 16135, ff. 99rb–103vb) 80 Roger Bacon’s De Signis iv.2 87 Boethius of Dacia’s Sophisma ohnea 92 Peter John Olivi’s Quaestiones logicales q.3 95 Anonymus Alani’s Solution to the Sophisma ohnea 98

vi

Contents 

Part 2 Signification in Logic and in Grammar 3 Names and Verbs in Priscian and in Aristotle 109 3.1 Priscian on the Constitution of Parts of Speech and Sentences 109 The Role of the Notion of Signification in the Division and Order of the Parts of Speech 111 Names and Verbs and Their Construction according to Priscian 115 3.2 Names and Verbs in Aristotle’s Perihermeneias 118 4 The Role of the Significate (significatum) in Grammar and in Logic 123 4.1 The Pre-Modist Tradition 123 4.2 The Modist Tradition 139 Conclusion 157 Bibliography 163 Index of Subjects 176 Index of Modern Authors 180 Index of Ancient and Medieval Authors 182

Introduction Hilary Putnam begins his article ‘The Meaning of “Meaning”’ by claiming that modern linguistic research has greatly illuminated the syntax of natural languages, but has left their semantic aspect as obscure as ever: Analysis of the deep structure of linguistic forms gives us an incomparably more powerful description of the syntax of natural languages than we have ever had before. But the dimension of language associated with the word ‘meaning’ is in spite of the usual spate of heroic if misguided attempts, as much in the dark as it ever was.1 According to Putnam, it is the traditional formulation of the problem in terms of ‘meaning’ that is responsible for this dismal progress in semantics: Just as the modern reformulation of the questions has improved the syntactic analysis of linguistic structures, so too has the poor state of the notion of ‘meaning’ perpetuated the obscurity around the semantic dimension of human language. In my opinion, however, the poor state of modern semantic enquiries is mainly due to an ambiguous notion of meaning. In fact, modern linguistic enquiries intend to answer fundamentally different questions in terms of ‘meaning’. Meaning plays an explanatory role in, at least, these functional aspects of language: first, communicative – how language transmits information; next, semiotic – what is a linguistic sign; finally, descriptive – how language accurately or truly represents the objective world. In other words, ‘the dimension of language associated with the word “meaning”’ is ‘as much in the dark as it ever was’, quite likely because there are not one but several dimensions of language associated with the notion of meaning. It seems this diagnosis also applies, with the necessary qualifications, to the medieval notion of signification (significatio). For the Latin verb ‘significare’ and its derivative noun ‘significatio’ involve at least the same ambiguity that underlies our modern notion of meaning. The medieval notion plays indeed different roles in some ancient traditions with quite different agendas. I intend to show that the different roles this notion plays in those ancient traditions are a determining factor in crucial discussions about significatio in thirteenthcentury linguistic literature. Yet, the analysis of the three most important ­thirteenth-century discussions about significatio reveals an increasingly coherent 1 H. Putnam, ‘The Meaning of “Meaning”’, in Mind, Language and Reality. Philosophical Papers. Vol. 2 (Cambridge: cup, 1975), pp. 215–271, at p. 215.

© koninklijke brill nv, leiden, ���� | doi 10.1163/9789004300132_002

2

Introduction

treatment of this notion: Towards the end of the thirteenth century, the notion plays a central role only in accounts of the institution of linguistic signs and next to no role in logical and grammatical accounts. In the early thirteenth-century, linguistic problems typically arise from the encounter of the Aristotelian Organon with other linguistic traditions, notably with Boethius’ logical corpus and with Priscian’s grammar. Most of these problems revolve around the role significatio plays in the institution of linguistic signs, the grammaticality of sentences and the truth or falsity of assertions, as  well as in the question of linguistic communication. In a span of about sixty years, these discussions lead us to the approach to signification of late ­thirteenth-century masters of Arts, such as Boethius of Dacia and Radulphus Brito. These authors, as we shall see, put forth a notion of signification that is disconnected from the analyses of grammaticality and truth or falsity, and that finally retains a central role only in the institution of simple linguistic items: in impositive grammar (grammatica impositiva). We shall see that the notion of signification evolving beyond medieval logic and grammar is the common feature these three discussions share, thus showing a tendency towards a more coherent use of the notion, which remains fundamental for medieval semiotics. This study should also incidentally show the limits of equating the medieval notion of signification with the modern notion of meaning. Thus, we should notice that thirteenth-century approaches to signification should not be considered as a stage in the historical development of the notion of meaning, but rather as part of a set of linguistic enquiries that arose in a scholarly context that has no exact counterpart in modern linguistic analyses. This last remark explains my heavily historical approach to these discussions.2 Rather than analyse from the point of view of modern problems, I intend to reconstruct the medieval philosophical problems as they were engaged at that time, using mostly the theoretical tools provided by the medieval philosophical background itself. Hence, I intend to reveal the medieval problems and their medieval solutions in their own doctrinal context, with the conviction that precisely the medieval way of positing and treating linguistic problems presents an alternative way of approaching human language that is philosophically interesting in itself.

2 For a collection of articles analysing different methodological approaches to the history of medieval logic, see M. Cameron and J. Marenbon (eds.), Methods and Methodologies. Aristotelian Logic East and West, 500–1500 (Leiden: Brill, 2011) (Investigating Medieval Philosophy 2).

3

Introduction

The medieval way is remarkable for at least three reasons: First, it describes and tries to regulate the use of ordinary human language, aiming to navigate its natural pitfalls in a well-defined set of linguistic situations (namely, teaching, textual interpretation and philosophical disputation). Second, it analyses human language mindful of its use in these particular situations – insofar as it is the singular tool for production, communication and interpretation of human knowledge. Last but not least, based on careful analyses of the linguistic problems that are to be engaged, it intentionally refines theoretical tools fit to engage those problems and attempts to fix the tools that were diversely used in e.g. the authoritative ancient literature. As we shall see, all of these aspects are involved in the thirteenth-century development of the notion of signification. *** Signification is a key notion in early thirteenth-century logic, which was mainly studied on the basis of some of Boethius’ logical opuscula3 and his translations of Aristotle’s Categories and Perihermeneias and of Porphyry’s Isagoge. In the academic context of the medieval faculties of Arts, Boethius’ logical works and translations had to interact with other linguistic treatises where the notion also played a central role and that had rich and long traditions. Signification is, in fact, also central in grammar, which in the same period was mainly studied on the basis of Priscian’s Institutiones grammaticae.4 Even within the medieval logical corpus, diverse distinctions are made based on this notion. For instance, the ten Aristotelian categories (substance, quantity, quality, relation etc.) follow a distinction of ten kinds of things that predicates signify, and the same goes for the distinctions between name and verb in the Perihermeneias and between the five Porphyrian predicabilia (genus, species, proper, difference and accident) in the Isagoge. 3 Boethius’ logical opuscula include his commentaries on Aristotle’s Categories and Perihermeneias and on Porphyry’s Isagoge, De divisione, De topicis differentiis, De syllogismo categorico and De hypotheticis syllogismis. The last two treatises, though, were not part of the academic program of the 13th-century universites of Paris and Oxford. 4 Donatus’ Ars grammatica was also popular, but for my concerns Priscian’s is by far the most influential grammatical treatise. For Donatus’ influence on the Middle Ages, see S. Ebbesen, ‘Theories of Language in the Hellenistic Age and in the Twelfth and Thirteenth Centuries’, in D. Frede and B. Inwood (eds.), Language and Learning: Philosophy of Language in the Hellenistic Age (Cambridge: cup, 2005), pp. 299–319 (repr. in S. Ebbesen, Collected Essays of Sten Ebbesen, 2 vols., Aldershot/Farnham: Ashgate, 2008–9, vol. 1, pp. 79–98).

4

Introduction

Boethius makes us aware of a first conflict in the ancient traditions of the Categories and the Perihermeneias, by pointing to an opposition between an immediate signification of things in the Categories and an immediate signification of concepts in the Perihermeneias. Consequently, medieval commentators on the Perihermeneias start to discuss the problem of what the immediate signification of words is – concepts or things – as early as the 1240s (in e.g. Nicholas of Paris’ commentary). The brief treatment of this problem by master Nicholas and his contemporaries develops, in the second half of the century, into a sophisticated debate that transcends the faculties of Arts.5 At the turn of the fourteenth century, John Duns Scotus describes it (in his Ordinatio of 1304) as a magna altercatio,6 a debate where at least three different linguistic problems are involved: a semiotic problem, a problem of linguistic communication and a problem of propositional verification. The quarrel about the immediate signification of words is closely related to other questions that deal with the consequences of advocating for an immediate signification of things. The most important for our purpose are the variants of the question ‘whether a word loses its signification with the destruction of the things it signifies’; a question at the centre of which are a semiotic problem and a problem of propositional verification. In the first part of this study, we will examine these two medieval debates (in sections  2.1 and 2.2), together with an introductory chapter (Section 1) about their ancient sources, namely Aristotle’s Perihermeneias and Boethius’ second commentary on the same treatise.7 Next, the interaction between Priscian’s grammar and Aristotelian logic raises a second conflict between grammaticality and truth and falsity. Medieval authors had to deal with two different accounts, one logical and one grammatical, of parts of speech (partes orationis) and, particularly, of names and verbs (nomina and verba). Indeed, the definitions of names and verbs given by Priscian in his Institutiones and those given by Aristotle in his Perihermeneias, both of them in terms of what and how they signify, differ in substantial respects. Consequently, questions about the different approaches to parts of 5 For an analysis of this development, see also A.M. Mora-Márquez, ‘A New Perspective on the Origin of the Debate on Signification at the End of the 13th Century’, in C.T. Thörnqvist and B. Bydén (eds.), Papers on Aristotelian Logic and Metaphysics Presented to the Danish-Swedish Network for the Aristotelian Tradition in the Middle Ages (2009–2011) (Toronto: pims, forthcoming). 6 Cf. Duns Scotus, Ordinatio, d. 27, q.1. 7 For the late ancient and medieval reception of Aristotle’s Perihermeneias, see M. Cameron and J. Marenbon (eds.), Aristotelian Logic East and West, 500–1500: On Interpretation and Prior Analytics in Two Traditions, Vivarium 48.1–2 (2010).

5

Introduction

speech in Aristotelian logic and in Priscian’s grammar are ubiquitous in commentaries on the Perihermeneias and on Priscian’s Institutiones. More precisely, medieval authors raise a question about the different roles played by the notion of signification in grammar and in logic; a question that we shall analyse, including its ancient sources Priscian and Aristotle, in Part 2. The three questions studied here testify to a somewhat common origin and development of thirteenth-century discussions about significatio: They all start by opposing certain positions of earlier authorities and develop into sophisticated linguistic discussions, where we can perceive the underlying intention, by remarkable authors, of giving the notion a more coherent treatment. This coherence, as we shall see, is achieved by letting signification play an important role only in the explanation of the institution of simple linguistic items and by radically reducing its use in the solution of logical and grammatical problems. *** Let us consider the corpus I use, the time frame it covers and the authors and texts it includes (and consequently the ones left out). For the sake of narrowing the research to a manageable and reasonable time frame, I have studied the logical tradition that began with the founding of the University of Paris in the early thirteenth century, until the emergence of nominalism in the first quarter of the fourteenth century. Thus, twelfth-century literature is included only when its influence on the thirteenth century is undeniable, as in the case of  Peter Helias’ Summa grammaticae. Regarding early thirteenth-century ­commentaries on the Ars vetus (i.e. the Isagoge, the Categories and the Perihermeneias), the direct and main source is Boethius’ logical corpus, and there is very little in the early commentary tradition that cannot be related in one way or another to Boethius’ works.8 This is the principal reason for not taking into consideration the logical literature from the twelfth century. Given the richness of the thirteenth-century scholarly corpus, my research will focus on texts and discussions that emerged in and stayed within the faculty of Arts, with the notable exception of the Franciscan theologians Roger Bacon and Peter John Olivi, who take a stand on the artistic discussions in a couple of remarkable linguistic treatises – the De signis and the Quaestiones logicales, respectively. I take into account the most important edited material 8 For a partial example, see the introduction in A.M. Mora-Márquez, ‘Anonymus Oxford, Commentary on De interpretatione 1 (ms Oxford, BodlL Can. misc. 403, ff. 31ra–34vb)’, cimagl 83 (2014), 135–206.

6

Introduction

that was produced in the faculties of Arts of the universities of Paris and Oxford and that contains one or more of the three questions that are analysed in Parts 1 and 2 – the three questions that, to the best of my knowledge, embrace problems directly associated with the notion of signification. I also take into account unedited material I regard as most influential and relevant (e.g. Nicholas of Paris’ and Robert Kilwardby’s commentaries on the Perihermeneias). Finally, for the second half of the century I focus on masters that represent important points of dissent and/or development, such as Martin of Dacia, Boethius of Dacia and Radulphus Brito, and leave aside or briefly mention other masters, such as Siger of Brabant and Simon of Faversham, whose ­positions do not command distinction. Thomas Aquinas’ commentary on the  Perihermeneias9 is also left aside, for it neither clearly influenced the ­thirteenth-century commentary tradition, nor took a noteworthy stand on the discussions with which this study is concerned. Furthermore, important discussions such as the ones about the possibility of naming God and the mechanisms of angelic communication are left aside, principally because of their theological nature. The theological material has been thoroughly and extensively studied and discussed by prominent scholars to whose work I refer the interested reader.10 Finally, since Augustine’s treatises were not part of the faculty of Arts’ curriculum, I do not include him in the presentation of ancient sources. However, I do briefly present his accounts when I introduce Roger Bacon and Peter John Olivi – the only two authors of the corpus whose logical discussions are beyond doubt influenced by Augustine’s linguistic accounts. *** 9

10

Thomas Aquinas, Expositio in librum Perihermeneias, ed. R.-A. Gauthier (Rome/Paris: Edition léonine/Vrin, 1989) (Opera omnia 1*/1). For the commentary’s relation with the 13th-century commentary tradition of the Perihermeneias, see J. Isaac, Le Perihermeneias en occident de Boèce à Saint Thomas (Paris: Vrin, 1953). In this book I shall focus on linguistic literature, but the notion of signification played also a role in other contexts. For instance, medieval readers and commentators on Peter Lombard’s Sentences had to discuss the problem of divine names when reading the distinctions xxi–xxvi of book I, where the Lombard raises questions about the possible ways of talking about God. In order to tackle these questions, theologians not only appealed to Aristotelian notions, but they also let these notions interplay with theological sources, of which the most influential are undoubtedly Augustine’s De doctrina christiana and De magistro, and (Pseudo) Dionysus the Areopagite’s On Divine Names. For this aspect of the discussion, see the work (see bibliography) of John Marenbon and Luisa Valente for the 12th century and of Jenniffer Ashworth and Irène Rosier-Catach for the 13th century.

7

Introduction

Although this study is mainly intended for historians of medieval philosophy, I strove to make it accessible to linguists, logicians and philosophers of language interested in this rich period in the history of linguistic enquiries. Accordingly, in the philosophical analyses I try to minimise my use of technical language, from the point of view of both modern and medieval terminology. In cases where use of medieval technical language proved unavoidable, I provide a brief introduction to the terms needed to properly understand the subsequent discussion. Translations are also provided (my own, unless otherwise indicated) of Latin and Greek passages essential to the analysis, with the Latin or Greek text in the footnotes. *** This monograph is the result of ten years of research and it was written with the financial support of the Danish Carlsberg Foundation. At a first stage of this research period (2004–2009), I was a Master and PhD student of Prof. Annick Jaulin and Prof. Christophe Grellard, with the financial support of the Ecole Doctorale de Philosophie, at University Paris 1, Panthéon-Sorbonne. Later (2010–2013), I was a postdoctoral researcher at the Saxo Institute (University of Copenhagen), under the supervision of Prof. Sten Ebbesen, with the financial support of the Carlsberg Foundation. During all these years I be­nefited in various ways from discussions with many friends and colleagues. My teachers Annick Jaulin, Christophe Grellard and Sten Ebbesen have given me their unconditional support and sensible advise throughout all these years and to them I owe a special debt of gratitude. Jocelyn Benoist, Joël Biard and Irène Rosier-Catach – opponents at my PhD defense – gave me a number of criticisms and suggestions that were fundamental for the improvement of the first stage of my work. Karin Margareta Fredborg, Chris Martin and Steffen Lund Jørgensen carefully read and commented upon a first version of this monograph that was discussed at a workshop in Copenhagen. Mary Sirridge and Paul Thom read parts of this monograph and provided me with great advise for its improvement. I also benefited from many good ideas thanks to informal discussions with Pierre Pellegrin, Michel Crubellier, Costantino Marmo, Claude Panaccio, Jenny Ashworth, Paolo Crivelli, Jean Baptiste Gourinat, David Bloch, Julie Brumberg-Chaumont, Aurélien Robert, Iacopo Costa, Catarina Dutilh Novaes, Sara Uckleman, Rodrigo Guerizoli, Alfredo Stork, Frédéric Goubier, Ernesto Perini Santos, Juliette Lemaire, Valeria Buffon, Gustavo Fernández Walker, Leone Gazziero and Taki Suto. An anonymous referee was the fairest and sharpest reader an author could hope for. John Marenbon, the editor of this series, was of great help during the process of

8

Introduction

submission and publication. The head of my department at the University of Gothenburg, Christina Thomsen Thörnqvist, has given me all the moral and financial support needed to bring this monograph to good term. Finally, Debbie Axlid and Kirez Reynolds did a tremendous work of revision of the English prose. To all these people I owe all my gratitude. Last but not least, I say again many thanks to my partner Søren, who during all these years never failed to be there for me, at times with remarkable stoicism.

PART 1 Signification of Concepts and Signification of Things



chapter 1

Ancient Sources 1.1 Aristotle’s Perihermeneias Aristotle’s Perihermeneias is generally considered to contain his account of signification, specifically its first chapter. Thirteenth-century scholars, who followed Boethius’ lead, were also of this opinion. Therefore, this chapter exerted a strong influence on their discussions on signification and hence we begin this study with a brief presentation of Perihermeneias 1. Most modern interpreters hold that Perihermeneias 1 deals with the question of how language represents the world.1 The idea has often been presented as if Aristotle started his inquiry by assuming that language is perfectly univocal and proceeded to describe how this language accurately represents the world through its representation of thoughts.2 It is no surprise then that Aristotle is considered to be the origin of a long history of misguided attempts to explain how language has univocal signification and reference.3

1 Cf. J.L. Ackrill, Aristotle. Categories and De interpretatione (Oxford: Clarendon Press, 1963), p. 113; L.M. de Rijk, Aristotle’s Semantics and Ontology. Volume I: General Introduction. The Works on Logic (Leiden: Brill, 2002) (Philosophia Antiqua XCI/I), p. 192; C.W.A. Whitaker, Aristotle’s De interpretatione. Contradiction and Dialectic (Oxford: oup, 1996), p. 8; P. Aubenque, ‘Sens et unité du traité aristotélicien De l’interprétation’, in S. Husson (ed.), Interpréter le De interpretatione (Paris: Vrin, 2009), pp. 37–50, at p. 41. 2 For instance P. Aubenque says [in ‘Sens et unité’, at p. 42 and p. 44]: ‘Ce sera un problème largement débattu au Moyen Âge de savoir si les fictions comme la chimère…ont une essence, et dès lors, sont un certain type d’êtres…Aristote anonce ce problème, mais n’en débat pas ici. Car il pare au plus pressé et le plus urgent pour lui…est de fixer des règles techniques qui permettront de traiter le langage comme s’il était parfaitement univoque, transparent, et n’avait pas bésoin d’être interprété’; and ‘En vérité, le traité se demande à quelles conditions le langage peut s’éléver au dessus de sa fonction symbolique – l’annuler si possible – pour devenir comme la représentation, une image de la chose, un décalque de la réalité, et autoriser ainsi une conclusion qui va des mots aux choses.’ For a similar claim, see also N. Kretzmann, ‘Medieval Logicians on the Meaning of the Propositio’, Journal of Philosophy 67 (1970), 767–787, at p. 768. 3 Cf. H. Putnam, ‘Meaning, Other People and the World’, in Representation and Reality (Cambridge ma: mit Press, 1988), p. 19.

© koninklijke brill nv, leiden, ���� | doi 10.1163/9789004300132_003

12

chapter 1

While it is true that the Perihermeneias relies on an implicit notion of signification,4 in fact it does not directly treat the problem of a true or accurate linguistic representation of the world,5 nor does it address verification. Rather, it aims to identify what is necessary for an assertion to be assigned a truthvalue, in order that it can be contradicted. This problem, essential to dialectics (a linguistic practice whose starting point is the assumption that its initial assertion is true and whose aim is its refutation), is not the same as the one that asks how one actually determines the truth value of an assertion, nor the one that asks what is necessary for an assertion to be true or false. Consequently, Perihermeneias 1 does not put forth a semantic analysis of an ideal univocal language. In fact the possibility of ambiguity is considered further in the treatise (in Chapters 86 and 11), with the underlying intention of introducing rules into the uses of a language that is naturally susceptible to ambiguity, in order to enable interlocutors to accurately communicate their thoughts about some topic. Ancient and medieval scholars grasped and used this communicational aspect of the treatise, as seen in their interpretation of the influential passage Perihermeneias 1.16a3–9. They rightly noticed that Aristotle’s use of an inexact notion of signification is here guided by his intent to explain how assertions serve their purpose of accurately communicating, to another, someone’s

4 There is only one use of the word σημασία (i.e. signification) in the whole Corpus Aristotelicum, namely in Pr. 919b36. When talking about the semantic properties of human language, most of the time Aristotle uses conjugations of the verb σημαίνειν, and occasionally of the verb δηλου᷑ν, but there is no equivalent of the Latin terms significatio, denotatio, suppositio and so on, or of the modern terms meaning, sense and reference. For an analysis of the most influential attempts to reconstruct Aristotle’s theory of meaning followed by his own reconstruction, see M. Wheeler, ‘Semantics in Aristotle’s Organon’, Journal of the History of Philosophy 37.2 (1999), 191–226. 5 This does not mean that attemps to reconstruct what would have been Aristotle’s (or any other ancient or medieval author for that matter) reply to this problem are futile. In my opi­ nion, such attempts are an interesting and legitimate philosophical exercise, as long as this attempt of reconstruction is explicitly stated and as long as ancient or medieval authors are not presented as intentionally putting forth such accounts. 6 For a thorough analysis of Chapter 8, see S. Bobzien, ‘Aristotle’s De interpretatione 8 is about ambiguity’, in D. Scott (ed.), Maieusis. Essays in Ancient Philosophy in Honour of Myles Burnyeat (New York: oup, 2007), pp. 301–321. For an exhaustive account of Aristotle’s notion of ambiguity, see C. Shields, Order in Multiplicity: Homonymy in the Philosophy of Aristotle (Oxford: oup, 1999).

Ancient Sources

13

thought that (some or every) A is B.7 Let us start, then, by spelling out the details of a communicational reading of Perihermeneias 1.8 Perihermeneias 1.16a3–8: Utterances as Symbols and Signs The passage Perihermeneias 1.16a3–8, where Aristotle introduces the communicative dimension of human language, starts with a general semiotic statement that describes utterances as both symbols (σύμβολα) and signs (σημει᷑α) of thoughts, as well as thoughts as likenesses (ὁμοιώματα) of objects.9 The passage goes like this: Utterances are symbols of affections in the soul, and written expressions are symbols of utterances; and just as letters are not the same for everyone, utterances are not the same either; however, the things of which these [utterances] are primarily10 signs, the affections of the soul, are the 7

8

9

10

For the case of Nicholas of Paris (1240s) and Radulphus Brito (1290s), see my case-study in S. Ebbesen, D. Bloch, J. Fink, H. Hansen and A.M. Mora-Márquez, History of Philosophy in Reverse. Reading Aristotle through the Lenses of Scholars from the Twelfth to the Sixteenth Centuries (Copenhagen: The Royal Danish Academy of Sciences and Letters, 2014) (Scientia Danica. Series H. Humanistica 8. vol. 7), at pp. 131–140. For an exhaustive study of Perihermeneias 1, see S. Noriega-Olmos, Aristotle’s Psychology of Signification. A Commentary on De interpretatione 16a3–18 (Berlin: De Gruyter, 2013). However, Noriega-Olmos hesitates all along his book about the emphasis that should be put on the communicational character of the chapter; for this, see A.M. Mora-Márquez, ‘(Review of) Simon Noriega-Olmos, A Commentary on De interpretatione 16a3–8’, The Classical Review 64.2 (2014), 402–404. For other discussions of this passage, see W. Belardi, ‘Riconsiderando la seconda frase del De interpretatione’, Studi et saggi linguistici 21 (1981), 79–83; N. Kretzmann, ‘Aristotle on Spoken Sounds Significant by Convention’, in J. Corcoran (ed.), Ancient Logic and its Modern Interpretation (Dordrecht: Reidel, 1974), pp. 3–21; E. Montanari, La Sezione linguistica del Peri hermeneias di Aristotele. 2 vols. (Florence: Studi et testi 5 & 8, 1988); J. Pépin, ‘Sumbola, Sèmeia, Homoiômata: A propos de De interpretatione 1.16a3–8 et Politique viii 5, 1340a6-39’, in J. Wiesner (ed.), Aristoteles. Werk and Wirkung. t. 1 (Berlin: 1985), pp. 22–44; Whitaker, Aristotle’s De interpretatione, pp. 9–25. In his article ‘Aristotle on Signification and Truth’ [in G. Anagnostopoulos (ed.), A Companion to Aristotle (Oxford: Blackwell, 2009), pp. 81–100] Crivelli interprets the passage as a token-oriented semiotic statement. Minio-Paluello’s edition, following the Latin translations by Boethius and William of Moerbeke, as well as the Armenian and the Syrian translations, reads πρώτων. N. Kretzmann, J. Pépin and W. Belardi, following the Greek manuscript tradition, choose the adverbial reading πρώτως. Kretzmann and Pépin suggest that the adverb πρώτως qualifies ‘signs’ so that utterances are primarily signs of affections of the soul and only secondarily their symbols. W. Belardi suggests that πρώτως qualifies ‘utterances’ so that utterances are primarily signs of thoughts, in opposition to written expressions that are secondarily signs of thoughts. Note that the reading

14

chapter 1

same for everyone; and the things of which these [affections] are likenesses, the things, are also the same.11 I translate ‘τὰ ἐν τῃ᷑ φωνῃ᷑’ as ‘utterances’ to cover both complex and simple utterances: sentences, assertions, names and attributes.12 The alternative translation, ‘words’, would misrepresent the general character of the passage, and the same goes for ‘written words’, to which I prefer ‘written expressions’. I shall, however, paraphrase ‘utterance’ as ‘word’ when it is clear that we are dealing with simple expressions such as ‘man’ or ‘runs’. ‘Affections in the soul’, in turn, can be paraphrased as ‘thoughts’, also for the sake of maintaining the general character of the passage. Aristotle introduces two key notions when he says that utterances are symbols and signs of thoughts.13 Let me start by elucidating the import of qualifying utterances as symbols (σύμβολα). When Aristotle qualifies utterances as symbols, he continues a discussion going back to at least Plato’s Cratylus.14 In this dialogue, Socrates and Hermogenes discuss the natural or conventional character of names. Socrates presents us with a position (not necessarily his own) according to which, since πρώτων, which is the one I (and medieval scholars) follow and read in an adverbial way, rules out both Kretzmann and Pépin’s as well as Belardi’s interpretations. 11 Arist., Int. 1.16a3–8: ‘Ἔστι μὲν οὐ�ν τὰ ἐν τῃ᷑ φωνῃ᷑ τω᷑ ν ἐν τῃ᷑ ψυχῃ᷑ παθημάτων σύμβολα, καὶ τὰ γραφόμενα τω᷑ ν ἐν τῃ᷑ φωνῃ᷑. καὶ ὥσπερ οὐδὲ γράμματα πα᷑σι τὰ αὐτά, οὐδὲ φωναὶ αἱ αὐταί· ὡ�ν μέντοι ταυ᷑τα σημει᷑α πρώτων, ταὐτὰ πα᷑σι παθήματα τη᷑ ς ψυχη᷑ ς, καὶ ὡ�ν ταυ᷑τα ὁμοιώματα πράγματα ἤδη ταὐτά.’ 12 A ‘rhêma’, as we shall see, refers to whatever can be said of a name, whether it is an action, a disposition, a quality and so on. So, the grammatical category of verbs covers only some of the possible rhêmata considered by Aristotle. Since I think that the modern use of the word ‘verb’ has a heavy grammatical value, I dismiss this translation as misleading and use instead the word ‘attribute’. 13 For another discussion of σύμβολα and σημει᷑α, see C. Chiesa, ‘Symbole et signe dans le De interpretatione’, in H. Joly (ed.), Philosophie du langage et grammaire dans l’antiquité (Bruxelles: Editions Ousia, 1986) (Cahiers de Philosophie Ancienne 5), pp. 203–218. 14 Hermogenes is supposed to be defending Democritus’ position that names are not significant by nature. Cf. Plato, Crat. 388a8–b13. For an account of language in Democritus, see A. Brancacci, ‘Les mots et les choses: La philosophie du langage chez Démocrite’, in H. Joly (ed.), Philosophie du langage et grammaire, pp. 9–28. For general studies on the Cratylus, see R. Barney, Names and Nature in Plato’s Cratylus (New York/London: Routledge, 2001) and D. Sedley, Plato’s Cratylus (Cambridge: cup, 2003). Shorter discussions are found in D. Sedley, ‘Plato’s Cratylus’, Stanford Encyclopedia of Philosophy, http://plato.stanford.edu/ entries/plato-cratylus/; and N. Kretzmann, ‘Plato on the correctness of names’, American Philosophical Quarterly 8 (1971), 126–138.

Ancient Sources

15

naming is an action and every action has a proper nature, names are natural tools for the action of naming and ought to be chosen according to the proper nature of this action. Aristotle is opposed to the position that names are natural tools for naming and instead proposes in the Perihermeneias that utterances in general, names included, are a matter of human convention. Therefore, his description of utterances as symbols is better understood against the background of the opposition between conventional and natural tools. In one of its most common uses, the Greek word ‘σύμβολον’ stands for some token that is used as a proof of a contract between two parties – usually a tally that is broken into two pieces – so that each contracting part keeps one piece as a proof that the contract was sealed.15 But the tally can be round or squared, metallic or wooden, red or blue, and so on, so that the material aspects of the tally (shape, colour, material, size etc.) are not univocally dictated by the nature of the act of sealing a contract. Different sorts of tallies can be the proof of the same sort of contract and the same sort of tally can be the proof of different sorts of contract. The indication of the contract by some sort of tally follows from a somewhat arbitrary choice rather than from the nature of an action. Accordingly, just as the tally that seals a contract is not a natural tool, utterances are not natural tools for human communication either, but rather symbols chosen by human convention. In the case of, let us say, stones, some utterance other than ‘stone’ could have served the purpose of naming them. A good indication of this, as Aristotle himself suggests, is that the same thing can be spoken of by means of different utterances in different languages. Aristotle’s qualification of utterances as symbols aims thus to underscore the conventional character of human language as a tool for communication.16 15 16

In another common use, the word ‘σύμβολον’ stands for tokens serving as proofs of identity, cf. e.g. Lysias, Orat. 19.25. The conventional character of language is recalled in Chapters 2 and 4. In Chapter 2, Aristotle claims that names are conventional because no utterance is a name unless it is first a symbol – that is, until it is conventionally chosen to indicate a certain thing. Hence, it is not the inferential character of utterances that makes them names (or attributes or sentences), but rather the fact that they were conventionally chosen. Conventionality separates thus human language from animal sounds, which also indicate something, but not in the same way as human linguistic signs do (cf. Int. 2.16a27–28). In Chapter 4, Aristotle recalls again the conventional character of language when he claims that sentences are significative, not as a tool, but by convention (cf. Int. 4.17a1–2). For a more thorough discussion of the conventional character of human language as opposed to animal utterances, see Whitaker, Aristotle’s De interpretatione, pp. 45–52. Although I basically agree with Whitaker’s claim that for Aristotle linguistic articulation and rationality go together, it seems nevertheless untrue that Aristotle reserves the verb ‘τὸ σημαίνειν’

16

chapter 1

After having stated the conventional character of human language in lines 1.16a3–6, Aristotle moves from qualifying utterances as symbols to qualifying them as signs. Aristotle’s most common use of the word ‘σημει᷑ον’ makes reference to a proof – a piece of evidence that something else is or has been the case. The sort of evidence that he has in mind is described in detail in the last chapter of Prior Analytics, where he distinguishes proofs from likelihoods and describes proofs as ‘that which coexists with something else, or before or after whose happening something else has happened’,17 and which can be one of the three terms of an enthymeme (the kind of syllogism used in rhetorical settings).18 Examples of proofs are for a woman to have milk, which is a piece of evidence of her having given birth, or for Pittacus to be good, which is a piece of evidence that the wise are good.19 Therefore, that an utterance is the sign of a thought ought to mean that a speaker’s utterance is a piece of evidence of the occurrence of a thought in her mind. In fact, it is revealing that here Aristotle emphasises the primary relation of signification between utterances and thoughts, while in other places (e.g. in the Categories20) he rather talks of words that signify things. The reason for this emphasis must be that ‘primarily’, in the Perihermeneias, points to the primary content that the production of an utterance intends to transmit. Of course, this does not rule out the possibility that the utterance also points to the thing that

17 18

19 20

only to human utterances. There is for instance a passage in the Topics where Aristotle mentions a fact ‘that…indicates and produces health’ (τὸ σημαντικὸν καὶ τὸ ποιητικὸν ὑγιείας…, cf. Top. I.15.107b8). Aristotle’s use of the verbs ‘τὸ δηλου᷑ν’ and ‘τὸ σημαίνειν’ seems to be interchangeable and almost always their intended sense is the same – ‘to indicate x’, i.e. to direct someone’s attention to x. He thereafter adds explicit qualifications to the way in which something indicates something else, e.g. when he adds that articulate human sounds indicate something by convention. Translation by Harold P. Cooke in Loeb Classical Library 325 at p. 525. Cf. Arist., Pr. An. ii.27.70a8–10. This use of the word σημει᷑ον is very common in his philosophical and scientific treatises, where after having stated that ‘S is P’, he goes on to provide a piece of evidence by saying: ‘and a proof (σημει᷑ον) of this is that…’. There is another use of σημει᷑ον that makes reference to a mark, such as the mark that is left on the wax by a ring. Cf. Arist., De an. ii.12.424a18–21. Aristotle uses the ring/wax example in order to illustrate how sensation amounts to the reception of a form without the matter; but this use is obviously ruled out by the qualification of utterances as symbols, because the form of the mark left by a ring on the wax is univocally dictated by the shape of the ring, and therefore such a use does not involve the conventionality of human utterances that Aristotle intends to put forth in the Perihermeneias. Cf. Arist., Pr. An. ii.27.70a11–23. Cf. Arist., Cat. 4.1b25–27.

Ancient Sources

17

is the content of that thought,21 when underscoring that relation happens to be useful (as in the case of e.g. the context of the Categories, where a division of attributes is made in terms of their signification of different sorts of things). Nevertheless, since in the context of Perihermeneias 1 the main concern is to explain how the accurate communication of thoughts takes place, we are told that thoughts (and not things) are primarily indicated by utterances. It is also noteworthy that the interpretation of the sign as a piece of evidence entails the intervention of someone for whom the sign is a piece of evidence, namely a listener in the case of utterances. Aristotle’s view on linguistic signification is thus inscribed in a context of communication of thoughts, where at least a speaker and a listener are involved, so that an utterance is significative if it is for a listener a piece of evidence of the occurrence of a thought in the speaker’s mind. This interpretation is supported by a passage in Chapter 3, where Aristotle himself considers the understanding by the listener a sufficient condition for an utterance (a name in this case) to be significative: Then, the attributes, when they are said themselves by themselves, are names and signify a certain thing; for the speaker stops [his] thinking and the listener came to rest.22 Aristotle is thus pointing out in the short and influential passage 1.16a3–8 of the Perihermeneias the following two features of human language: It primarily 21

Thoughts (both simple, such as the concept of man, and compound, such as the thought that every man is an animal) are described as likenesses of things. This roughly means that thoughts are somehow their objects (cf. De an. iii.6.431b22–23). In other words, when grasping an object, the soul becomes not exactly its object, but similar to it. This similarity consists in the thought sharing something with its object, namely its formal aspect (Arist., De an. iii.8.431b26–30: ‘The sensitive and knowing faculties of the soul are in potency [the objects] – the object of knowledge and the sense-object. And it is necessary that they are either the objects themselves or the forms. But not the objects themselves; for the stone is not in the soul, but its form.’) Now, since objects (and more precisely their formal aspect) remain always the same, different human beings can have formally the same thought of a determinate object, so that in their act of communication it is possible that they attribute the same predicate to the same object of thought – it is possible for two interlocutors to talk about the same thing. For a thorough reconstruction of the Aristotelian account of intellection, see S. Noriega-Olmos, Aristotle’s Psychology; and D. Charles, Aristotle on Meaning and Essence (Oxford: Clarendon Press, 2000). 22 Arist., Int. 3.16b20–21: ‘αὐτὰ μὲν οὐ�ν καθ’ αὑτὰ λεγόμενα τὰ ῥήματα ὀνόματά ἐστι καὶ σημαίνει τι, – ἵστησι γὰρ ὁ λέγων τὴν διάνοιαν καὶ ὁ ἀκούσας ἠρέμησεν, – …’.

18

chapter 1

signifies thoughts – it gives the listener evidence of a thought occurring in the speaker’s mind – and it does so, not as a natural tool for communication, but as the result of human convention. Perihermeneias 1.16a9–18: Simple and Compound Linguistic Items After having stated that utterances are conventionally and primarily significative of thoughts, Aristotle proceeds to a division of linguistic utterances according to the sort of thought they indicate: And it is the case that, just as in the soul sometimes there is a thought without truth or falsity and sometimes [a thought] where there is already necessarily one of them, in the same way also in the utterance. For both truth and falsity are about composition and division. Therefore, the names themselves and the attributes are like the thought without composition and division, e.g. ‘man’ or ‘white’, when nothing is added; for it is not yet true or false. And here is a proof: ‘Goat-stag’ signifies something, but is not yet true or false, unless that it is or that it is not is added either simply or with respect to time.23 Since thoughts can be either compound or simple, utterances too are either compound or simple. Compound thoughts are the opinion or the judgment that something is the case or is not the case (e.g. that Homer is a poet), and they can be given a truth-value precisely because they are compound – because they state that something is or is not the case about an object.24 These thoughts are communicated by means of assertoric sentences, i.e. assertions that can also be given a truth-value if and only if they preserve the compound form of the thought that they communicate, i.e. if and only if they are composed of a name and an attribute that indicate the simple thoughts involved in the compound thought. Simple thoughts that are not susceptible of truth or falsity are just the intellectual understanding of a kind of being (e.g. being a man or being white). These thoughts are indicated by means of simple utterances, such as names 23 Arist., Int. 1.16a9–18: ‘ἔστι δέ, ὥσπερ ἐν τῃ᷑ ψυχῃ᷑ ὁτὲ μὲν νόημα ἄνευ του᷑ ἀληθεύειν ἢ ψεύδεσθαι ὁτὲ δὲ ἤδη ῳ�� ἀνάγκη τούτων ὑπάρχειν θάτερον, οὕτω καὶ ἐν τῃ᷑ φωνῃ᷑· περὶ γὰρ σύνθεσιν καὶ διαίρεσίν ἐστι τὸ ψευ᷑δός τε καὶ τὸ ἀληθές. τὰ μὲν οὐ�ν ὀνόματα αὐτὰ καὶ τὰ ῥήματα ἔοικε τῳ᷑ ἄνευ συνθέσεως καὶ διαιρέσεως νοήματι, οἱο� ν τὸ ἄνθρωπος ἢ λευκόν, ὅταν μὴ προστεθῃ᷑ τι· οὔτε γὰρ ψευ᷑δος οὔτε ἀληθές πω. σημει᷑ον δ’ἐστὶ του᷑δε· καὶ γὰρ ὁ τραγέλαφος σημαίνει μέν τι, οὔπω δὲ ἀληθὲς ἢ ψευ᷑δος, ἐὰν μὴ τὸ εἰ �ναι ἢ μὴ εἰ �ναι προστεθῃ᷑ ἢ ἁπλω᷑ ς ἢ κατὰ χρόνον.’ 24 For Aristotle’s account of truth, see P. Crivelli, Aristotle on Truth (Cambridge: cup, 2004).

Ancient Sources

19

and attributes.25 Names and attributes are not susceptible of truth-value because they indicate thoughts that do not state the occurrence or nonoccurrence of a state of affairs. For instance, when someone says ‘man’, she is not stating that there is actually some man. But although names and attributes indicate simple thoughts that are not susceptible of truth-value, they can produce assertions that are susceptible, when a simple utterance – an attribute – is added to another simple utterance – a name. Accordingly, assertions preserve the thought’s susceptibility of truth-value if and only if they preserve its logical compound form by being composed of a name and an attribute that indicate the parts of the compound thought. An assertion transmits a judgment or an opinion, which can be true or false because it states a fact by comparing (composing or dividing) simple thoughts. Here Aristotle seems again to make an anti-platonic statement, for his point seems to be that names do not by themselves state the existence of their significate, e.g. ‘Good’ does not by itself state the existence of the Good. Statements can only result from composing something with or separating it from a bearer. Since by saying ‘goat-stag’ – a non-existing entity – we do not state its nonexistence and it is still necessary to say ‘goat-stags do not exist’, a fortiori by simply uttering ‘man’ or ‘white’ we do not state anything about man or white, unless we use a compound statement like ‘a man is running’ or ‘Socrates is white’. In sum, for an utterance to be an assertion that can be given a truth-value, it has to be composed of a name and an attribute – it has to preserve the logical form of a judgment or opinion that can be given a truth-value, because it is a compound of the thing we think something about (what is indicated by the name) and what we think is the case or is not the case about this thing (what is indicated by the attribute). Consequently, this whole chapter is about the possibility for assertions to preserve the compound logical form – and hence the susceptibility of truth-value – of the thoughts they intend to communicate to someone else. This concern is most important in the case of dialectics, where an assertion’s susceptibility of being granted as true or rejected as false is the condition sine qua non for its refutation by its contradictory assertion. Hence, after having thus introduced the name, the attribute, the sentence and the affirmation and negation in Chapters 2–6, Aristotle proceeds to introduce the contradictory pair and the principle of bivalence (PoB) in Chapter 7.

25

Note that names and attributes are simple not because they are one word, but rather because they indicate a simple thought. In other words, their simplicity follows from the simplicity of the thought that they indicate, rather than from the simplicity of their utterance.

20

chapter 1

Contradiction and Equivocation The contradictory pair is defined by Aristotle as the set of an affirmation and a negation of the same thing about the same thing. This definition entails a further requirement a name and an attribute ought to meet in order to bring about an assertion that complies with the PoB – the name and the attribute ought to indicate exactly the same thing both in the affirmation and in the negation (i.e. they have to indicate the same thing both for the one who grants a proposition as true and for the one who aims to reject it as false). This requirement is crucial for Aristotle’s analysis because of a feature of human language that he consi­ ders both natural and the most common cause of linguistic misunderstandings: ambiguity.26 Thus, when Aristotle says that the affirmation and the negation ought to be of the same thing about the same thing, he intends to guard his readers against one of the possible cases in which a contradictory pair will not comply with the PoB, namely when the name or the attribute in the affirmation and in the negation is used equivocally by the speaker and the listener. In Chapter 8 of the Perihermeneias,27 Aristotle tells us that an affirmation (or a negation) is a unity when it indicates that one thing holds of one thing, e.g. ‘every man is white’.28 An affirmation (or a negation) that fails to meet this requirement is not one, but plural. Aristotle illustrates the case of a plural affirmation by means of the following situation. Let us suppose that the name ‘cloak’ (‘ἱμάτιον’) was given to both man and horse. In this case, the affirmation ‘a cloak is white’ is not one but two affirmations, as it states both that a man is white and that a horse is white. Now, when someone says ‘a cloak is white’ and her thought is about a horse, while her interlocutor thinks of a man, their exchange can go on without there being real contradiction. For if the speaker claims that ‘a cloak is white’ and her interlocutor claims that ‘a cloak is not white’, and one is talking about a man while the other is talking about a horse, the assertions do not form a contradictory pair that complies with the PoB, and hence the latter will not refute the former.29 Consequently, for a name and an attribute to bring about an assertion that complies with the PoB, not only do they need to indicate simple thoughts, they also have to be used by the speaker and the listener as indicating the same simple thoughts. § 26 27 28 29

Cf. Arist., se 1.165a3–13. For a thorough analysis of Chapter 8, see Bobzien, ‘Aristotle’s De interpretatione 8…’. Cf. Arist., Int. 8.18a12–18. Cf. Arist., Int. 8.18a18–27.

Ancient Sources

21

The treatment that Aristotle gives to linguistic signification both in the Perihermeneias and in the dialectical treatises to which he explicitly refers (i.e. the Topics and the Sophistical Refutations) shows that Aristotle’s semiotic statement in Perihermeneias 1.16a3–8 does not intend to explain the nature of a truthful linguistic representation of the world, but rather the assertion’s function of conveying information to someone else. Aristotle wants to explain how there can be unambiguous linguistic exchange about our opinions or judgments, so that assertions can be susceptible of affirmation as true or rejection as false and of subsequent contradiction. For an assertion to be granted as true or rejected as false, it has to be composed of a name and an attribute, and both its name and its attribute ought to signify one and the same simple thoughts to both interlocutors. Only when these conditions are satisfied can the assertion be given a truth-value and comply with the PoB – it can be the starting point of a dialectical discussion. Thirteenth-century scholars were well aware of the communicational dimension behind the claim that utterances are primarily signs of thoughts, as well as of its importance for dialectical practices, and hence this dimension surfaces repeatedly in medieval developments of Aristotelian logic. However, readers of the Perihermeneias at least as early as Porphyry also begin to wonder about the truthful linguistic representation of the world and start to produce interpretations of the treatise in that sense. Boethius, in turn, incorporates these interpretations and transmits them to the late Middle Ages in his own commentaries on the treatise. Since Boethius was undoubtedly the source for thirteenth-century scholars, his interpretation greatly shaped theirs and endowed them with a verificational aspect that was to co-exist with the communicational one, albeit not without some confusion. In order to see this more clearly, let me now turn to Boethius’ reception of Aristotle’s Perihermeneias 1. 1.2

Boethius’ Second Commentary on the Perihermeneias

Anicius Manlius Severinus Boethius (ca. 486–ca. 525) provided thirteenthcentury scholars with two commentaries on and a translation of the Perihermeneias, as well as with translations of and commentaries on Porphyry’s Isagoge and Aristotle’s Categories. These translations and commentaries had an enormous impact on the way scholastics interpreted the linguistic ideas in the Organon. They also transmitted to the Middle Ages pieces of information of long-lost commentaries on the Organon belonging to the peripatetic and neo-platonic traditions, such as those of Alexander of Aphrodisias, Aspasius, Herminus and Porphyry. Alexander and Porphyry’s commentaries exerted a

22

chapter 1

strong influence on Boethius, and through him important peripatetic and neoplatonic notions were passed to the thirteenth century.30 In what follows we shall focus on Boethius’ interpretation of Perihermeneias 1, taking his second commentary on the Perihermeneias as the main source, but drawing from other texts when necessary.31 The Subject Matter of the Perihermeneias according to Boethius Boethius is responsible for the translation of the Greek Perihermeneias into the Latin De interpretatione. An interpretation (i.e. a ‘rendering’), which is according to him the subject matter of the Perihermeneias, is defined as an articulate sound that is significative by itself.32 The Roman commentator reaches this definition after a process of division that goes from sounds (a blow through the throat, e.g. a coughing), to utterances (voces) (a sound produced by beating the air with the tongue, e.g. a barking), to expressions (locutio) (an articulate utterance that can be written with letters, e.g. ‘blityri’ or ‘et’), and down to interpretations (expressions that are significative by themselves, e.g. ‘homo’).33 Hence, being significative by itself is the feature that separates interpretations from other kinds of expressions, and then the subject matter of the Perihermeneias are articulate utterances that are significative by themselves (i.e. names, verbs and assertions).34 Boethius stresses that the treatise is not only about expressions (locutio/lexis), because being significative by itself is an essential part of its subject matter, and it is not only about sentences (oratio/logos) either, 30

See S. Ebbesen, ‘Boethius as an Aristotelian Commentator’, in R. Sorabji (ed.), Aristotle Transformed. The Ancient Commentators and their Influence (London: Duckworth, 1990), pp. 373–391. 31 The main studies about Boethius’ commentaries on the Perihermeneias are John Magee’s Boethius on Signification and Mind (Leiden: Brill, 1989) (Philosophia Antiqua 52) and Taki Suto’s Boethius on Mind, Grammar and Logic: A Study of Boethius’ Commentary on Peri hermeneias (Leiden: Brill, 2012) (Philosophia Antiqua 127). For a comprehensive study of Boethius’ philosophical thinking see J. Marenbon, Boethius (Oxford: oup, 2003). For his life and influence, see also J. Marenbon, Boethius. 32 Boethius, In Perih.2 (ed. Meiser, 4:26–28): ‘Illa quoque potest esse definitio vocis, ut eam dicamus sonum esse cum quadam imaginatione significandi.’ In: Commentarii in librum Aristotelis ΠΕΡΙ ΕΡΜΗΝΕΙΑΣ, ed. C. Meiser, 2 vols. (Leipzig: Teubner, 1877–1880). 33 Boethius, In Perih.2 (ed. Meiser, 6:1–5): ‘Concurrentibus igitur his tribus: linguae percussione, articulato vocis sonitu, imaginatione aliqua proferendi fit interpretatio. Interpretatio namque est vox articulata per se ipsam significans.’ 34 Boethius, In perih.2 (ed. Meiser, 6:20–21): ‘Interpretationis vero partes hoc libro constituit nomen et verbum.’ Note that by Boethius’ time, name and verb (nomen and verbum) – his translations of onoma and rhêma – are already well established grammatical categories.

Ancient Sources

23

because names and verbs are also significative by themselves and therefore should be treated as interpretations in their own right. Accordingly, the treatise deals with any kind of utterance that is significative by itself, putting thus the focus of Perihermeneias 1 on the very notion of signification. Boethius on Perihermeneias 1.16a3–8 Boethius narrows his analysis of Perihermeneias 1.16a3–8 to only names (nomina) and verbs (verba) on the grounds that the passage is the logical continuation of Perihermeneias 1.16a1–2,35 where Aristotle stated the intention of the treatise: It must be established what is a name and a verb, and then what is a negation and an affirmation, as well as an assertion and a sentence.36 This would mean that Aristotle intends to open his enquiry with a description of the significative character of names and verbs. At the very beginning of his own enquiry about names and verbs, the Roman commentator introduces the influential ‘semantic’ triangle res–intellectus–vox: Whether in the case of a question and an answer, or in the case of the continuous sequence of a sentence and someone else’s understanding and listening, or in the case that [the master] teaches and [the pupil] learns, the whole arrangement of speech (ordo orandi) is achieved with these three elements: things, thoughts and utterances.37 He proceeds to explain the relations between these elements, claiming that:

35 Boethius, In Perih.2 (ed. Meiser, 31:9–14): ‘Quare non est disiuncta sententia sed primae propositioni continua. Nam cum quid sit verbum, quid nomen definire constituit, cum nominis et verbi natura sit multiplex, de quo verbo et nomine tractare vellet clara significatione distinxit.’ Cf. Arist., Int. 1.16a1–2. Recall that in Aristotle the extent of this passage is general and not restricted to only names and verbs; see Section 1.1. 36 Aristoteles latinus, Int. 1.16a1–2: ‘Primus oportet constituere quid sit nomen et quid verbum, postea quid est negatio et adfirmatio et enuntiatio et oratio.’ (Boethius’ translation). 37 Boethius, In Perih.2 (ed. Meiser, 20:12–17): ‘Sive enim quaelibet interrogatio sit atque responsio, sive perpetua cuiuslibet orationis continuatio atque alterius auditus et intelligentia, sive hic quidem doceat ille vero discat, tribus his totus orandi ordo perficitur: rebus, intellectibus, vocibus.’

24

chapter 1

The thing is conceived by the intellect, the utterance signifies the thoughts and concepts of the soul and the thoughts both conceive the things that are their objects and are signified by the utterances.38 Things, Boethius tells us, are there in the world presumably as the result of some providence and have a nature that is distinctively arranged. Thoughts, in turn, are always concomitant to and produced by things that have a distinct nature and that stand before the human senses. Finally, utterances are always the expression of thoughts.39 These elements are arranged40 so that: The thing precedes the thought, the thought [precedes] the utterance, and the utterance [precedes] the letters – but this cannot be converted.41 38 Boethius, In Perih.2 (ed. Meiser, 20:17–20): ‘Res enim ab intellectu concipitur, vox vero conceptiones animi intellectusque significat, ipsi vero intellectus et concipiunt subiectas res et significantur a vocibus.’ 39 Cf. Boethius, In Perih.2 (ed. Meiser, 20:28–21:5): ‘…intellectus autem concipiant res, quae scilicet habent quandam non confusam neque fortuitam consequentiam sed terminata naturae suae ordinatione constant. Res enim semper comitantur eum qui ab ipsis concipitur intellectum…Rebus enim ante propositis et in propria substantia constitutis intellectus oriuntur. Rerum enim semper intellectus sunt, quibus iterum constitutis mox significatio vocis exoritur. Praeter intellectum namque vox penitus nihil designat.’ 40 In his book Boethius on Signification and Mind, John Magee raises the question of what sort of priority is at stake here and comes to the conclusion that it is a logical priority, so that things are a sufficient, but not a necessary, condition for thoughts and utterances (cf. Magee, Boethius on Signification, pp. 72 and 92). However, Magee’s conclusion cannot follow from the second commentary on the Perihermeneias alone; for in fact things are independent of our knowledge of them, as it is suggested in the following claim [Boethius, In Perih.2 (ed. Meiser, 22:9–11)]: ‘Cum res est, eius quoque esse intellectum, quod si non apud homines, certe apud eum, qui propriae divinitate substantiae in propria natura ipsius rei nihil ignorat.’ So, Boethius clearly considers the possibility for a thing to exist without being known by men, even if it is never unknown to God, who always has the thoughts of all things in his divine mind. It is also clear that a thing is not a necessary condition for a thought; for in fact Boethius accepts the occurrence of empty thoughts, such as the thought of chimeras and of centaurs [Boethius, In Perih.2 (ed. Meiser, 22:3–5)]: ‘Sunt enim intellectus sine re ulla subiecta, ut quos centauros uel chimeras poetae fixerunt. Horum enim sunt intellectus quibus subiecta nulla substantia est.’ Consequently, things are neither a necessary nor a sufficient condition for thoughts, so no logical priority between things and thoughts can be established on the basis of the second commentary on the Perihermeneias alone (and thereby not between things and utterances either). 41 Boethius, In Perih.2 (ed. Meiser, 21:28–30): ‘Praecedit autem res intellectum, intellectus vero vocem, vox litteras – sed hoc converti non potest.’

Ancient Sources

25

What is at stake here is a natural priority, so that when a thing of a determinate sort, and whose existence is independent of and prior to its knowledge by men, stands before a human mind, it triggers a natural process of understanding it – natural in the sense that this process can be described by a natural discipline, such as psychology. Significative utterances, in turn, which are given as names of things by a conventional choice (ad placitum), are produced because thoughts trigger in their possessor the will to communicate them and, thereby, the psychological process leading to this communication. Hence, Boethius’ description of this ordo orandi has to do with the natural and temporal arrangement of the items res–intellectus–vox. After his discussion of the ordo orandi, Boethius goes on to comment on the passage 1.16a3–8 of the Perihermeneias, which in his own translation goes as follows: Sunt ergo ea quae sunt in voce earum quae sunt in anima passionum notae et ea quae scribuntur eorum quae sunt in voce. Et quemadmodum nec litterae omnibus eaedem, sic nec voces eaedem. Quorum autem haec primorum notae, eaedem omnibus passiones animae et quorum hae similitudines, res etiam eaedem. De his quidem dictum est in his quae sunt dicta de anima, alterius est enim negotii. (al, Int. 1.16a3–8, Boethius’ translation) Perhaps the most remarkable feature of Boethius’ translation of this passage is the Latin ‘nota’ as the rendering of both ‘σύμβολον’ and ‘σημει᷑ον’. As Taki Suto shows, ‘nota’ covers both the idea of being a conventional token (i.e. being a σύμβολον) and the idea of being a piece of evidence that something else is the case (i.e. being a σημει᷑ον), and hence it does not seem to be a poor translation choice for both terms.42 But Boethius had other possibilities at hand, namely to render ‘σύμβολον’ as ‘symbolum’ and ‘σημει᷑ον’ as ‘signum’, so why did he not do so? ‘Symbolum’ for ‘σύμβολον’, Suto explains, was not a suitable choice, because at Boethius’ time the word had a loaded meaning, i.e. ‘symbolum’ stood either negatively for a mystical sign in a pagan cult or positively for the Apostles’ Creed.43 ‘Signum’ for ‘σημει᷑ον’ did not look like a good choice either (at least 42 43

In his article ‘Medieval Logicians on the Meaning of the Propositio’, Journal of Philosophy 67 (1970), 767–787, N. Kretzmann considers this translation choice as ‘a serious mistake’. Cf. Suto, Boethius on Mind, pp. 52–58. Suto bases her argument on a careful analysis of the use of the word in passages by Plautus, Pliny, Tertullian, Augustine, Firmicus Maternus and Cyprian.

26

chapter 1

not to Boethius). In fact, ‘signum’ was Boethius’ translation of ‘σημει᷑ον’ in the Prior Analytics, but he saw an important difference between a ‘sign’ in the context of the Prior Analytics, where he takes it to stand only for propositional items, and in the context of the Perihermeneias, where he takes it to stand for any utterance significative by itself, names and verbs included.44 Suto proposes that the Roman commentator avoided ‘signum’ to prevent his readers here (i.e. in the Perihermeneias) from an association between signa and propositional items; an association that she claims to be ‘widely acknowledged in the ancient traditions of logic, not only among the peripatetics but also the stoics’.45 There are, however, ancient uses of the word ‘signum’ – ‘signum’ as a proof or piece of evidence that something else is or has been the case – that would cover properly what Aristotle intends to express in the Perihermeneias with the word ‘σημει᷑ον’, and then it would have been a suitable translation choice. Suto quotes indeed a passage from Cicero’s De inventione where the sense of ‘signum’ is the one at stake in the Perihermeneias: A sign is something that falls under some of the senses and indicates something that seems to arise from it; [something] that may have happened before or at the same time or that may have followed it; and yet it requires proof and stronger corroboration – e.g. blood, flight, paleness, dust and things that are similar to these.46 ‘Signum’, as Cicero uses it here, would have properly rendered the sense of Aristotle’s ‘σημει᷑ον’ in the Perihermeneias’ passage – the significative utterance as a piece of evidence of the occurrence of a thought in the speaker’s mind. Boethius’ reason for not using ‘signum’ could rather be that this Latin word usually stands for an event that is the proof or piece of evidence of some other event being the case, and it rarely, if ever, applies to linguistic expressions. But it did not have to apply only to linguistic expressions, since Aristotle’s point is precisely that utterances, just as any other sign, are pieces of evidence that something else is the case. At any rate, perhaps rendering both ‘σύμβολον’ and ‘σημει᷑ον’ with ‘nota’ was not as big a mistake as Kretzmann thought it to be, as

44 Cf. Magee, Boethius on Signification, p. 63 and Suto, Boethius on Mind, p. 59. 45 Suto, Boethius on Mind, p. 64. 46 Cicero, De inv. I.30, 48: ‘Signum est quod sub sensum aliquem cadit et quiddam significat quod ex ipso profectum videtur, quod aut ante fuerit aut in ipso negotio aut post sit consecutum et tamen indiget testimonii et gravioris confirmationis, ut cruor, fuga, pallor, pulvis, et quae his sunt similia.’

Ancient Sources

27

the word seems to have properly conveyed the idea of being a conventional linguistic indication of thoughts. When commenting on Perihermeneias 1.16a3–8, Boethius raises the question why Aristotle inserted this elliptic and elusive passage at the beginning of the treatise. He also tells us that before him, at least Herminus, Alexander of Aphrodisias and Porphyry had raised the same question. After having dismissed Herminus’ explanation, we are given Alexander’s, who claimed that Aristotle inserted the passage ‘because the value of significative [words] comes from what is signified’ – Aristotle intended to state what is the semiotic value of words.47 We are also told that Porphyry’s explanation is the most accurate; Aristotle was, in fact, taking a position in an old discussion about the value of significative words: But Porphyry expounded the cause and the origin of this statement more fully, he who retraced the whole controversy and argument of the ancient philosophers concerning the value of signification. He says that the opinions of the ancient philosophers had been uncertain as to what words exactly signified. Some believed that words signified things and that what resounded in words were their names. Others thought of incorporeal natures… Others believed that sensations [were signified], others again that images were signified by words.48 Presumably, the discussion regarded the nature of the words’ semiotic content. In his passage, Aristotle would be dismissing the positions according to which words signify: (a) external things, (b) Platonic Ideas49 and (c) mental items, 47 Boethius, In Perih.2 (ed. Meiser, 26:9–11): ‘…quocirca quoniam significantium momentum ex his quae significantur oritur…’. 48 Boethius, In Perih.2 (ed. Meiser, 26:17–27:6): ‘Sed Porphyrius ipsam plenius causam originemque sermonis huius ante oculos conlocavit, qui omnem apud priscos philosophos de significationis vi contentionem litemque retexuit. Ait enim dubie apud antiquorum philosophorum sententias constitisse quid esset proprie quod vocibus significaretur. Putabant namque alii res vocibus designari earumque vocabula esse ea quae sonarent in vocibus arbitrabantur. Alii vero incorporeas quasdam naturas meditabantur…Alii vero sensus, alii imaginationes significari vocibus arbitrabantur.’ This translation is a slight modification of Arens’ translation. Cf. H. Arens, Aristotle’s Theory of Language and its Tradition (Amsterdam: John Benjamins, 1984), p. 167. 49 Boethius does not discuss this possibility, not because he denies the existence of incorporeal natures, but because for him, following Porphyry, logic is a discipline about human language, whose extension is the sensible world. See S. Ebbesen, ‘Porphyry’s Legacy to Logic: a Reconstruction’, in Sorabji (ed.), Aristotle Transformed, pp. 141–172. Suto and

28

chapter 1

such as sensations or images. Boethius proceeds to explain first the reason why Aristotle dismisses sensations and images as the significates of names and verbs. Sensations cannot be the passions of the soul Aristotle is talking about in Perihermeneias 1.16a3–8, because for him sensations are not passions of the soul but passions of the body.50 The situation is more complicated in the case of images;51 for they seem to be a legitimate instance of ‘passions of the soul’.52 Boethius goes about this problem by recalling a passage in De anima,53 where Aristotle considers images as something incomplete with respect to concepts,54 because, even though images are a necessary condition for the formation of concepts, it is not possible to produce a mental judgment out of images. Hence, following in the footsteps of Aristotle, Boethius tells us that, since images cannot produce compound thoughts, they cannot be the significates of names and verbs either: Wherefore the image is something incomplete, but names and verbs signify something complete and not something that falls short.55 The passions of the soul to which Aristotle is referring in this passage must, then, be concepts. The common feature of names and verbs is, thus, to be utterances conventionally significative of concepts.56 An utterance such as ‘garalus’ may, because of its appearance (figura), look like a name to the grammarian, Magee propose that Boethius has Proclus’ theory of names in mind, according to which ‘man’ signifies primarily the Idea of Man and only secondarily instantiated men. See Magee, Boethius on Signification, p. 95, n. 9 and Suto, Boethius on Mind, p. 28. 50 Boethius, In Perih.2 (ed. Meiser, 27:18–20): ‘Sed qui passiones animae a uocibus significari dicit, is non de sensibus loquitur. Sensus enim corporis passiones sunt.’ 51 Note that ‘image’ refers here to the impression left by a form in the imaginative part of the soul – the phantasia. 52 Boethius, In Perih.2 (ed. Meiser, 27:25–28:1): ‘Sed quoniam imaginatio quoque res animae est, dubitauerit aliquis ne forte passiones animae imaginationes, quas Graeci *phantasias* nominant, dicat.’ 53 Cf. Arist., De an. iii.8.432a10–14. 54 Cf. Boethius, In Perih.2 (ed. Meiser, 29:6–10): ‘Nam cum res aliqua sub sensum vel sub cogitationem cadit, prius eius quaedam necesse est imaginatio nascatur, post vero plenior superveniat intellectus cunctas eius explicans partes quae confuse fuerant imaginatione praesumptae.’ 55 Boethius, In Perih.2 (ed. Meiser, 29:11–13): ‘Quocirca imperfectum quiddam est imaginatio, nomina uero et uerba non curta quaedam sed perfecta significant.’ 56 Cf. Boethius, In Perih.2 (ed. Meiser, 31–32).

Ancient Sources

29

yet is not a name for the philosopher as it has not yet been imposed in order to signify a concept.57 So far, however, there is no mention of possibility (a) – that the significates of words are things. Perhaps in order to fill this void, Boethius proceeds to introduce the crucial idea that the immediate significates of names and verbs are concepts, things being their significates only secondarily: [U]tterances signify thoughts and things, but thoughts principally, [and they signify] the things that the intelligence itself understands by a secondary signification with the mediation of thoughts.58 Boethius, who seems to be more interested than Aristotle himself in explaining the nature of the link between language and the world, reads the primary link between words and concepts as a semantic link that entails a secondary one between words and things, thereby introducing, for the first time in an extant Latin commentary on Aristotle’s Perihermeneias, the idea of a secondary signification of things by words. Surely, it is possible to account for the signification of things by means of the signification of concepts. Suto nicely explains how this can happen by proposing two situations in which we see something through something else: (a) When we see a planet through the lens of a telescope. (b) When we see someone approaching through a mirror. In the first case we see both the lens and the planet, but the planet through the lens. In the second case, by contrast, it would be fair to say that we only see the mirror and that we see the person approaching in the mirror. The signification of things through concepts in Boethius would be analogous to the situation of the telescope:59 through the signification of the concept the name also signifies the concept’s object and does not stay stuck in the concept itself. However, this would also suggest that, just as in the situation of the telescope the main target of vision is the planet, in the same way the main target of the signification of words would be the external world. Consequently, the reader is invited to lay an emphasis on the relation between language and world that is not at all evident in the Aristotelian passage. Boethius neither develops a positive argument for the immediate signification of concepts by names and verbs nor explains why Aristotle would dismiss 57 Cf. Boethius, In Perih.2 (ed. Meiser, 32). 58 Boethius, In Perih.2 (ed. Meiser, 33:27–31): ‘…ea quae sunt in uoce res intellectusque significent, principaliter quidem intellectus, res uero quas ipsa intellegentia comprehendit secundaria significatione per intellectuum medietatem.’ 59 Cf. Suto, Boethius on Mind, p. 33.

30

chapter 1

things as their immediate significates. As to the signification of things, Boethius only tells us that Alexander of Aphrodisias had already raised the question why Aristotle says that names are primarily significative of concepts, even though they are names of things. Thereafter, Boethius spends a few lines telling us that, according to Alexander, names name things but are primarily significative, i.e. indicative, of concepts, because their primary function is to make thoughts public.60 Note that Alexander’s explanation would point to a communicational reading of the passage in question similar to the one that was given in Section 1.1; an explanation that, in passing, Boethius neither grants nor rejects. As to the immediate signification of concepts, there is a passage at the very beginning of his commentary where Boethius says that: Insofar as the word itself signifies a concept, it is divided into two parts, as it was already said, i.e. into name and verb; but insofar as the word signifies, by means of the concepts, the things that are the objects of thought, Aristotle divides the number of significative words into ten ca­­ tegories…Name and verb are a certain quality of the word and, of course, they signify the ten categories; for the ten categories are never uttered without the quality of the name or of the verb. Wherefore, the intention of this book is to treat significative words insofar as they signify concepts of the soul and of the intellect. The intention of the Categories was described in its commentary: Since it is about words that signify things, their signification can be divided into as many parts as things that are objects of thought can be indicated by words through sensations and concepts.61 The approach to significative simple utterances – names and verbs – in the Perihermeneias is thus explained in opposition to the approach to significative 60 Cf. Boethius, In Perih.2 (ed. Meiser, 41). 61 Boethius, In Perih.2 (ed. Meiser, 7:12–8:7): ‘Et quantum vox ipsa quidem intellectus significat, in duas (ut dictum est) secatur partes, nomen et verbum, in quantum vero vox per intellectuum medietatem subiectas intellectui res demonstrat, significantium vocum Aristoteles numerum in decem praedicamenta partitus est…Vocis enim quaedam qualitas est nomen et verbum, quae nimirum ipsa illa decem praedicamenta significant. Decem namque praedicamenta numquam sine aliqua verbi qualitate vel nominis proferentur. Quare erit libri huius intentio de significativis vocibus in tantum quantum conceptiones animi intellectusque significent. De decem praedicamentis autem libri intentio in eius commentario dicta est, quoniam sit de significativis rerum vocibus, quot partibus distribui possit earum significatio in tantum quantum per sensuum atque intellectuum medietatem res subiectas intellectibus voces ipsae valeant designare.’

Ancient Sources

31

simple utterances – the ten Aristotelian categories – in the Categories. A word is qualified as a name or a verb insofar as it indicates a concept, but the same word falls under a certain Aristotelian category depending on the sort of thing that it indicates through that concept. Here again, Boethius does not explain why for a significative word to be qualified as a name or as a verb it has to indicate a concept. However, his opposition between signification of things (in the Categories) and signification of concepts (in the Perihermeneias) will provide medieval scholars with the starting point of more developed explanations of the immediate signification of concepts by names and verbs. Concepts and Likenesses in Boethius Boethius’ suggestion that names and verbs signify things secondarily by means of their signification of concepts leads us to the question whether concepts could somehow distort the way we understand things and, thereby, the way we speak about them. In his commentaries on both the Isagoge and the Perihermeneias, following in Aristotle’s footsteps Boethius describes the concept as a likeness: [A]ristotle called the passion of the soul a likeness, which, according to Aristotle, grasps nothing different from the essential feature of the thing that is the object [of thought].62 Which sort of likeness is at stake here? Let us consider the case of the concept of man and raise the problem in the following way: (a) Is the concept of man an understanding of the human nature itself or is it the understanding of an agreement between several men?63 (b) Is the concept of man formally identical with the form of any given man?

62 Boethius, In Perih.2 (ed. Meiser, 43:12–15): ‘Similitudinem uero passionem animae uocauit, quod secundum Aristotelem nihil aliud intellegere nisi cuiuslibet subiectae rei proprietatem.’ 63 As could be suggested by this passage from Boethius’ commentary on the Isagoge [In Isag.2 (ed. Brandt, 166)]: ‘Then, their likeness is gathered from the individuals in which they [i.e. the species] exist – just as the likeness of humanity [is gathered] from the individuals of men –, likeness which becomes species when it is thought by the rational soul and truly known.’ [‘Tunc ex singulis in quibus sunt eorum similitudo colligitur – ut ex singulis hominibus inter se dissimilibus humanitatis similitudo, quae similitudo cogitata animo ueraciterque perspecta fit species.’ In: In Porphyrii Isagogen commentarium, ed. G. Schepps and S. Brandt (Vindobonae: F. Tempsky, 1906) (csel xxxxviii)] Note that this passage does not entail that the likeness cannot be gathered from one individual alone.

32

chapter 1

In his book L’art des généralités, Alain de Libera puts forth the following reply to these questions: (a) For Boethius a specific concept, for instance, represents an agreement between individuals of the same species: La similitude des espèces exprime leur convenientia – une convenance, c’est-à-dire à la fois une rencontre et un accord, fondés dans l’ordre de l’essence…la similitudo est un concept exprimant une relation entre plusieurs choses (la rencontre de plusieurs en une même essence) ou le fait que des choses x1, x2,…, xn, possèdent chacune une propriété essentielle P1, P2,…, Pn, qui détermine leur rencontre avec les autres…la similitude conceptuelle est moins collectée à partir d’elles que collective.64 Consequently, (b) the concept is a likeness insofar as it represents in a unique understanding the agreement of several individuals as regards their essential features. This would seem to prevent the concept from being formally identical with the form of the external thing, since it would be a mental construct rather than the impression or the transfer of a form. De Libera explains that there is in Boethius the cohabitation of two models of concept formation – one inductive and one non-inductive – a cohabitation that is already present in Alexander of Aphrodisias and that, whether consciously or not, passed on to Boethius. The latter would finally adopt the inductive model, in order to make the concept the representation of a relation of essential agreement between individuals numerically different. However, this reply to the questions is problematic. De Libera grounds his position in the following passage from De divisione: It is particularly useful to know that a genus is in a sense a unified likeness between a number of species, a likeness such as points to the substantial agreement between them all. Hence a genus is collective of a plurality of species whereas the species are disjunctive of a single genus.65 (Trans. Magee, p. 33) Admittedly, Boethius talks here of the genus as expressing a substantial agreement between species, which could also apply to species that are the 64 A. de Libera, L’art des généralités. Théories de l’abstraction (Paris: Aubier, 1999), p. 275. 65 Boethius, De divisione (ed. Magee, 32): ‘Illud autem scire perutile est, quoniam genus una quodammodo multarum specierum similitudo est quae earum omnium substantialem convenientiam monstret, atque collectivum plurimarum specierum genus est, disiunctivae vero unius generis species.’ In: De divisione, ed. J. Magee (Leiden: Brill, 1998) (Philosophia Antiqua 77).

Ancient Sources

33

expression of a substantial agreement between individuals. Nevertheless, this move does not seem to be valid regarding the species that are immediately related to individuals – the species specialissimae. In his commentary on Porphyry’s Isagoge, Boethius indeed puts forth a difference between species that can be further divided into subspecies and species that cannot be further divided: For some species can also be genera, but some other remain only in the species’ essential feature and do not move on to the nature of the genus.66 The latter – the species specialissimae: …are called ‘species specialissimae’, because the whole name of a certain thing is borne by such entities, which are constituted purely and without admixture by the essential property in case.67 In other words, the name of a species specialissima, e.g. ‘man’, applies to all the individuals whose essential feature is to be a man. These species represent the substance of all their individuals: In fact, since the species points out to one substance, which would be the substance of all the individuals that are put under the species, thus [the species] was preferable, if someone wanted to look upon the substance.68 Furthermore, some of these species are associated with an essential feature, e.g. to be a sun or to be a phoenix, which is actually instantiated in only one individual: Sometimes, the species can be at the top of only one individual; for if the phoenix is unique, as the tale goes, the species of the phoenix is predicated 66 Boethius, In Isag.2 (ed. Brandt, 166): ‘Aliae enim sunt species, quae et genera esse possunt, aliae, quae in sola speciei permanent proprietate neque in naturam generis transeunt.’ 67 Boethius, In Isag.2 (ed. Brandt, 207): ‘Specialissimae species appellantur, idcirco quoniam integrum cuiuslibet rei vocabulum illa suscipiunt, quae pura inmixtaque in ea de qua quaeritur proprietate sunt constituta.’ 68 Boethius, In Isag.2 (ed. Brandt, 166): ‘Nam cum species substantiam monstret unam, quae omnium individuorum sub specie positorum substantia sit, quodammodo nulli praeposita est, si ad substantiam quis velit aspicere.’

34

chapter 1

of only one individual. It is thought that the species of the sun also has only one sun as a subject, so that the species does not contain in itself any multiplicity.69 Consequently, genera and superior species would represent something – an essential agreement – that does not seem to be the same as what is represented by a species specialissima, because such agreement would require the existence of at least two individuals of the species, and, as we just saw, some species specialissimae represent a substance that exists in only one individual. Accordingly, Boethius seems to put forth an account that articulates two models of concept formation – one inductive and the other non-inductive – so that the non-inductive model serves to explain the formation of concepts of species specialissimae, while the inductive model serves to explain the formation of genera and of superior species. Therefore, concepts can be either the understanding of a nature as if it were separated from individuals or the understanding of the agreement of several species in one essential feature. In the case of the non-inductive model, since there is a sharing of form between object and concept, ‘for the same thing, [when it is] in the soul, is an affection, and it is a likeness of the thing’,70 the formal identity follows evidently, which would thereby assure and ground the secondary signification of external things. § Boethius’ logical works considerably shaped the reception of Aristotle’s Perihermeneias 1 and, thereby, the development of the thirteenth-century notion of signification. The main Boethian elements that will be retained in the late medieval period and that are not evident features of the Aristotelian treatise are: (a) the strong focus on the notion of signification itself; (b) the strong focus on the semiotic content of simple words – names and verbs – in the passage 1.16a3–8; and (c) the idea of a secondary signification of things 69 Boethius, In Isag.2 (ed. Brandt, 215): ‘At vero species etiam uni aliquando individuo praeesse potest. Si enim unus, ut perhibetur, est phoenix, phoenicis species de uno tantum individuo praedicatur; solis etiam species unum solem intelligitur habere subiectum. Ita nullam multitudinem species per se continet.’ 70 Cf. Boethius, In Perih.2 (ed. Meiser, 35:15–21): ‘Quare quoniam passiones animae quas intellectus uocauit rerum quaedam similitudines sunt, idcirco Aristoteles, cum paulo post de passionibus animae loqueretur, continenti ordine ad similitudines transitum fecit, quoniam nihil differt utrum passiones diceret an similitudines. Eadem namque res in anima quidem passio est, rei uero similitudo.’

Ancient Sources

35

through the primary signification of concepts by virtue of a formal identity between concepts and things. These Boethian elements will be to a great extent accountable for the semiotic and verificational approaches to signification, which are superposed over the communicational one and which will be at the centre of the discussions about signification of concepts and signification of things that we analyse in the following two sections.

chapter 2

Medieval Discussions about Signification of Concepts and Things 2.1

Whether Words (Qua Names) Signify Concepts or Things

As John Duns Scotus relates in his Ordinatio (ca. 1304), the question whether words immediately signify concepts or things was highly debated in his time.1 In fact, the question is extant in at least twelve logical treatises from the thirteenth century, two of them by the Franciscans Roger Bacon (De signis, ca. 1267) and Peter John Olivi (Quaestiones logicales, ca. 1287) and the rest by masters of Arts. It is usually raised in commentaries on Aristotle’s Perihermeneias, but also in commentaries on the passages in Metaphysics iv.4 about the signification of names and on the passage 1.165a6–8 of the Sophistical Refutations, where we are told that we use words in place of things. The main extant versions of the question are by Martin of Dacia (In Perih., 1270s), Peter of Auvergne (In Perih., 1270s), Siger of Brabant (In Met., 1270s), the Incertus sf (In se, late 1270s), Simon of Faversham (In Perih., 1280s), Radulphus Brito (In Perih. and In Met., 1290s) and John Duns Scotus (In Perih.1 and In Perih.2, perhaps 1290s). Scotus’ commentaries, late witnesses to the thirteenth-century debate, present us with an extensive summary of diverse and opposite positions. The debate, however, seems to be rather ancient, for in his second commentary on the Perihermeneias, Boethius gives us the impression that the controversy goes back to at least Alexander of Aphrodisias, who asked himself why Aristotle claims that names signify concepts, even though they are names of things: 1 Duns Scotus, Ordinatio d.27 q.1 (ed. Vatican, 97:3–5): ‘Licet magna altercatio fiat de voce, utrum sit signum rei vel conceptus, tamen breviter concedo quod illud quod signatur per vocem proprie est res.’ For other discussions of this question, namely the ones that were the source of inspiration for my own study, see G. Pini, ‘Signification of Names in Duns Scotus and Some of his Contemporaries’, Vivarium 39 (2001), 20–51; and id., ‘Species, Concept and Thing: Theories of Signification in the Second Half of the Thirteenth Century’, Medieval Philosophy and Theology 8 (1997), 21–52. For signification in Scotus, see C. Marmo, ‘Ontology and Semantics in the Logic of Duns Scotus’, in U. Eco and C. Marmo (eds.), On the Medieval Theory of Signs (Amsterdam: John Benjamins, 1989), pp. 143–193; D. Perler, ‘Duns Scotus’s Philosophy of Language’, in T. Williams (ed.), The Cambridge Companion to Duns Scotus (Cambridge: cup, 2003), pp. 161–192; and id., ‘Duns Scotus on Signification’, Medieval Philosophy and Theology 3 (1993), 97–120.

© koninklijke brill nv, leiden, ���� | doi 10.1163/9789004300132_004

SIGNIFICATION OF CONCEPTS AND THINGS

37

But since Aristotle says: ‘The things of which these are primarily marks – the passions of the soul – are the same for every one’, Alexander asks: If the names are [names] of things, why would Aristotle say that utterances are primarily marks of the intellect? For the name is given to a thing, as when we say ‘homo’, we signify certain concept, but [‘homo’] is the name of a thing, i.e. of a mortal rational animal. Why then wouldn’t utterances be primarily marks of the things to which they are given, rather than of the concepts? But perhaps someone would object that utterances are names of things, however we do not use utterances in order to signify things, but in order to [signify] those passions of the soul that are born in us out of things. Whence, since utterances themselves are uttered in order to signify them, [Aristotle] rightly said that they are primarily marks of [the passions].2 Alexander would reply to his own question by establishing a distinction between the act of naming and the act of signifying, so that a name signifies that which it primarily intends to communicate when it is uttered – a concept.3 Furthermore, as we already saw, Boethius tells us that Porphyry, in his lost commentary, relates about an ancient controversy about the signification of words – whether they signify things, ideas, sensations or imaginations.4 So it is probable that the two remarks about this controversy by Albert the Great and Roger Bacon, in the 1260s, make reference to that ancient controversy, and that the debate mentioned by Scotus is indeed its thirteenth-century revival.5 2 Boethius, In Perih.2 (ed. Meiser, 40:28–41:13): ‘Sed quoniam ita dixit Aristoteles: quorum autem haec primorum notae, eaedem omnibus passiones animae sunt, quaerit Alexander: si rerum nomina sunt, quid causae est ut primorum intellectuum notas esse voces diceret Aristoteles? Rei enim ponitur nomen, ut cum dicimus homo significamus quidem intellectum, rei tamen nomen est, id est animalis rationalis mortalis. Cur ergo non primarum magis rerum notae sint voces quibus ponuntur potius quam intellectuum? Sed fortasse quidem ob hoc dictum est, inquit, quod licet voces rerum nomina sint, tamen non idcirco utimur vocibus, ut res significemus, sed ut eas quae ex rebus nobis innatae sunt animae passiones. Quocirca propter quorum significantiam voces ipsae proferentur, recte eorum primorum esse dixit notas.’ 3 Note that this reading is very similar to the one that I proposed in Section 1.1. 4 Cf. Boethius, In Perih.2 (ed. Meiser, 26:17–27:6); see full quotation above, in Section 1.2, p. 27, n. 48. 5 Albert the Great, In Perih. (ed. Borgnet, 381): ‘Consideratur enim vox significativa ad placitum dupliciter, scilicet secundum institutionem et usum et secundum causam institutionis. (Nota hanc distinctionem: quia per eam solvuntur multa argumenta et salvatur quod voces primo significent res, et etiam quod primo significent conceptus sub diversis respectus).’ In: Liber Perihermeneias, ed. A. Borgnet (Paris: Vivès, 1890) (Opera Omnia 1). Roger Bacon, De signis (ed. Fredborg et al., 162–163): ‘Et est difficilis dubitatio utrum vox significet species

38

chapter 2

As I have shown in Section 1, there are conflicting claims in the authoritative literature of the period about the signification of words. Boethius, for instance, tells us on the one hand that according to Aristotle’s Perihermeneias, words signify concepts; on the other hand, he claims in his commentary on the Categories that this treatise deals with words that signify things.6 Furthermore, in the Sophistical Refutations, Aristotle himself claims that we use words in place of things, and he is usually taken to be saying in Metaphysics iv.4 (1066a31–35) that names signify the essence of things. Finally, Priscian claims too that words represent concepts,7 but he also defines e.g. the name as a word that signifies a substance with a quality (both of which are Aristotelian categories of things or res praedicamentales).8 While Priscian’s claims are only rarely brought into the discussion, Aristotle and Boethius are systematically quoted in arguments of authority in favour of both the immediate signification of concepts and the immediate signification of things. Now, a closer look at the different objections against the immediate signification of concepts in Scotus’ commentary reveals that this position is attacked from at least three different angles, which I shall call the semiotic angle, the categorial angle and the verificational angle. As to the categorial angle, a number of arguments against the immediate signification of concepts are based on the Aristotelian categories. If, as Boethius suggests in his commentary on the Categories, a word’s category is determined by the ontological category of its significate, then, if words signified concepts, all words would fall under the category of quality; for, ontologically speaking, concepts are qualities of the human soul.9 apud animam an res, et quomodo potest significare species et habitus et conceptus mentis…’. In: De signis, ed. K.M. Fredborg, L. Nielsen and J. Pinborg, in ‘An Unedited Part of Roger Bacon’s Opus Maius: De signis’, Traditio 34 (1978), 75–136. 6 Cf. Boethius, In Cat. (pl 64, 160A): ‘Est igitur huius operis intentione vocibus res significantibus in eo quod significantes sunt pertractare.’ In: In Categorias Aristotelis libri quattuor, ed. J.-P. Migne (Paris: Migne, 1860) (Patrologia Latina 64). 7 Priscian, ig xi.2 (ed. Hertz, 552): ‘Quid enim est aliud pars orationis nisi vox indicans mentis  conceptum, id est cogitationem?’ In: Institutiones grammaticae libri xviii, 2 vols., ed. M. Hertz (Leipzig: Teubner, 1855–1859) (gl ii–iii). 8 For this definition of names and the particular questions that emerge from its comparison with Aristotle’s definition, see below, Part 2. 9 Cf. Duns Scotus, In Perih.1 (ed. Bonaventure, 50:4–5): ‘Tum quia tunc omne nomen significaret accidens, quia illa species est in anima ut in subiecto, sicut species visibilis in oculo.’ In: Quaestiones in duos libros Perihermeneias, ed. R. Andrews et al. (St. Bonaventure, ny: The Franciscan Institute, 2004) (Opera philosophica 2). Recent translation in E. Buckner and J. Zupko, Duns Scotus on Time and Existence. The Questions on Aristotle’s ‘De interpretatione’ (Washington DC: CUAP, 2014).

SIGNIFICATION OF CONCEPTS AND THINGS

39

To the best of my knowledge, only two authors use a categorial argument against the immediate signification of concepts, namely Radulphus Brito and Walter Burley. Brito, for instance, says: Likewise, if every name signified a concept, then, since the concept is an accident, the name would signify an accident, and since it falls under a category by means of what is signified, then the name would fall under the category of the accident.10 Yet, as Scotus remarks in his commentary, this argument is easily refuted as concepts can be seen from two perspectives: (i) As accidents of the soul, and as such they do not determine the category of the word. (ii) As likenesses of external things, so that the category of the concept’s object determines the category of the word: 10

Radulphus Brito, In Met. (ed. Ebbesen, 111): ‘Item, si omne nomen significaret intellectum, tunc, cum intellectus sit accidens, tunc nomen significaret accidens, et cum per id quod significatur per aliquid reponatur in praedicamento, tunc nomen esset in praedicamento accidentium.’ In: Quaestiones super Metaphysicam, partial edition in S. Ebbesen, ‘Words and Signification in 13th-century Questions on Aristotle’s Metaphysics’, cimagl 71 (2000), 107–114. For a reconstruction of Brito’s theory of signification, see A.M. Mora-Márquez, ‘Radulphus Brito on Common Names, Concepts and Things’, in J. Fink, H. Hansen and A.M. Mora-Márquez (eds.), Logic and Language in the Middle Ages: A Volume in Honour of Sten Ebbesen (Leiden: Brill, 2013) (Investigating Medieval Philosophy 4), pp. 357–372. See also S. Ebbesen, ‘Radulphus Brito: The last of the great masters, or Philosophy and Freedom’, in J.A. Aertsen and A. Speer (eds.), Geistesleben im 13. Jahrhundert (Berlin: De Gruyter, 2000), pp. 231–251. Cf. Walter Burley, In Perih. (ed. Brown, 51): ‘Vel aliter, ut dicit Boethius super Praedicamenta quod in libro Praedicamentorum determinatur de vocibus ut res significant et hic in isto libro determinatur de vocibus ut significant intellectus. Et sic in libro Praedicamentorum determinatur de vocibus primae impositionis sed in isto libro determinatur de vocibus secundae impositionis. Voces primae impositionis sunt quae significant res, distinguendo res contra voces, et dicuntur nomina rerum. Sed voces secundae impositionis sunt quae significant voces… et dicuntur nomina nominum. Unde nomina primae intentionis et nomina secundae intentionis sunt nomina primae impositionis, quia significant res; sed “nomen”, “verbum” et hoc quod dico “terminum”…dicuntur nomina secundae impositionis.’ In: Commentarius in librum Perihermeneias, in S.F. Brown, ‘Walter Burley’s Middle Commentary on Aristotle’s Perihermeneias’, Franciscan Studies 33 (1973), 45–134. For other discussions of Burley’s theory of signification, see J. Biard, Logique et théorie du signe au 14e siècle (Paris: Vrin, 1989), pp. 136–161; L. Cesalli, ‘Meaning and Truth’, in A. Conti (ed.), A Companion to Walter Burley (Leiden: Brill, 2013), pp. 87–134; id., Le réalisme propositionnel: sémantique et ontologie des propositions chez Jean Duns Scot, Gauthier Burley, Richard Brinkley et Jean Wyclif (Paris: Vrin, 2007); and A. Conti, ‘Significato e verità in Walter Burley’, Documenti e studi sulla tradizione filosofica medievale 11 (2000), 317–350. For Walter Burley in general, see A. Conti (ed.), A Companion.

40

chapter 2

To the first [position], one must answer that it is not inconvenient for the name to immediately signify an accident, not insofar as it [i.e. the accident] is something in itself, but insofar as it is a sign of a thing. And thus, some utterances signify a substance as the ultimate significate, and thus they are said to signify substance absolutely.11 Another objection proposes that if names signified concepts (which are accidents of the soul) and, through concepts, things that fall under any of the ten Aristotelian categories (e.g. a substance), every name would be equivocal as it would signify both a substance and an accident. However, a further reply remarks that names are equivocal when they have different significates in different acts of signification, but a word signifies the concept and the thing in the same act of signification (i.e. the thing by means of the concept), and thus there is no equivocation: Against these [arguments]: No substantial notion is the same for the substance and for the accident, since they do not have the same general genus. But the likeness of a substance is an accident. Therefore, if these [i.e. the substance and the accident] are signified with some name, both of them would have ‘only the name in common, but diverse substantial notion’. Hence, every name would be equivocal… To this, it is said that the equivocal name signifies many things in different acts of signification; but the utterance signifies the thing and the likeness with respect to the same act, because with respect to the same act the utterance is sign of the sign qua sign and of its content (signati).12 No trace of this argument can be found in the extant versions of the question about the immediate signification of words, nor could I find a discussion of such an issue in the commentaries on the Categories from the same period. 11

12

Duns Scotus, In Perih.1 (ed. Bonaventure, 52:21–53:4): ‘Ad primam respondetur quod non est inconveniens omne nomen significare accidens immediate, non in quantum est quid in se, sed ut signum rei. Et ita aliquae voces significant substantiam ut ultimum significatum, et ita dicuntur absolute significare substantiam.’ Duns Scotus, In Perih.1 (ed. Bonaventure, 51:15–52:9): ‘Contra ista: substantiae et accidenti nulla ratio substantialis est eadem, cum nec habeant idem genus generalissimum; sed rei quae est substantia, similitudo illius est accidens; ergo si haec significentur per aliquod nomen, istis duobus erit “solum nomen commune et ratio substantiae diversa”; igitur omne nomen erit aequivocum… Ad illud dicitur quod aequivocum diversis actibus significandi significat multa; sed vox est significans rem et similitudinem eodem actu, quia eodem actu est vox signum signi in quantum signum et signati eius…’.

SIGNIFICATION OF CONCEPTS AND THINGS

41

There is indeed a discussion about the double categorisation of words such as ‘body’ and ‘science’, but in this discussion the problem at stake is different; for the problem of double categorisation deals with the possibility that one word with one and the same significate falls under more than one category (e.g. quantity and substance in the case of ‘body’ and quality and relation in the case of ‘science’). Hence, ‘body’, for instance, is not considered an equivocal word – it does not have different significates – but rather a univocal word with a double categorisation. Medieval commentators on the Categories are concerned with the notion of predicability, or with sayables (dicibiles), or with significant words insofar as they are susceptible of being attributed to a subject.13 Furthermore, there is some agreement that metaphysical categories and categories of attribution are not necessarily the same, so that things qua beings only fall under one category of being, even though words can have a double category of attribution. The way to tackle the problem of the double categorisation of words differs from one author to another and goes beyond the purpose of this study – it is a complex logico-metaphysical problem that, to the best of my knowledge, has no outstanding implications for the development of the notion of signification in the thirteenth century. I shall therefore leave the categorial argument aside and focus on the semiotic and verificational angles. The Semiotic Angle The Immediate Signification of Concepts From the semiotic angle the question of the immediate signification of words is closely related to the question of their imposition on their content and hence the question amounts to asking whether words qua signs are related to concepts or to things. In the case of the so-called common names of first imposition (e.g. ‘man’, ‘horse’ and ‘stone’),14 which are at the core of most of the discussions from the second half of the century, the conundrum is their applicability to many things.15 The question, then, asks on which kind of item the 13

14 15

See e.g. Nicholas of Paris, In Praed. L1 ([Rationes super Praedicamenta, ms Münich, Bayerische Staatsbibliothek Clm. 14460, ff. 42ra–62ra] ed. Hansen, forthcoming): ‘Est ergo causa materialis sive subiectum huius libri dicibile incomplexum ordinabile.’ Names of second imposition are e.g. ‘name’, ‘verb’, ‘subject’, ‘predicate’ etc. Contrary to some medieval theologians (e.g. Thomas Aquinas) for whom the object of imposition and the object of signification are not necessarily the same thing, for the authors under consideration in this study a word’s significate is that on which the utterance is imposed. The distinction between object of imposition and object of signification is at least as old as Boethius, for whom names are imposed on the things they name, although they signify thoughts. In the 12th century a similar distinction is made by e.g. Peter Abelard; see C.J. Martin, ‘The Development of Logic’, in R. Pasnau (ed.), The Cambridge History of

42

chapter 2

common name of first imposition is imposed – a thing or its concept – and how one further explains its applicability to many things. Contrary to Boethius, who in the beginning of his commentary on the Categories claims that names of first imposition are imposed on things,16 most authors from the first half of the thirteenth century hold that words are imposed on, and consequently signify, concepts. The question they raise is rather why concepts are signified by an object of the sense of hearing – an utterance – and not by the object of any other sense.17 Their common answer to the question is that a sign and its significate ought to have the same sort of Medieval Philosophy. 2 vols. (Cambridge: cup, 2010), Vol. 1, pp. 129–132; and id., ‘Imposition and Essence: What’s new in Abelard’s theory of meaning’, in T. Shimizu and C. Burnett (eds.), The Word in Medieval Logic, Theology and Psychology (Turnhout: Brepols, 2009), pp. 173–214. It is also common to find this distinction in 12th-century theological discussions about language, as well as in 12th-century grammar. For 12th-century grammar, see I. RosierCatach, ‘Grammar’, in R. Pasnau (ed.), The Cambridge History, Vol. 1, pp. 196–218, at pp. 196– 200. For 12th-century theologians, see L. Valente, Logique et théologie (Paris: Vrin, 2008); and J. Marenbon, ‘Gilbert of Poitiers’s Contextual Theory of Meaning’, in Fink et al. (eds.), Logic and Language, pp. 49–64. For Thomas Aquinas, see J.E. Ashworth, ‘Signification and Modes of Signifying in Thirteenth-Century Logic: A Preface to Aquinas in Analogy’, Medieval Philosophy and Theology 1 (1991), 39–67; and I. Rosier-Catach, ‘Res significata et modi significandi: les implications d’une distinction médiévale’, in S. Ebbesen (ed.), Sprachteorien in Spätantiken und Mittelalter (Tübingen: Gunter Narr Verlag, 1995), pp. 135–168. 16 Boethius, In Cat. (pl 64, 159B–C): ‘Prima igitur illa fuit nominum positio, per quam vel intellectui subiecta vel sensibus designaret. Secunda consideratio, qua singulas proprietates nominum figurasque perspicerent, ita ut primum nomen sit ipsum rei vocabulum: ut, verbi gratia, cum quaelibet res homo dicatur. Quod autem ipsum vocabulum, id est homo, nomen vocatur, non ad significationem nominis ipsius refertur sed ad figuram, idcirco quod possit casibus inflecti. Ergo prima positio nominis secundum significationem vocabuli facta est, secunda vero secundum figuram: et est prima positio, ut nomina rebus imponerentur, secunda vero ut aliis nominibus ipsa nomina designarentur.’ 17 Robert Kilwardby, Notulae super Perihermeneias ([mss Cambridge, Peterhouse 206, ff. 13va–21ra; Madrid, Biblioteca Universitaria 73, ff. 44ra–66va; Venezia, Biblioteca Marciana L.vi.66, 2r–18v] M46rb; P67vb; V3r): ‘Quaeritur postea propter quid passio sive intellectus significatur per obiectum auditus sicut per vocem, et per obiectum visus sicut per litteram, et non per obiectum aliorum sensuum…’; Nicholas of Paris, In Perih. (ed. Hansen and Mora-Márquez, 27:23–24): ‘Quaeritur propter quid per vocem intellectus repraesentatur potius quam per aliud sensibile signum.’ In: H. Hansen and A.M. Mora-Márquez, ‘Nicholas of Paris on Aristotle’s Perihermeneias 1-3’, cimagl 80 (2011), 1–88. Cf. AnOx, In Perih. (ed. Mora-Márquez, 187:10–15): ‘Quia primo dicit quod figura est signum vocis et quod vox sit signum passionis, cum ergo vox sit obiectum auditus, figura autem visus, cum sint alia sensibilia, propter quid magis ista duo (sensibile visus et sensibile auditus) imponuntur ad significandum quam sensibilia aliorum sensuum imponuntur.’ In: Mora-Márquez,

SIGNIFICATION OF CONCEPTS AND THINGS

43

being. Since the word’s significate (i.e. the concept) is a mental being, the being of signs has to agree with the mental being of concepts. Yet only utterances (and more precisely articulate utterances) are as such; for, contrary to things whose only cause is nature (e.g. men and horses), the causes of articulate utterances are both nature and the rational faculty of the soul, because humans produce articulate utterances as the result of their rational will to communicate. Utterances are thus partly mental by virtue of their efficient cause and are, therefore, suitable to be imposed on concepts. In Robert Kilwardby’s words (perhaps 1240s): It must be said that, since the sign and its content are proportional, and the content is a likeness of the thing in the rational soul, it is necessary for this sign to have something on the part of the rational soul and something on the part of the thing. But sense-objects other than the utterance and the letter are totally by nature and in no way by will or by the rational soul, and because of this it is manifest that they could not be signs of concepts…And thus, it is evident that the utterance is a sign of the concept; for it has something on the part of the thing – matter; for air is its matter – and something on the part of the rational soul, since [the rational soul] is its efficient cause…18

18

‘Anonymus Oxford’. I owe thanks to Alessandro Conti for having given me access to Lewry’s transcription of Kilwardby’s commentary. Robert Kilwardby, Not. sup. Perih. (M46rb; P67vb; V3r): ‘Dicendum quod cum signum et signatum sit proportionalia, et signatum est similitudo rei in anima rationali, oportet quod signum huiusmodi habet aliquid ab anima rationali et aliquid a re. Set sensibilia alia a voce et figura omnino sunt a natura et nullo modo a voluntate sive ab anima rationali, et propter hoc manifestum est quod non poterant esse signa intellectuum… Et sic patet quod vox sit signum intellectus: haec enim aliquid habet a re, scilicet materiam, materia enim eius est aer, et aliquid ab anima rationali, cum sit eius principium effectivum.’ Cf. AnOx, In Perih. (ed. Mora-Márquez, 190:1–8): ‘Ad illud quod primo quaeritur, dicendum quod sunt quaedam sensibilia quorum natura est principium solum, quaedam vero quorum natura et anima. Sensibile illud cuius natura est principium solum est sicut color et durum et molle, huiusmodi non imponitur ad significandum; aliud est autem sensibile cuius natura est principium et anima, sicut vox, quia vox est sonus procedens ab ore animalis, et huiusmodi potius debet imponi ad significandum, eo quod signum et signatum convenientiam debent habere inter se…’. For Robert Kilwardby’s semantics, see P.O. Lewry, ‘Robert Kilwardby on Meaning: A Parisian Course on the “Logica Vetus”’, in J.P. Beckmann et al. (eds.), Sprache und Erkenntis im Mittelalter (Berlin: De Gruyter, 1981), pp. 376–384; and id., Robert Kilwardby’s Writings on the Logica Vetus with Regard to their Teaching and Method (Oxford: Ph.D. dissertation, 1978). For a recent collection of articles on Robert Kilwardby, see H. Lagerlund and P. Thom (eds.), A Companion to the Philosophy

44

chapter 2

Some years later Ps.-Kilwardby addresses this issue and explains in detail the mechanisms of the imposition of utterances on concepts. The first question he raises is whether the imposition of utterances with the purpose of signification is possible at all.19 There is first the claim that such imposition is not possible, because to impose an utterance is to provide it with signification by joining it to a significate. Yet the union of an utterance and a concept is impossible, because the utterance is a material thing and the concept is a mental thing, and things that are substantially different, and in different places, cannot be joined. Therefore, the imposition of utterances on concepts is not possible at all.20 Ps.-Kilwardby opposes that it is evident that words have signification by imposition – words do not come to being naturally, but they are coined; for, as Aristotle himself remarks, words are not the same for all people, therefore they do not have natural signification, and hence they signify by convention (ex institutione): Likewise, the Philosopher says in Perihermeneias 1 that utterances are signs of concepts. Concepts and things are the same for everyone, but utterances are not the same for everyone. From this, it is argued: If significative utterances, as names and verbs, are not the same for everyone, of Robert Kilwardby (Leiden: Brill, 2013) (Brill’s Companions to the Christian Tradition 37); for Kilwardby’s philosophy, see J.F. da Silva, Robert Kilwardby on the Human Soul. Plurality of Forms and Censorship in the Thirteenth Century (Leiden: Brill, 2013) (Investigating Medieval Philosophy 3). 19 Ps.-Kilwardby, In Prisc. mai. (ed. Fredborg et al., 49): ‘Quaeritur ergo primo an sit possibilis  institutio vocis ad significandum.’ In: Commentum super Priscianum maiorum, ed. K.M. Fredborg, N.J. Green-Pedersen, L. Nielsen and J. Pinborg, in ‘The commentary on “Priscianus Maior” ascribed to Robert Kilwardby’, cimagl 15 (1975), 1–146. Cf. John of Dacia, Summa (ed. Otto, 178:14–28), in: John of Dacia, Summa grammatica, in Johannis Daci Opera, ed. A. Otto (Copenhagen: gad, 1955) (CPhD I.1–2). See also C. Marmo, Semiotica e linguaggio nella Scolastica: Parigi, Bologna, Erfurt 1270–1330. La semiotica dei modisti (Rome: Istituto Storico Italiano per il Medioevo, 1994), Ch. 1. 20 Ps.-Kilwardby, In Prisc. mai. (ed. Fredborg et al., 49–50): ‘Videtur quod non. Ad hoc quod sit institutio vocis ad significandum necesse est speciem intelligibilem uniri voci. Sed haec unio est impossibilis, ergo institutio est impossibilis. Maior patet quia instituere vocem ad significandum nihil aliud est quam dare voci significationem quam prius non habebat. Dare autem voci significationem quam prius non habebat est ei unire significationem. Significatio autem non est nisi species intelligibilis… Minor patet quia quaecumque necesse est distare secundum substantiam et situm impossibile est ea coniungi. Sed species intelligibilis quae dicitur significatio et vox sunt huiusmodi…quod patet quia species intelligibilis est in anima, vox autem sensibilis in aere extra… Quare impossibili est ea coniungi…’. Note that this is the same concern that we find in Kilwardby and AnOx.

SIGNIFICATION OF CONCEPTS AND THINGS

45

therefore they do not signify naturally, because natural things are the same for everyone. Therefore they only signify by institution.21 Having determined that the imposition of utterances on concepts is indeed possible, Ps-Kilwardby goes on to discuss the character of the union that results from the imposition of an utterance on a concept. Several possibilities are put forth, the first being that the concept and the utterance are joined just as form is joined to matter. This possibility is rejected because utterances perish but concepts remain, which would not be the case if concepts were joined to utterances as their form.22 Another possibility is that they are joined just as a thing and its likeness, and this can happen either immaterially or materially. Immaterially, just as the objects of intellectual knowledge and their intellectual representation; materially, just as an object and its image in a sensitive organ or in a mirror. Yet the utterance is not an object of intellectual knowledge because it is a sensible thing, and as a result it cannot be immaterially joined to the concept; and the concept is not corporeal, so it cannot be materially joined to the utterance.23 21 Ps.-Kilwardby, In Prisc. mai. (ed. Fredborg et al., 51): ‘Item: I. Perihermeneias dicit Philosophus quod voces sunt signa intellectuum. Sed intellectus et res sunt eadem apud omnes, voces autem non sunt eaedem apud omnes. Ex quo arguitur: si voces significativae ut sunt nomina vel verba non sunt eaedem apud omnes, ergo non significant naturaliter, quia naturalia sunt eadem apud omnes; ergo ex institutione significant solum.’ Note that for Ps.-Kilwardby institution and imposition are equivalent; see e.g. the following passage [Ps.-Kilwardby, In Prisc. mai. (ed. Fredborg et al., 51–52)]: ‘Item: patet per rationes; quia illud videtur possibile seu contingens quo non existente positum in esse nullum accidit inconveniens. Sed dato quod nulla facta fuisset impositio sive institutio vocis ad significandum sicut quondam fuit, ex ipsa institutione facta nullum accideret inconveniens sed potius multae commoditates. Ergo ipsa institutio ad significandum est possibilis.’ 22 Ps.-Kilwardby, In Prisc. mai. (ed. Fredborg et al., 53–54): ‘Sed tunc supposito quod intellectus uniatur aliquo genere unionis quaeritur quid sit illud. Aut enim unitur ei secundum substantiam ut forma materiae… Non primo modo quia cum intellectus naturaliter praecedat ipsam vocem, quare non unitur ei ut forma materiae, tum quia vox fluxibilis est et intellectus manet, quod non contingeret si esset eius forma…’. 23 Ps.-Kilwardby, In Prisc. mai. (ed. Fredborg et al., 54–55): ‘…aut secundum sui similitudinem et hoc dupliciter: aut per modum spiritualem ut res cognoscibiles ab anima uniuntur sibi per sui similitudinem vel speciem spiritualem, aut per modum corporalem et hoc dupliciter: aut sicut idolum in organo cognoscendi…aut sicut idolum in speculo materiali…quod non contingit, quia nihil unitur alicui per modum spiritualem nisi ipsum sit sensus spiritualis. Vox autem sensibilis non est huiusmodi… Si uniretur sicut idolum organo cognoscendi, tunc illud iterum esset impossibile, cum ipsa vox non sit sicut organum cognoscendi. Item: solum corporale idolum habet. Species autem intelligibilis non

46

chapter 2

Along the lines of Kilwardby and of AnOx, Ps.-Kilwardby places the concept and the utterance in the mental realm by introducing the idea of a mental utterance, so that the external utterance would be imposed through the imposition of the mental utterance on a concept: [I]t is asked…with respect to which utterance is the institution made and how – whether with respect to the internal mental utterance or to the external utterance that is a sense-object; [and] it must be said that with respect to both, but the institution of the exterior [utterance is made] by means of the interior [utterance].24 However, Ps.-Kilwardby takes a step further by explaining in detail the psychological mechanisms that take place during the production of a word. Once a speaker has acquired the concept of a thing and once the will to communicate it to someone else has arisen, the rational part of the soul devises a mental utterance that is joined to the concept, thus producing a mental word. Thereafter, since a listener cannot understand an utterance unless it has a material bearer, the rational soul provokes the movement of the bodily organs that participate in the production of utterances, so that the mental word takes an external utterance as a sensible vehicle capable of reaching the intellect of a listener: Hence, once the soul has an intention susceptible of being signified, immediately a pre-invention of the utterance is produced by which such intention or intellection must be signified; and the intention susceptible of being signified is attached to the intention of the utterance, just as the end [is attached] to that which is directed towards the end. Consequently, since a sensible sign is required in order to make public this concept to est corporalis, ergo non unitur voci ut idolum organo cognoscendi. Nec etiam ut species obiecti corporis speculo materiali, quia tunc contingeret quod spirituale per sui unionem cum materiali fieret materiale.’ The possibilities that the utterance and the concept are joined as a means to the end, as the mover to the moved, as the wax to the ring and as a sign to its significate are also taken into consideration. 24 Ps.-Kilwardby, In Prisc. mai. (ed. Fredborg et al., 57): ‘…quaeritur…scilicet in qua voce fiat institutio et qualiter an scilicet in voce mentali interiori aut in voce sensibili exteriori, dicendum quod in utraque, sed in exteriori fit institutio interiori mediante.’ See also C. Panaccio, ‘Grammar and Mental Language in the Pseudo-Kilwardby’, in S. Ebbesen and R. Friedman (eds.), Medieval Analyses in Language and Cognition: Acts of the symposium “The Copenhagen School of Medieval Philosophy, January 10–13, 1996” (Copenhagen: Royal Danish Academy of Sciences and Letters, 1999), pp. 397–414.

SIGNIFICATION OF CONCEPTS AND THINGS

47

someone else (because there is nothing in the intellect that was not in the senses before), the rational soul moves by its desire and imagination the bodily parts whose purpose is the formation of the utterance… and the perceptible utterance is formed along with the intention of the utterance that was pre-invented and pre-conceived in the mind of the speaker, as well as is produced the external speech which embraces and represents the same thing that was embraced and represented by the internal speech.25 The external utterance qua word signifies the same thing that is signified by the mental utterance qua mental word. This external utterance is therefore related to: (i) the mental utterance of which it is a likeness as a natural sign (because the effect is a natural sign of its cause), (ii) the mental word composed of the mental utterance and the concept, (iii) the concept and (iv) the external thing. Regarding the relations (ii), (iii) and (iv), the external word is in general a sign by imposition,26 since the act of the will by which the mental word was produced plays an essential role in the coming into being of these three relations. Thus, Ps.-Kilwardby resolves the problem raised by earlier authors by splitting the production of the utterance into a mental part and an external part and by spelling out in great detail the mechanisms of the utterance’s transition from the mental realm to the external world. The requirement that utterances must be imposed on concepts undoubtedly lies behind Martin of Dacia’s argument in favour of the immediate signification of concepts (in question 3 of his commentary on the 25 Ps.-Kilwardby, In Prisc. mai. (ed. Fredborg et al., 59): ‘Apud animam igitur statim cum habet intentionem significabilem fit praeexcogitatio vocis, qua talem intentionem sive intelligentiam deceat vel oporteat significari, et illi intentioni vocis applicatur intentio significabilis sicut finis ei quod est ad finem. Consequenter quia ad hoc quod huiusmodi intellectus alii manifestetur exigitur aliquod signum sensibile, quia nihil est in intellectu quod prius non fuerit in sensu, movet anima rationalis per appetitum et imaginationem membra deputata ad formationem vocis…et formatur vox sensibilis iuxta intentionem vocis praecogitatae et praeconceptae apud animam proferentis eam, et fit sermo exterior idem continens et repraesentans quod per sermonem interiorem continebatur et repraesentabatur.’ 26 Ps.-Kilwardby, In Prisc. mai. (ed. Fredborg et al., 59): ‘Et igitur vox exterior sensibilis habet quadruplicem comparationem: unam ad intentionem vocis interioris ad cuius similitudinem figuratur, aliam ad intellectum seu similitudinem rei, tertiam ad ipsum sermonem interiorem complentem tam speciem significabilem quam vocis intentionem, quartam ad rem extra quae per vocem significatur intellectu movente. Respectu primi est signum naturale…sed respectu secundi, tertii et quarti est significativum ex institutione.’

48

chapter 2

Perihermeneias – perhaps the earliest extant version of the question about the immediate signification of words in a question-commentary).27 In his reply to the question, Martin first puts forth Boethius’ claim that names of first imposition are imposed on things as an argument in favour of the immediate signification of things.28 Thereafter, Martin briefly defends his own position of an immediate signification of concepts by saying that names of first imposition signify both a concept and an external thing, but primarily the concept, and the thing through the concept, because names of first imposition are imposed on things according to their mental being: If [the question regards] the thing according to its third mode [i.e. the thing that exists outside the soul, such as wood, stone etc.], then I say that the utterance signifies both the passion in the soul and the thing outside the soul, however the signification of the passion is prior. The proof of this is evident from Metaphysics iv, where it is said that utterances are imposed on things only according to their being imagined in the intellect. Therefore, [utterances] signify priorly the passions.29 27

28

29

Martin of Dacia, In Perih. (ed. Roos, 242:1–2): ‘Quaestio est utrum voces primo significant passionem in anima aut res extra animam.’ In: Quaestiones super Peri hermeneias, in Martini de Dacia opera, ed. H. Roos (Copenhagen: gad, 1961) (CPhD ii). For an exhaustive analysis of Martin’s question, see A.M. Mora-Márquez, ‘Martinus Dacus and Boethius Dacus on the Signification of Terms and the Truth-value of Assertions’, Vivarium 52 (2014), 23–48; see also J. Pinborg, ‘Bezeichnung in der Logik des xiii. Jahrhunderts’, in A. Zimmermann (ed.), Der Begriff der repraesentatio im Mittelalter (Berlin: De Gruyter, 1971) (Miscellanea Mediaevalia 8), pp. 238–281. Martin of Dacia, In Perih. (ed. Roos, 242:8–10): ‘Oppositum patet per Boethium qui dicit, quod intellectus rebus quas vidit nomen imponit. Sed res quas vidit sunt extra animam. Ergo et cetera.’ Cf. Boethius, In Cat., cited above, p. 42, n. 16. Martin of Dacia, In Perih. (ed. Roos, 243:4–9): ‘Si de re tertio modo (i.e. res extra animam ut lignum, lapis et huiusmodi), sic dico quod vox significat et passionem in anima et rem extra animam, per prius tamen est significatum passionis. Cuius probatio patet ex iv. Metaphysicae, ubi dicitur, quod voces non imponuntur rebus nisi secundum quod habent esse imaginatum in intellectu. Ergo per prius significant passiones.’ Martin supposedly refers to Met. Δ4.1015a11–13, where there is neither mention of imposition nor of imagined being. Met. Δ4 deals with the different senses of ‘nature’, and the passage in question talks about metaphor, i.e. the transfer of a name. However, Aristotle explicitly says in De an. ii.8.420b29–421a6 that significative utterances always indicate an image (a phantasia). It is also worth mentioning that Thomas Aquinas indeed talks about prior and posterior imposition in his own commentary on the Metaphysics. Cf. Thomas Aquinas, In Met. V, c.4, l.5, n. 824: ‘Nomina enim imponuntur a nobis secundum quod nos intelligimus, quia nomina sunt intellectuum signa. Intelligimus autem quandoque priora ex posterioribus.

SIGNIFICATION OF CONCEPTS AND THINGS

49

Despite the elusive character of Martin’s argument, it suggests not only that it is possible to impose names of first imposition on concepts, but also that for names to be applied to things they must first be imposed on their concepts – an argument that stands very close to the explanation that we saw in earlier authors such as Kilwardby and Ps.-Kilwardby. This is also what is at stake for Albert the Great, who rejects the possibility of a cognitive access to the essences of external things and, thereby, rejects the possibility of directly imposing names on them. Albert claims that according to their imposition, utterances signify concepts, because at the moment of imposition the intellect only has cognitive access to the concept of the thing: In fact, the conventional significative utterance is considered in two ways: according to its institution and use, and according to the cause of its institution… According to its institution and use, it is a mark of the concept in the soul. However, when instituting [it], the one who institutes only has a [cognitive] relation to this or that which he already conceives in the soul.30 Nevertheless, he grants that according to the cause of the imposition (i.e. the final cause), names signify things, because they are imposed in order to allow human communication about the world: The cause of the institution: Since we cannot carry things with us, articulate utterances were invented in order to communicate with each other by means of a [re]presentation of the thing; [utterances] by means of which we talk about the things themselves or about their intentions, so that we communicate with each other with them. And in this last way, utterances are referred to the signification of things.31

30

31

Unde aliquid per prius apud nos sortitur nomen, cui res nominis per posterius convenit: et sic est in proposito.’ See also Mora-Márquez, ‘Martinus Dacus and Boethius Dacus’, 28–31. Albert the Great, In Perih. (ed. Borgnet, 381): ‘Consideratur enim vox significativa ad placitum dupliciter, scilicet secundum institutionem et usum et secundum causam institutionis… Secundum institutionem quidem et usum nota est conceptus qui est in anima. Instituens autem non habet respectum in instituendo nisi ad id vel illud quod jam in anima concepit.’ In: Liber Perihermeneias, ed. A. Borgnet (Paris: Vivès, 1890) (Opera Omnia 1). For logic at the time of Albert the Great, see J. Brumberg-Chaumont (ed.), L’Organon dans la translatio studiorum à l’époque d’Albert le Grand (Turnhout: Brepols, 2013). Albert the Great, In Perih. (ed. Borgnet, 381): ‘Causa autem institutionis, quia cum res efferre non possumus, ut nobis invicem communicemus rerum praesentatione, inventae sunt voces articulatas, quibus ipsas res sive intentiones rerum exprimamus, ut nobis invicem

50

chapter 2

Hence, Albert maintains that from a semiotic point of view, utterances are immediately imposed on concepts because imposition on x requires cognition of x, but the intellect has direct cognitive access only to concepts. However, from the point of view of the purpose of communication, utterances signify things, because words are coined with the purpose of public communication about the world. In this way, Albert articulates the ancient communicational approach to the question of the immediate signification of utterances32 with the semiotic approach to the question, which had gained the upper hand in the first half of the thirteenth century. Consequently, Albert presents us with a semiotic explanation of the immediate signification of concepts, while at the same time accepting that insofar as words are used in order to talk about the world, they signify things. Thus, ‘to signify x’ has both the sense of ‘to be a conventional sign of x (primarily a concept)’ and the sense of ‘to refer to or to mark out x as the object of communication (primarily a thing)’. Moreover, Albert associates the first imposition on concepts with the applicability of names to many external things; a problem that is essentially related to the logico-ontological problem of universals. In his commentary on Porphyry’s Isagoge, Albert tells us that the intellect has cognitive access only to the concept of the thing, because the object of intellectual knowledge needs to be universal, yet for Albert everything that exists outside the intellect is ineluctably singular: The utterance, insofar as it is referred to the concept of someone who wants to discover what is unknown by means of what is known, is divided into common (or appellative or universal) and proper (or singular); for the utterance is neither of them [i.e. neither common nor singular], insofar as it is referred to the thing that was marked out. In fact, all the things that fall under the senses, and according to their natural constitution, are singular, and whatever is common in them is taken by the intellect.33

32

33

vocibus communicemus. Et hoc modo ultimo voces ad significata rerum referuntur.’ Here Albert makes reference to Aristotle’s se 1.165a10–13, where we are told that we use names in the place of things. With the difference that Albert’s focus, following in Boethius’ footsteps, is on the transmission of information about the world, and not exclusively on the transmission of judgments or opinions, which was Aristotle’s focus. Albert the Great, In Isag. (ed. Santo Noya, 9:68–10:3): ‘Adhuc autem vox, secundum quod refertur ad intellectum eius, qui quaerit invenire ignotum per notum, habet quod dividitur in commune sive appellativum sive universale, et proprium sive singulare. Talium enim nihil accidit ei, secundum quod ad rem designatam refertur. Res enim omnes sunt singulares, quae cadunt sub sensu et secundum quod constituuntur a natura, et commune, quod est in eis, accipitur ab intellectu.’ In Albert the Great, Super Porphyrium De quinque

SIGNIFICATION OF CONCEPTS AND THINGS

51

Hence, the names that are used as terms in a syllogism – the logical method that leads to the discovery of what is unknown from what is known – are said to be universal only insofar as they indicate universal concepts, because there is nothing universal in material things and, therefore, names cannot be universal insofar as they stand for things.34 Accordingly, Albert also strongly suggests that a common name can be applied to many things through its signification of something that is common – a concept – which is formally identical to the singular essences of external things: It must be said that it is true that the universal essence is one in number in the soul, in itself and in the singular, and it only differs according to the being that determines it as this and that.35

34

35

universalibus, ed. M. Santos Noya (Münster: Aschendorff, 2004) (Opera Omnia, Editio Coloniensis, 1.1). Here Albert appeals to Ps.-Kilwardby’s idea that the external utterance is the expression of an internal utterance, but taking as his immediate source Damascena’s De fide orthodoxa. See Albert the Great, In Isag. (ed. Santos Noya, 7:7–13): ‘Tali autem sermone, secundum quod sic designativus est concepti, utitur homo et ad se ipsum et ad alium. Propter quod dicit Damascenus quod in duo dividitur, scilicet in endiadentum, hoc est sermonem interius in mente dispositum, et in eum qui, exterius prolatus, angelus intelligentiae est sive cordis nuntius, qui conceptus cordis nuntiat ad alterum.’ Cf. Damascenus, De fide orthodoxa. Versions of Burgundio and Cerbanus, E.M. Buytaert ed. (New York: The Franciscan Institute, 1955), p. 135. The formation of the universal concept consists in a series of separations from matter, which result in the reception of an intelligible form in the possible intellect – the cognitive faculty of the soul – where essences exist as universals and as objects of knowledge; cf. Albert the Great, In De anima (ed. Stroick, 101:62–102:15): ‘Dicimus igitur, quod omne apprehendere est accipere formam apprehensi, non secundum esse, quod habet in eo quod apprehenditur, sed secundum quod est intentio ipsius et species, sub qua aliqua sensibilis vel intellectualis notitia apprehensi habetur. Haec autem apprehensio, ut universaliter loquendo, quattuor habet gradus. Quorum primus et infimus est, quod abstrahitur et separatur forma a materia… Secundum autem gradus est, quod separatur forma a materia et a praesentia materiae… Tertius gradus apprehensionis est, quo accipimus non tantum sensibilia, sed etiam quasdam intentiones quae non imprimuntur sensibus, sed tamen sine sensibilibus numquam nobis innotescunt… Quartus autem et ultimus gradus est, qui apprehendit rerum quiditates denudatas ab omnibus appendiciis materiae nec accipit ipsas cum sensibilium intentionibus, sed potius simplices et separatas ab eis. Et ista apprehensio solius est intellectus.’ In: De anima, ed. C. Stroick, omi (Münster: Aschendorff, 1968) (Opera Omnia, Editio Coloniensis 7.1). For Albert the Great’s theory of abstraction, see A. de Libera, Métaphysique et noétique. Albert le Grand (Paris: Vrin, 2005), Ch. V. Albert the Great, In Isag. (ed. Santos Noya, 34:50–53): ‘Dicendum pro certo quod universale unum numero essentiae est in anima et in seipso et in singulari, nec differt nisi secundum esse determinans ipsum ad hoc vel illud.’

52

chapter 2

There is a formal identity between the essence of an external thing and the form of its concept, which allows the application of the name to all the external things where that essence is individuated. Hence, not only names are imposed on concepts because they somehow share a sort of existence – mental existence – but also because in order to be applicable to many things names must signify something universal and only concepts are as such. This is how the semiotic problem of the articulation of utterance and concept in order to produce a linguistic sign – a word – gets entangled with the problem of the applicability of common names to many things. Albert the Great provides us, indeed, with a somewhat coherent explanation of the imposition of common names on concepts and their applicability to many things. Yet, he also presents us with at least a double use of the notion of signification: a use derived from the notions of imposition and universal applicability and a use derived from the idea of linguistic communication about the world. As we shall see this double use occurs in most of the authors who deal with the questions about signification during the second half of the century. The Modist Rejection of the Immediate Signification of Concepts Most masters of Arts from the second half of the thirteenth century reject the position that utterances can only be imposed on concepts, without necessarily rejecting the idea that the universality of common names is somehow grounded in a certain universality of their significate. Peter of Auvergne, for instance, in his question whether utterances signify affections of the soul or the things themselves,36 puts forth an argument that recalls Albert’s claim about the universality of concepts and the singularity of individuated essences: Furthermore, the same [position] about the common term is argued in this way: The common term signifies something common; in fact, it is only said to be common, because it signifies something common. But if it signified something real, it would not signify something common. There­ fore, it seems that it signifies something in the intellect.37 36

37

Peter of Auvergne, In Perih. (ed. Ebbesen, 152): ‘Et tunc quaeritur de primo, utrum ipsae voces significent passiones intellectus vel significent ipsas res.’ In: S. Ebbesen, ‘Animal est omnis homo: Questions and sophismata by Peter of Auvergne, Radulphus Brito, William of Bonkes and Others’, cimagl 63 (1993), 145–208, at pp.150–155. For Peter of Auvergne’s semiotics, see also Marmo, Semiotica e linguaggio, Ch. 3. Peter of Auvergne, In Perih. (ed. Ebbesen, 152): ‘Praeterea, hoc idem arguitur de termino communi sic: Terminus communis significat aliquid commune, immo ex alio non dicitur communis nisi quia significat aliquid commune; sed si significaret aliquid quod esset in

SIGNIFICATION OF CONCEPTS AND THINGS

53

Peter’s own argument in favour of the immediate signification of essences is based on the commonly used idea that ‘signification follows understanding’ (significare sequitur intelligere); an idea that, as we shall see in Part 2, is deeply engrained in thirteenth-century accounts of the grammatical categorisation of words.38 Since signification follows understanding, and since the intellect understands things by means of concepts, utterances also signify things by means of concepts; for the intellect understands the thing itself through its likeness, and therefore it also imposes the utterance on the thing itself through its likeness: About the second question, it must be understood…that just as it goes in understanding, it goes in signifying, because the act of signifying follows the act of understanding. Hence, it must be said that the utterances signify the same thing that the intellect understands; but the intellect understands by means of likenesses… Thus, when the intellect understands the thing itself, and this by means of its likeness, the intellect imposes the utterance in order to signify the thing itself, and this by means of the thing’s likeness… Hence, Aristotle said meaningfully that utterances are marks of the passions in the soul, because utterances signify nothing, unless by means of the passions or likenesses.39

38

39

re, non significaret aliquid commune; quare videtur quod significet aliquid quod est apud intellectum.’ This idea is also widely used in theology. It is indeed used by Alexander of Hales, Bonaventure and Thomas Aquinas and is rejected by Henry of Ghent and Duns Scotus. According to Rosier-Catach, the common source of these authors would be Pseudo-Denys’ claim that the different ways of understanding a thing result from different intellectual faculties, rather than from different features of the thing understood. For instance, in the case of God, who is essentially simple, our understanding of him is necessarily compound, because our cognitive faculties prevent us from understanding him as simple. Consequently, our way of naming him cannot be simple either. Cf. Ps.-Denys, De divinis nominis I.I. For the theological use of this principle and its rejection by Henry of Ghent and Duns Scotus, see J.E. Ashworth, ‘“Can I speak more clearly than I understand?” A Problem of Religious Language in Henry of Ghent, Duns Scotus and Ockham’, Historiographia linguistica 7 (1980), 29–38; Rosier-Catach, ‘Res significata’; and id., ‘Henri de Gand, le De dialectica d’Augustin et l’institution des noms divins’, Documenti e studi sulla tradizione filosofica medievale 6 (1995), 145–253. The principle in question, however, is rarely (if ever) referred to Ps.-Denys in grammatical or logical treatises from the faculties of Arts. In my opinion, one could also explain the wide use of this idea by masters of Arts from the second half of the 13th century as a logical development of earlier medieval interpretations of Priscian’s Institutiones that go in this direction. For these interpretations, see below, Part 2, Section 2. Peter of Auvergne, In Perih. (ed. Ebbesen, 154): ‘De secundo quaesito intelligendum… quod sicut est in intelligendo, ita est in significando, quia significare sequitur intelligere.

54

chapter 2

Thus, the likeness – the concept – is not what is immediately signified, but rather a cognitive requirement for the immediate imposition of the name on the thing that is the concept’s content. This is to say that the concept is an epistemological, and not a semiotic, condition for the signification of an essence. Peter closes his discussion with an objection to the argument that appeals to the singularity of individuated essences. A common name like ‘homo’ does not need to signify something actually common, but only something potentially common – the essence itself. An individuated essence, although it is actually singular, is also potentially universal, thus it can be understood and signified by a common term as universal: When it is said that the common term signifies something common, it must be said that the common term, e.g. ‘man’ and ‘animal’, does not signify something actually common, but only potentially [common]. In fact, it signifies only the quidity of man or of animal, and this quidity is not actually universal or particular, but it is universal or particular only potentially. Hence, it must be said that nothing prevents the common term, even when it signifies something that exists in reality, from signifying something that is potentially common.40 Hence, Albert’s position that external things are singular is not challenged. What is challenged is his position that there cannot be cognitive access to external essences, because external essences are potentially universal and this potentiality grounds their cognition as universal. Peter’s position is found in a more developed way in an anonymous author – the Incertus sf – who in his commentary on the Sophistical Refutations provides

40

Et ideo dicendum quod ipsae voces significant idem quod intellectus intelligit; sed intellectus intelligit res per similitudines, sicut visum est… Ita quod cum intellectus intelligit ipsam rem, et hoc per similitudinem ipsius rei, intellectus imponit ipsam vocem ad significandum ipsam rem, et hoc per similitudinem ipsius rei… Et ideo significanter dicit Aristoteles quod voces sunt notae earum passionum quae sunt in anima, quia voces nihil significant nisi per passiones sive per similitudines.’ Peter of Auvergne, In Perih. (ed. Ebbesen, 155): ‘Ad aliud quod dicitur quod terminus communis aliquid commune significat, dicendum quod terminus communis, ut “homo” vel “animal”, non significat aliquid commune in actu sed tantum in potentia; significat enim tantum quiditatem hominis vel animalis, et ista quiditas non est universalis nec particularis in actu, sed tantum in potentia se habet ad hoc quod sit universalis vel particularis. Et ideo dicendum quod nihil prohibet quin terminus communis, et quamvis significet aliquid quod sit in re, significet aliquid commune in potentia.’

SIGNIFICATION OF CONCEPTS AND THINGS

55

us with a thorough description of the mechanisms of the immediate signification of real essences. As most masters from his period, the Incertus sf presents us with the Porphyrian division of names into names of first and second imposition. Names of second imposition signify second intentions – logical notions that qualify our knowledge of external things. For instance, ‘universal’ is a name of second imposition that signifies the universal understanding of things, and ‘genus’ and ‘species’ signify their generic and specific understanding, and so on.41 In contrast, names of first imposition signify real essences in the external world.42 The proof that words of first imposition signify real essences, the author says, is that the significate of a word is the content of the concept that is formed in the mind of the listener when she understands the word. Now, when a speaker utters ‘a man runs’, the listener does not understand that the concept of man runs, but that a real man runs, because ‘man’ provokes the formation of the concept of a real man and hence it signifies a real man: The proof comes from Aristotle in Perihermeneias I [i.e. Int. 3]: ‘To signify is to form a concept’…Hence, the utterance signifies that whose concept is formed, insofar as the listener grasps [it] by means of the [utterance]. However, the listener grasps a real external thing by means of a word of first imposition; for when someone says ‘a man runs’, if by means of ‘man’ were formed the concept of a likeness or of a passion that only exists in the soul, and not the concept of a real external thing, it would follow that such intellectual compositions would be impossible, and similarly the propositions and the compositions that signify them; for it is impossible

41 Incertus sf, In se (ed. Ebbesen, 279:28–33): ‘Ad hoc (i.e. utrum possibile sit vocem rem veram significare) dicendum quod quaedam sunt nomina quae dicuntur esse nomina secundae impositionis quae significant res secundarum intentionum, cuiusmodi sunt illa quae significant non rem veram, sed intentiones quasdam quas fundat anima in rebus ipsis secundum quod intellectae, ut “genus” “species” et “universale” et sic de aliis.’ In: Auctores incerti, Quaestiones super Sophisticos Elenchos, ed. S. Ebbesen (Copenhagen: gad, 1977) (CPhD vii). Jan Pinborg attributed Incertus sf’s commentary to Boethius of Dacia, but Sten Ebbesen rejected this attribution, as well as a possible attribution to Peter of Auvergne, in his article ‘Questions and Sophismata: Tracking Peter of Auvergne’, in H.A.G. Braakhuis and al. (eds.), Aristotle’s Peri hermeneias in the Latin Middle Ages: Essays on the Commentary Tradition (Groningen: Ingenium, 2003), pp. 31–49. 42 Incertus sf, In se (ed. Ebbesen, 279:33–35): ‘Alia autem sunt nomina primae impositionis significantia res primarum intentionum, ut “homo” “asinus” “lapis” et similia et illa rem veram extra significant.’

56

chapter 2

that a real predicate, such as ‘to run’ inheres in the species or likeness in the intellect.43 Notice, however, that here the author moves from the problems of imposition and of the universal applicability of a name to the problem of the signification of a word in an assertoric context. If the name ‘man’ signified the concept as such, and not the content of the concept insofar as it exists in an external thing, it would be impossible for the listener to understand the assertion ‘a man runs’, let alone to grant it as true or reject it as false, because the composition of a concept and a real predicate like running is impossible. Therefore, at least in the case of singular accidental predication, it would be evident that the terms of an assertion stand for – signify – external things. The anonymous author underscores in this way that some linguistic utterances, namely some assertions composed of names of first imposition, intend to transmit information about real facts – i.e. they are fundamentally constative. Nonetheless, names do not signify real essences insofar as they exist, but insofar as they are understood – a name signifies an essence insofar as the latter is an object of imposition, and as such it has been intellectually separated from the existence that it has in individuals. The essence has indeed been separated from any sort of existence: It must be understood that names [of first imposition], although they signify a real external thing, however they do not signify it…insofar as it has…actual existence; for then, it would follow that with a change in the thing with respect to actual existence, the signification of the word would change… But a name signifies a real external thing as it is understood by the intellect; for the utterance becomes significative only because of the relation of signification that is given to it by the one who imposes it. But the one who gives a significate to the utterance cannot attribute it, unless she first conceives that significate in the mind, since she is a cognitive 43 Incertus sf, In se (ed. Ebbesen, 279:35–48): ‘Cuius probatio nam per Aristotelem primo Perihermeneias significare est intellectum constituere… Igitur vox illud significat cuius intellectum constituit secundum quod audiens per eam comprehendit. Per illas autem voces primae impositionis audiens rem veram extra apprehendit, nam dicendo “homo currit”…si per “hominem”…constituitur intellectus non rei verae sed alicuius similitudinis vel passionis existentis in anima tantum, sequeretur quod omnes tales compositiones intellectus essent impossibiles, et similiter compositiones et propositiones illas significantes, nam impossibile est aliquod praedicatum reale, cuiusmodo est “currere”…inesse speciei vel similitudini existenti in intellectu.’

SIGNIFICATION OF CONCEPTS AND THINGS

57

agent, and hence, just as the one who imposes understands a thing, in the same way she imposes [the utterance] on the thing in order to signify it.44 Along the lines of Avicenna’s (980–1037 ad) theory of the indifference of the essence, masters from the last quarter of the thirteenth century maintain that essences are understood and signified in themselves, and in themselves they are indifferent both to actual existence and to mental existence, even though actual existence or mental existence (or both) are always their concomitant accidents. Thus, existence outside the soul and existence in the soul are not essential features of essences, but essences always exist either outside the soul or in the soul (or both). Nevertheless, the signification of external essences that follows from the imposition of the utterance does not involve the actual existence of the essence; for this would entail the instability of the word’s signification, given that external things are constantly changing, not with respect to what they are essentially, but with respect to the events in which they are accidentally involved. At the turn of the fourteenth century, Radulphus Brito presents us with the most philosophically sophisticated version of this position. Just as Peter of Auvergne and the Incertus sf, Brito rejects the immediate signification of concepts and defends the immediate signification of essences, and does so from a semiotic perspective. However, he prefers to explain the reference to actual things in the external world in assertoric contexts (e.g. the reference to an individual of man in ‘a man runs’) in terms of second intentions, rather than in terms of signification. Brito’s argument for the immediate signification of essences is basically the same that we found in the Incertus sf: (a) To signify x is to provoke the understanding of x in the mind of the listener. (b) But when someone hears the utterance ‘x’, the essence of x is understood, and not its concept, because what is understood (i.e. the essence of the 44 Incertus sf, In se (ed. Ebbesen, 279:50–280:62): ‘Intelligendum tamen est quod nomina illa, etsi rem veram extra significent, ipsam tamen non significant…secundum quod… existentiam actualem habet, sic enim sequeretur quod re ipsa transmutata quantum ad esse in effectu transmutatur vocis significatio… Sed significant nomina rem veram extra ut ipsa ab intellectu est apprehensa, non enim redditur vox significativa nisi per rationem significandi sibi concessam ab imponente, imponens autem significatum voci non potest tribuere nisi prius illud significatum in mente conceperit, cum sit agens per cognitionem, et ideo sicut imponens rem aliquam intelligit, sic ei imponit vocem ad significandum.’

58

chapter 2

thing itself) is different from that through which it is understood (i.e. the concept). (c) Since the essence (through a concept) is primarily understood, it is also primarily signified by ‘x’.45 Brito defends the idea that the essence is the first object of the intellect in his commentary on the De anima in a passage that reveals his position as regards a debate about the first object of understanding. This thirteenth-century debate46 is placed in the crossroads of a metaphysical debate and an epistemological debate, debates that overlap and make the round trip between the Parisian faculties of Arts and Theology. The metaphysical debate has its theological origin in the sacrament of Eucharist and in the related discussion about the separation of accidents from substances. The epistemological debate, on the other hand, revolves around the possibility of knowing essences without their accidents. Contrary to Franciscan authors (who claim that it is impossible to have immediate cognitive access to essences, which can only be known in a mediate way through their accidents), Brito defends the possibility of an immediate cognitive access to essences, even if this access requires the knowledge of their proper accidents.47 45

46

47

Radulphus Brito, In Perih. q. 3 (ed. Pinborg, 276): ‘Modo in nominibus primae impositionis voces significant res et non conceptus rerum… Maior patet, quia sicut dicit Phylosophus… significare est intellectum constituere… Item probatur: essentia rei est quod intelligitur, quia illud quod est primum obiectum intellectus est illud quod intelligitur per vocem. Modo conceptus non est primum obiectum intellectus, immo rei essentia et ipsum quod quid est est primum obiectum intellectus, sicut apparet ii. De anima. Ergo essentia rei est quod intelligitur, et per consequens est illud quod significatur per vocem primae impositionis.’ Question edited in: J. Pinborg, ‘Bezeichnung in der Logik des xiii. Jahrhunderts’, pp. 275–281. For a thorough analysis of Brito’s account of the signification of common names, see A.M. Mora-Márquez, ‘Radulphus Brito on Common Names, Concepts and Things’. Aurélien Robert provides a thorough study of this debate in his unpublished dissertation ‘Penser la substance: étude d’une question médiévale’ (University of Nantes, 2005). I owe him thanks for giving me access to it. Radulphus Brito, In De an. (quoted by Robert, p. 197): ‘Aliter dicitur a quibusdam, et istam viam alias tenui, supponendo primo quod quando virtutes sunt connexae et ordinatae ad invicem, una illarum, scilicet inferiori, existente in sua operatione ex hoc statim superior potest aliquid cognoscere quod non est cognitum a virtute inferiori… [M]odo virtus intellectiva superior est ad virtutem fantasticam et ideo quando virtus fantastica est in fantasiando aliquod proprium accidens fantasiat, virtute intellectus agentis abstrahentis aliquam rationem intelligendi a fantasmatibus, tunc intellectus possibilis intelligit ipsum quod quid est sine hoc quod intelligat aliquod accidens prius et hoc est quod dicit

SIGNIFICATION OF CONCEPTS AND THINGS

59

Knowledge of the essence can be analysed into a neutral cognitive access to it (which allows the imposition of a word of first imposition) and the cognitive modality of this neutral access. These modalities are cognitive approaches – modes of understanding – that are ultimately grounded in potential properties of the essence itself, i.e. its possible modes of being. Accordingly, words of first imposition have further logical properties – e.g. predicability, universality, being a species and being a genus – because of these cognitive approaches to the essence itself. Thus, in his analysis of the term ‘intentio’ in his commentary on the Isagoge, Brito tells us that the term ‘intentio’ stands for four related, albeit notionally different, things: (a) a real external essence insofar as it is understood, which he calls ‘a concrete first intention’; (b) the concept of this essence as separated from any cognitive modality, which he calls ‘an abstract first intention’; (c) the essence itself insofar as it is understood under a certain cognitive modality, which he calls ‘a concrete second intention’; and (d) this cognitive modality itself, which he calls ‘an abstract second intention’: For instance, the intellect must first understand the man, absolutely, as a reasoning being, or the animal as a perceiving being, and this first and essential thought of the thing is called a first abstract intention, and the thing thus understood is called a first concrete intention. Then, once the intellect has understood the thing absolutely, it can afterwards understand it with a secondary thought, as [the thing] falls under some mode of being or under some regard, e.g. being predicable of many things. For Philosophus in littera sic: “cum enim habeamus tradere secundum fantasiam de accidentibus aut omnibus aut pluribus, tunc de substantia habebimus dicere optime”. Unde non vult quod fantasia cognoscat ipsum quod quid est, nec quod intellectus cognoscat ipsum accidens, sed quando fantasia fantasiatur ipsa fantasmata, tunc intellectus fertur in ipsum quod quid est virtute intellectus agentis abstrahentis aliquam rationem intelligendi ab ipsis fantasmatibus.’ This is also related to the notion of modus essendi or apparens, whose earliest proponent is perhaps Siger of Brabant. Brito adopts this notion, according to which the intellect grasps the essences of things, not in themselves, but through the sensible properties and operations by means of which an essence presents itself to the intellect. These properties and operations are also called modi essendi or apparentia. This notion also plays an important role in Brito’s explanation of the formation of second-order concepts, as well as in his argument in favor of the unity of the notion of being. For this latter use, see S. Donati, ‘Apparentia and Modi essendi in Radulphus Brito’s Doctrine of the Concepts: The Concept of Being’, in Fink et al. (eds.), Logic and Language, pp. 337–355; for the former use, see S. Ebbesen, ‘Radulphus Brito on the Metaphysics’, in J. Aertsen, K. Emery and A. Speer (eds.), Nach der Verurteilung von 1277 (Berlin: De Gruyter, 2001) (Miscellanea Mediaevalia 28), pp. 450–492 (repr. in Ebbesen, Collected Essays, vol. 2, pp. 197–208).

60

chapter 2

instance, when the intellect understands the man, which it had formerly understood in itself, as it is sayable of several things numerically different, the thing thus understood is a concrete species and such intention of the thing is the abstract species; and when it understands the man as it is sayable of several things specifically different as regards the quidity, it grounds there the intention ‘genus’. And the other intentions must also be understood in this way.48 Take for instance Socrates. (a) When the intellect understands Socrates as a rational being – as a human being – then Socrates’ humanity is a concrete first intention, and (b) the understanding of Socrates’ humanity is an abstract first intention. (c) When the intellect understands Socrates’ humanity as something that can be found in several numerically different individuals, Socrates’ humanity is a concrete second intention. Finally, (d) the understanding of humanity as something that can be found in several numerically different individuals is an abstract second intention – a species. It is thanks to (a) and (b) that we have a neutral cognitive access to the essence of human beings, which allows the imposition of a name of first imposition – e.g. ‘man’ – on it. However, it is thanks to (c) and (d) that this name is, for instance, a specific and universal term. For it is thanks to (c) and (d) that the essence falls under a certain mode of understanding that determines the mode of signifying whereby ‘man’ is a specific universal term – a term that is applicable to and predicable of many things. It is noteworthy that the different cognitive modalities are grounded in real properties of things – their modes of being – and thus the logical properties of a name of first imposition are ultimately grounded in the ontological structure of reality:

48

Radulphus Brito, In Isag. (ed. Pinborg, 105): ‘Verbi gratia sic oportet quod intellectus primo intelligat hominem absolute ut ratiocinantem vel quod intelligat animal ut sentiens, et iste intellectus primus et essentialis de re sive primus conceptus de re dicitur prima intentio in abstracto, et res sic intellecta dicitur prima intentio in concreto. Et tunc cum intellectus intellexit rem absolute potest postea ipsam intelligere secundario intellectu ut est sub aliquo modo essendi sive sub aliquo respectu ut scilicet est praedicabilis de pluribus. Verbi gratia, sicut intellectus intelligit hominem, quem prius secundum se intellexit, ut est dicibilis de pluribus differentibus numero, et res sic intellecta est species in concreto, et talis intellectio rei est species in abstracto; et intelligit animal ut est dicibile de pluribus differentibus specie in eo quod quid, et sic fundat ibi intentionem generis. Et sic intelligendum est de aliis intentionibus.’ In J. Pinborg, ‘Radulphus Brito on Universals’, cimagl 35 (1980), 56–142.

SIGNIFICATION OF CONCEPTS AND THINGS

61

The things themselves understood in this way are the concrete universal. All these notions of understanding are taken from certain modes of being in things, just as this notion of understanding the thing as it is in several numerically or specifically different individuals is taken from the mode of being ‘to be in several individuals different in form or in quantity’. Then, I argue: Just as the mode of being from which the universal intention in general is taken is related to the modes of being from which these five universal intentions are taken, the second intentions taken from those modes of being are related in the same way.49 Hence, Brito’s logic of second intentions – which accounts for logical properties of thought and language, such as universality, universal predicability, predicability, capability of being a subject, inferential validity etc. – is entirely grounded in ontological properties of the external world. More importantly, the signification of a word – the one that follows from its imposition – does not seem to play an immediate role in the explanation of those logical properties in Brito’s logic. He seems then to take a step further with respect to his predecessors when he narrows the use of the notion of signification to the logically neutral first institution of linguistic signs, the logical properties thereof in turn being systematically explained in terms of second intentions. Let us now see how these same issues were engaged in fundamentally different ways by the Franciscan authors Roger Bacon and Peter John Olivi. Roger Bacon and Peter John Olivi’s Rejection of the Immediate Signification of Concepts The positions taken by Roger Bacon and Peter John Olivi differ from the modists’ positions, largely because of the friars’ strong commitment to Augustinian semiotics. In De doctrina Christiana, Augustine (354–430 ad) introduces a division of signs into natural signs (signa naturalia) and given signs (signa data), depending on whether they are intentionally produced or not. Natural signs are those that provoke the knowledge of something else without someone’s intention of 49

Radulphus Brito, In Isag. (ed. Pinborg, 70): ‘Universale autem in concreto sunt ipsae res sic intellectae. Modo omnes istae rationes intelligendi sumuntur ab aliquibus modis essendi in re, sicut ista ratio quae est ratio intelligendi rem ut est in pluribus differentibus numero vel specie sumitur ab isto modo essendi qui est esse in pluribus differentibus formaliter vel per quantitatem. Tunc arguo: sicut se habet modus essendi a quo sumitur intentio universalis in communi et modi essendi a quibus sumuntur istae intentiones 5 universalium ad invicem, sic se habent istae intentiones secundae sumptae ex illis modis essendi.’

62

chapter 2

doing so. An example of this sort of sign is the smoke that indicates fire.50 Given signs, in turn, are intentionally produced by animated beings in order to communicate to each other their thoughts and sensations,51 human words being the given signs par excellence.52 The distinctive feature of Augustine’s definition of signs with respect to the Aristotelian tradition lies in Augustine’s strong focus on their reception in someone’s intellect rather than in the purpose of their production – a sign is not a sign unless it is interpreted as such through an understanding of its content. Thus, according to Augustine: A sign is a thing that, besides the form (species) that it presents to the sense, evokes something different in the mind.53 In a passage from De magistro, Augustine describes words as the union of utterance and signification. An utterance is that whereby the word can present something to the senses of an interpreter. Signification, in turn, takes place when the interpreter evokes an understanding of the thing that the utterance indicates, thereby providing the word with a value: Therefore, the understanding of words is achieved when the things are understood, and not when the words are heard; for we do not learn [the signification of the] words that we [already] know, nor can we claim to have learned those that we don’t know, unless we have understood their signification, which doesn’t happen with the hearing of uttered words, but with the understanding of the things signified.54 50 Augustine, De Doct. Ch. ii.I.1: ‘Naturalia sunt, quae sine voluntate atque ullo appetitu significandi praeter se aliquid aliud ex se cognosci faciunt, sicut est fumus significans ignem.’ In: Augustine, De doctrina christiana, in La doctrine chrétienne (Paris: Institut d’Etudes Augustiniennes, 1997) (Oeuvres de Saint Augustin 11/2). For Augustine’s semiotics, see E. Bermon, La signification et l’enseignement. Texte latin, traduction française et commentaire du De magistro de saint Augustin (Paris: Vrin, 2007); and R.A. Markus, ‘St. Augustine on Signs’, Phronesis 2 (1957), 60–83. 51 Augustine, De Doct. Ch. ii.ii.3: ‘Data vero signa sunt, quae sibi quaeque viventia invicem dant ad demonstrandos, quantum possunt, motus animi vel sensa aut intellecta quaelibet.’ 52 Augustine, De Doct. Ch. ii.iii.4: ‘Verba enim prorsus inter homines obtinuerunt principatum significandi quaecumque animo concipiuntur, si ea quisque prodere velit.’ 53 Augustine, De Doct. Ch. ii.i.1: ‘Signum est enim res praeter speciem, quam ingerit sensibus, aliud aliquid ex se faciens in cognitationem venire…’. 54 Augustine, De mag. xi.36: ‘Rebus ergo cognitis verborum quoque cognitio perficitur; verbis vero auditis nec verba discuntur; non enim ea verba, quae novimus discimus, aut quae

SIGNIFICATION OF CONCEPTS AND THINGS

63

The understanding of the value of a word by an interpreter is thus the essential feature of Augustine’s account of linguistic signs.55 The two crucial features of Augustinian linguistic signs, namely that they are produced intentionally with the purpose of communication and that they need to be interpreted in order to be signs, are widely used in the linguistic treatises by the Franciscan authors Roger Bacon and Peter John Olivi. Roger Bacon presents us with an account of signification largely associated with the problem of linguistic communication. Bacon claims that the relation to an interpreter is an essential feature of the sign, so that without this relation a material item is a sign only potentially: The sign is in the category of relation, and it is called [a sign] essentially in relation to the one for whom it signifies, because [the interpreter] is in action, when the sign itself is in action, and in potency when [the sign] is in potency, because if someone could not conceive [something] through the sign, [the sign] would be empty and vain – indeed it would not be a sign.56 Accordingly, for Bacon a sign is a perceptible or intelligible item whose signification depends on a double relation: an accusative relation to its significate and a dative relation to its interpreter. However, the accusative relation, which in the case of linguistic signs is established by imposition, is only potential until its actualisation by an interpreter. The relation between word and significate is accidental and subordinated to the more essential relation between word and interpreter: non novimus didicisse non possumus confiteri, nisi eorum significatione percepta, quae non auditione vocum emissarum, sed rerum significatarum cognitione contingit.’ In: E. Bermon, La signification et l’enseignement. 55 In De magistro we are also told that the knowledge of the value of a sign can only be acquired by means of other signs, in a process where the master is only leading his pupil to the recollection of a knowledge that he already has in his mind. Thus, properly speaking the master is not teaching anything but only leading his pupil to the recollection of something that he knew already. See Bermon, La signification et l’enseignement, 2nd part. 56 Roger Bacon, De signis (ed. Fredborg et al., 81): ‘Signum est in praedicamento relationis et dicitur essentialiter ad illud cui significat, quoniam illud ponit in actu cum ipsum signum sit in actu, et in potentia cum ipsum est in potentia. Quia nisi posset aliquis concipere per signum, cassum esset et vanum, immo non erit signum.’ For Bacon’s semiotics, see I. Rosier-Catach, La parole comme acte. Sur la grammaire et la sémantique au xiiie siècle (Paris: Vrin, 1994), Chs. 3–4; Marmo, Semiotica e linguaggio, Ch. 1; Joël Biard, Logique et théorie du signe (Paris: Vrin, 1989), Ch. 1; and T.S. Maloney, ‘The Semiotics of Roger Bacon’, Medieval Studies 45 (1983), 120–154.

64

chapter 2

This verb ‘I signify’ regards principally and most essentially the one by whom something is acquired – the thing signified by the dative rather than [the one signified] by the accusative.57 Finally, utterances are intentionally imposed on whatever can be understood, whether it exists or not, whether it is a concept, a fictional entity, or an external thing, and whether it is universal or singular: In fact, if we can understand, we can at once impose a name by will. And further, names are upon our own decree (ad placitum), and hence we can impose names by will and give them to beings or non-beings.58 With this semiotic background, Bacon replies to the question whether names immediately signify concepts or external things. However, he rephrases the question as follows: Supposing that the name was imposed on an external thing, does the name signify the thing or the concept whereby the imposition was made? Bacon’s solution puts a strong focus on the intentional imposition of the utterance by the speaker. His argument goes as follows: (a) Let us suppose that the name was intentionally imposed on an external thing. (b) Let us also suppose that because of this imposition it is a sign by will. (c) Could then the name be a sign by will (ad placitum) of the thing’s concept? (d) No, because in that case it would not be a sign by will of the thing on which it was intentionally imposed. (e) And since the name was not intentionally imposed on the concept, it is not a sign by will of the concept either.59 57

58

59

Roger Bacon, De signis (ed. Freborg et al., 82): ‘Hoc verbum “significo” essentialius et principaliter rescipit illud cui adquiritur aliquid, hoc est rem per dativum significatam, quam per accusativum. Et propter hoc ad rem quam habet significare non refertur nisi per accidens, hoc est sicut scibile ad scientiam.’ Roger Bacon, De signis (ed. Fredborg et al., 87–88): ‘Si enim possumus intelligere, possumus nomen imponere secundum quod ipse vult ibidem. Et iterum nomina sunt ad placitum nostrum, et ideo pro voluntate possumus nomina imponere et dare sive enti sive non enti.’ Roger Bacon, De signis (ed. Fredborg et al., 132–133): ‘Et certum est inquirenti quod facta impositione soli rei extra animam, impossibile est quod vox significet speciem rei tamquam rei signum datum ab anima et significativum ad placitum, quia vox significativa ad placitum non significat nisi per impositionem et institutionem. Sed concessum est vocem soli rei imponi et non speciei.’

SIGNIFICATION OF CONCEPTS AND THINGS

65

Thus, the potential signification of a word – its accusative relation to a significate – entirely depends on the intention of communication of the one who imposes the name. That is, it regards the thing about which the speaker intends to transmit information to someone else. Bacon also reads Priscian’s claim that parts of speech signify concepts taking ‘conceptum’, not as the accusative of the noun ‘conceptus, −us’, but as the neuter accusative of the passive participle ‘conceptus, −a, −um’ (i.e. that which is conceived): If one objects with Priscian that being a part of speech is nothing different from signifying a concept of the mind…it must be said that ‘concept of the mind’ in this authority is not taken as the accusative case of the noun ‘concept’, but as the substantivated adjective ‘conceived’, and thus ‘concept’ is the same as the thing conceived and understood in this way.60 This means that once a name has been intentionally given to an external thing, it potentially signifies that external thing and not the concept whereby the imposition was made. On the other hand, contrary to the modists for whom the universality of a common term depends partly on a cognitive modality and partly on an ontological modality, for Bacon it depends entirely on the ontological structure of the thing that is conceived and signified. A name is universal – it can be applied to many things – because it was imposed on a universal nature that inheres in several singular things of the external world: Further, the species of the universal thing is truly universal just as the species of the singular thing is singular; for, just as the universal man is predicated of singular men, in the same way the species of the universal man is predicated of the species of singular men, and it inheres in them, just as the universal thing inheres in the singular things.61 60

61

Roger Bacon, De signis (ed. Fredborg et al., 134–135): ‘Si obiciatur per Priscianum quod nihil aliud est esse partem orationis quam mentis conceptum significare…dicendum quod hoc quod dico conceptum mentis in illa auctoritate non sumitur secundum quod est accusativi casus huius nominis substantivi “conceptus”, sed huiusmodi adiectivi substantivati “conceptus”, et sic conceptus idem est, quod res concepta et intellecta quantum huiusmodi.’ Roger Bacon, Communia naturalium c. 10 (ed. Steele, 103:22–27): ‘Ceterum secundum veritatem species rei universalis est universalis sicut species rei singularis est singularis. Nam sicut homo universalis praedicatur de singularibus hominibus, sic species hominis universalis praedicatur de speciebus hominum singularium, et est in eis, sicut res universalis

66

chapter 2

Thus, the logical properties of language are not just loosely grounded in ontological properties of external reality, but univocally determined by the ontological structure of the thing signified by a name. How is it then possible to use the name ‘homo’ to talk about both a common nature as in e.g. ‘homo est species’ and a singular man as in e.g. ‘a man is running’? Because, Bacon tells us, at every use of the name ‘homo’, there is a renewed act of imposition – a re-imposition – that re-determines the signification intended by a particular speaker at a particular time.62 In the assertion ‘man is a species’ there is a re-imposition of ‘man’ on the common nature of man, and in ‘a man is running’ there is a re-imposition of ‘man’ on the singular form of some man who is actually running. Consequently, Bacon does not leave to the concept any role to play in the explanation of the logical features of language, other than to provide a cognitive access to the things we intend to talk about. It is no surprise, then, that the idea that signification follows understanding (significare sequitur intelligere) is entirely absent from his contributions to the debates discussed in this study. That is, not only does Bacon put forth a highly token-oriented logic, he also totally dismisses the logic of second intentions that we saw reaching its peak in Radulphus Brito. Peter John Olivi, in turn, also draws heavily on Augustine’s semiotics without going all the way along Bacon’s lines. In his Quaestiones logicales, in a discussion about equivocation, Olivi establishes a difference between the potential relation of signification that follows from the imposition of a name and the actual relation of signification that takes place when the name is used in a context of communication: It must be said that the signification of a name is taken in two ways: first, dispositionally, as it were; second, actually or in exercise, as it occurs when [the name] is actually uttered out of an intention to signify or when it is written in some narration. [The name] has [signification] in the first mode from its imposition; and in the second [mode] from its application to the things to which it is now applied and had already been imposed.63

62

63

est in rebus singularis.’ In: Communia naturalium, ed. R. Steele, in Opera hactenus inedita Rogeri Baconi (Oxford: Clarendon Press, 1909–13). The same holds for the use of ‘homo’ when we speak of an actual existing man and of a man who does not exist anymore. In the next section, we shall come back to this issue and to its relation with Bacon’s radical extensionalism. Peter John Olivi, ql, q.7 (ed. Brown, 353): ‘Dicendum quod significatio nominis dupliciter accipitur: primo scilicet quasi habitualiter; secundo in actu aut exercitio, sicut fit cum ex intentione significandi actualiter profertur aut cum in certa narratione est scriptum. Primum modum habet ex impositione; secundum vero ex applicatione ad quae iam est et

SIGNIFICATION OF CONCEPTS AND THINGS

67

When they are actually used in contexts of communication, names are intentionally given to – they signify – the thing that the speaker intends to talk about. For instance, in ‘the man runs’, the intention of the speaker is to talk about a real man and not about a thought, and hence when it is used in this assertion, ‘man’ signifies a real man and not its concept: [Utterances] signify priorly the thing that is a man, because the intention of men more directly and principally aims to signify the thing that is a man than something else…In fact, when we say ‘a man runs’, we do not intend to say that the concept of man, or that the man as it is in our intellect, runs, or that the abstract humanity runs, but rather that the thing that is a man concretively taken runs.64 Hence, ‘man’ not only potentially signifies the thing on which it has been conventionally imposed, but it also signifies – in a sense closer to ‘to refer to’ – the thing that a speaker wants to mark out when she actually uses the name in a context of linguistic communication. Nevertheless, Olivi departs from Bacon when he adheres to the idea that the thing is signified in a way that follows the mode of understanding that was intentionally chosen by the speaker when she uses the name: I reply that, even though the names of things were imposed in order to signify things, they were imposed on things only insofar as they were

64

erat impositum.’ For Olivi on signification, see A.M. Mora-Márquez, ‘Pragmatics in Peter John Olivi’s Account of Signification of Common Names’, Vivarium 49 (2011), 150–164; and I. Rosier-Catach, La parole efficace. Signe, rituel, sacré (Paris: Seuil, 2004), Ch. 2.7. For Olivi in general, see A. Boureau and S. Piron, Pierre de Jean Olivi (1248–1298). Pensée scolastique, dissidence spirituelle et société (Paris: Vrin, 1999). Peter John Olivi, ql, q.4 (ed. Brown, 346): ‘… prius significant res illas quas significandas sunt institutae. Et hoc modo hoc nomen “homo” prius significat rem quae est homo, quia intentio hominum directius et principalius intendit per hoc nomen “homo” significare rem quae est homo quam aliquid aliud…Cum enim dicitur “homo currit” non intendimus dicere quod conceptus quam habeo de homine vel homo prout est in intellecto nostro currit aut quod humanitas abstracta currit, sed potius quod res quae est homo concretive acceptus currit.’ Cf. Walter Burley, Comm. m. in Perih. 1.15 (ed. Brown, 55): ‘Similiter si vox significaret passionem seu idolum rei haec esset vera: “Sortes est idolum Sortis vel passio Sortis”, quia isti termini “Sortes” et “idolum Sortis” eandem rem significant primo, quia uterque significat primo idolum Sortis… Propter quod credo quod non oportet quod vox primo significet passionem animae. Sed in libera potestate imponentis est quid vox debeat primo significare, an rem an passionem.’

68

chapter 2

grasped by the intellect and under the mode under which the intellect wants to signify the things understood.65 While Bacon’s ontology allows him to anchor the universality of names directly on the forms of external things, Olivi, who rejects the external existence of universals, explains this logical feature in terms of modes of signifying and of understanding. Also, while for the modists the modes of signifying and of understanding are ultimately grounded in the modes of beings of things, Olivi explicitly and radically rejects the possibility that things or their properties affect in any way whatsoever the human soul. As a result, the logical features of both human knowledge and human language depend to a great extent on the intentional modalities of those who think and talk. In fact, Olivi rejects the need to posit intelligible forms (species intelligibiles), both as the starting point of cognitive processes (as in e.g. Thomas Aquinas) and as their result (as in e.g. Radulphus Brito). The reason behind Olivi’s rejection is his radical commitment to an anthropology based on the notion of free will, as opposed to an anthropology (e.g. Aquinas’) based on the human rational faculty. Olivi’s epistemological argument against intelligible forms (species intelligibiles) is that in either case (i.e. as the starting point or as the result of cognitive processes), they would posit an obstacle to a direct cognitive access to the external world.66 He proposes instead that the human intellect is directed by its will to focus its attention on certain features of things – on real 65

66

Peter John Olivi, ql, q.1 (ed Brown, 338): ‘Respondeo quod licet nomina rerum sint impo­ sita ad significandum res, non tamen sunt imposita rebus nisi prout sunt apprehensae ab intellectu et sub illo modo sub quo intellectus vult res se intellectas significari.’ Peter John Olivi, De verbo (ed. Pasnau, 144): ‘Nulla enim est necessitas aut utilitas ponere tale verbum (i.e. the concept or the intelligible species). Quia vel res et earum reales habitudines quas intellectus intelligit sunt praesentes intellectui in seipsis aut in speciebus memorialibus, et ideo sive res earumque habitudines sint intellectui in seipsis praesentes sive ipsae sint absentes per memoriales tamen species intellectui praesentatae, nulla est necessitas alterius obiectivi speculi in quo res ipsi intellectui praesentatur. Immo, potius esset ad impedimentum.’ For an exhaustive study of Olivi’s account of human cognition, see R. Pasnau, Theories of Cognition in the Later Middle Ages (Cambridge: cup, 1997); see also, id., ‘Petri Iohannis Olivi Tractatus de Verbo’, Franciscan Studies 53 (1993), 121–153; and J. Toivanen, Perception and the Internal Senses. Peter of John Olivi on the Cognitive Functions of the Soul (Leiden: Brill, 2013) (Investigating Medieval Philosophy 5). For Olivi’s rejection of the species intelligibilis, see L. Spruit, Species intelligibilis: From Perception to Knowledge. t. 1: Classical Roots and Medieval Discussions (Leiden: Brill, 1994). For Olivi’s anthropology, see R. Pasnau, ‘Olivi on Human Freedom’, in Boureau et al. (eds.), Pierre de Jean Olivi, pp. 15–26; and F.X. Putallaz, Insolenté liberté. Controverses et condamnations au xiiie siècle (Fribourg/Paris: Presses Universitaires /Le Cerf, 1995).

SIGNIFICATION OF CONCEPTS AND THINGS

69

notions (reales rationes) – so that concepts are nothing but the very act of focusing the intellect’s attention on those notions. Universal knowledge results from the intentional consideration of the real notion of common natures, without this implying that there is a bearer of universality other than the intentional act itself: The second [position] is evident, because the first abstraction of uni­ versal notions is made only as regards the act of consideration that separates the real notion of the common or specific nature, without regarding and considering the notion of individuation. However, through this no real object is separated or formed that is different from the act of consideration…67 Olivi rejects both the external existence and the mental existence of common natures. Universality is a cognitive approach whose object – the real notion – is not common. Hence, he puts forth a nominalism of sorts where the real notion of common natures can be understood as the intellectual consideration of a likeness between individual things, without the consideration of their individuating features – a likeness that is only an act of reason, which is not grounded in something real and commonly shared by individual things: In fact, things are not made similar with respect to a third thing, unless you take the third thing only according to the intellect or in the mode of the exemplar, as all the things that are thought through the same idea can be said to be similar with respect to it. What is said thereafter – that individuals are not similar as regards their proper nature – is true, if the sense is that they are not made similar with respect to it as a third thing commonly shared by them. In fact, the likeness does not require such third thing, but rather the opposite.68 67

68

Peter John Olivi, De verbo (ed. Pasnau, 145): ‘Ad secundam etiam patet quia prima abstractio rationum universalium fit in solo actu abstractivae considerationis realem rationem naturae communis vel specificae absque ratione suae individuationis attendentis et considerantis. Per hoc autem nihil obiectivum realiter abstrahitur vel formatur quod differat a praefactae considerationis actu…’. For Olivi’s account of intentionality, see D. Perler, Théories de l’intentionnalité au moyen âge (Paris: Vrin, 2003), Ch. 2. Peter John Olivi, In sent. ii (ed. Jansen, 248): ‘Non enim aliqua sibi assimilantur in aliquo tertio, nisi illud tertium accipias secundum intellectum tantum aut per modum exemplaris, eo modo quo omnia ideata ab eadem idea possunt dici sibi esse similia in illa. Quod autem dicitur post, quod individua non sunt sibi similia in natura propria: si sit sensus quod non assimilantur sibi in ea tanquam in alio tertio communiter ab eis participata,

70

chapter 2

Through the consideration of this likeness, the intellect can afterwards apply a common name to all the things that are in a relation with it. It can also establish further logical relations between things, such as predication and inference: To the fourth [argument], it must be said that, just as the intellect distinguishes between two notions that do not have real or essential distinction in things, and just as such two notions can be ordered as regards priority and posteriority through a relation to the intellect that they do not have in reality, in the same way similar predications, inferences and relations can be formed from such mode of understanding.69 To sum up, just like Bacon, Olivi presents us with a semiotic account along Augustinian lines, but the logical theory that it supports differs strongly both from Bacon’s realism and token-oriented logic and from the modists’ somewhat moderate realism and logic of intentions. Olivi breaks up, indeed, the epistemological constraint that lies behind Albert’s rejection of the imposition of common names on external things. However, he also rejects the idea that universality as a logical feature of language is immediately dictated by real properties of external essences, thus putting forth the only known thirteenthcentury anticipation of fourteenth-century nominalism. The Verificational Angle The final angle that is taken in arguments against the immediate signification of concepts (i.e. in Scotus’ commentary on the Perihermeneias) regards propositional verification. This approach to the question can also be traced back to the logical commentaries from the first half of the century, namely to the dubitationes about the different subject matters of the Categories and the Perihermeneias. Due to Boethius’ claim in his commentary on the Categories that words signify things,

69

vera est. Sed similitudo non exigit tale tertium, immo potius contrarium.’ In: Quaestiones in secundum librum Sententiarum, ed. B. Jansen (Quarachi: Collegium. S. Bonaventurae, 1922–26) (Biblioteca Franciscana Scholastica 4–6). Peter John Olivi, In sent. ii (ed. Jansen, 250): ‘Ad quartum dicendum quod sicut intellectus distinguit inter duas rationes quae secundum rem non habent inter se diversitatem realem seu essentialem et sicut duae tales rationes possunt habere inter se ordinem prioritatis et posterioritatis per respectum ad intellectum quem non habent in re: ita et possunt ex tali modo intelligendi consimiles praedicationes, consequentiae et habitudines formari.’ For Olivi on relations, see A. Boureau, ‘Le concept de relation chez Pierre de Jean Olivi’, in Boureau et al. (eds.), Pierre de Jean Olivi, pp. 41–56.

SIGNIFICATION OF CONCEPTS AND THINGS

71

authors such as Nicholas of Paris, Robert Kilwardby and Albert the Great ask why Boethius claims in his commentary on the Categories that words signify things, while in his commentary on the Perihermeneias his claim is that words signify concepts.70 These authors generally accept that Boethius is right in assessing a difference between the approaches to signification in the two treatises.71 The reason is that in the Categories Aristotle deals with simple sayables (dicibilia incomplexa) that are divided into the ten categories corresponding to the ten categories of being, while in the Perihermeneias he deals with simple sayables that can be the subject – that of which something is said – and the predicate – that which is said of something else – of an assertion. Regardless of the ontological category of its significate, a sayable is a subject or a predicate (a name or a verb) because of a mode of understanding this significate – insofar as the significate is understood as a subject or as a predicate: Likewise, the Categories deals with the term insofar as it is predicable, and thus is the thing; hence, there the analysis of the term is into utterance and thing. But this science [i.e. the Perihermeneias] deals with the term insofar as it is said of something else or as something else is said of it, and thus it is something according to the intellect; for the intellect takes whatever is said of something else and puts it under the notion /verb/; and [it takes] whatever about which other things are said and puts it under the notion /name/. This is why in this science the term is analysed into utterance and concept, rather than into utterance and thing. Hence, [Aristotle] says that utterances are marks of concepts, and so of things, but in the Categories [he says] that [utterances] are marks of things.72 70

71 72

See e.g. Nicholas of Paris, In Perih. (ed. Hansen and Mora-Márquez, 18:13–16): ‘Item. De diversitate dicibilis incomplexi in libro Praedicamentorum et hoc quaeritur. Dicit Boethius quod terminatur hic de vocibus prout significant intellectus, in libro Praedicamentorum prout significant res.’ Cf. Nicholas of Paris, In Perih. (ed. Hansen and Mora-Márquez, 20:28): ‘Posset etiam sustineri differentia quam assignat Boethius.’ Robert Kilwardby, Not. sup. Perih. (M46ra; P67va; V3r): ‘Item, in Praedicamentis est sermo de terminis ut terminus est praedicabilis, et sic est res; unde est ibi resolutio termini in vocem et rem. Set in scientia ista est sermo de termino secundum quod dicitur de altero sive alterum de ipso, et sic est aliquid secundum intellectum. Accipit enim intellectus quaecumque de altero dicuntur, et ponit sub intentione verbi; et de quo alia, et ponit sub intentione nominis. Et propter hoc in scientia ista resolvitur terminus in vocem et intellectum magis quam in vocem et rem. Unde dicit quod voces sunt notae intellectuum, et sic rerum; in Praedicamentis quod sunt notae rerum.’

72

chapter 2

However, to be understood as a subject or as a predicate is not a property of real things, but rather a way of understanding a significate when it is composed with or divided from something else in the mental acts of composition or division. For example, Nicholas of Paris tells us that: The Perihermeneias deals with utterances that signify concepts, i.e. modes of understanding; for, [an utterance] is not said to be a name because of the thing signified, but because of its mode of signifying.73 The name – the subject – and the verb – the predicate – of an assertion indicate two concepts that are considered as that of which something is said and that which is said of something else, respectively.74 Therefore, the name and the verb can form an assertion that preserves the logical form of a mental composition and division insofar as they signify as subject and as predicate the concepts that are mentally composed or divided. Moreover, in their analyses of Aristotle’s Perihermeneias, Nicholas and Kilwardby e.g. do not consider the possible reference to the external world by names and verbs, because their purpose is to explain what are names and verbs in the context of the treatise. They rightly notice that Aristotle’s names and attributes are the linguistic counterparts of the simple thoughts that are compared in a mental composition (or division) – a composition (or division) that is susceptible of truth-value precisely because of its compound form: In fact, in the Categories utterances signify things that exist, because a non-being is neither a category nor it falls under a category, but in this book [i.e. the Perihermeneias] [utterances] signify thoughts (i.e. intelligible things, largely taken, or imaginable things), whether they are beings

73 E.g. Nicholas of Paris, Rat. sup. Praed. L1 (ed. Hansen, forthcoming): ‘...in libro Perihermeneias de vocibus significantibus intellectum, id est modos intelligendi; non enim dicitur nomen propter rem significatam, sed propter modum significandi.’ 74 Cf. AnOx, In Perih. (ed. Mora-Márquez, 173:2–8): ‘Scientia autem Praedicamentorum est de vocibus inquantum sunt significativa intellectuum, qui intellectus sunt signa rerum materialium (sicut substantiae et qualitates). In libro autem isto determinat vocibus secundum quod sunt signa intellectuum, qui intellectus sunt signa rei secundum animam, sicut de nomine et verbo, quia huiusmodi voces (nomen et verbum) significant res quae sunt secundum animam…’. Cf. Albert the Great, In Isag. (ed. Santos Noya, 9:41–45): ‘Complexio et incomplexio non accidunt rei, secundum quod res est, nec etiam voci, secundum quod est vox, sed accidunt voci, secundum quod refertur ad intellectum simplicem vel compositum.’

SIGNIFICATION OF CONCEPTS AND THINGS

73

or non-beings, because an assertion can be formed indifferently by a name that signifies a being and a non-being.75 Although these early authors maintain, indeed, that logic in general deals with truth and falsity, as well as with syllogistic validity, they also separate Aristotle’s correspondence between assertions and mental compositions (or divisions) in the Perihermeneias from the problem of truth-determination (which they would rather discuss in e.g. their commentaries on the Categories). Nonetheless, at least from Martin of Dacia onwards there are explicit questions about propositional truth and falsity in the context of commentaries on the Perihermeneias (e.g. where truth and falsity are as in a subject; and whether a real composition corresponds to any true mental composition). These questions, in one way or another, amount to asking how it would be possible to make true statements about the external world (e.g. scientific statements), if words signified concepts. Martin of Dacia follows his predecessors in claiming that the truth or falsity of an assertion is immediately determined by the truth or falsity of the mental composition or division that it signifies. For truth and falsity are primarily a feature of mental items and only secondarily a feature of assertions through their representation of true or false mental compositions or divisions. Martin thereby pushes the problem of propositional verification to the mental level: It must be said that truth is the good of the intellect, just as Aristotle takes it in book vi of the Ethics. For truth is a disposition of the intellect according to which it is made similar to the thing. Therefore, since truth is the good of the intellect, nothing will be true unless it is related to the intellect. But something is related to the intellect in three ways: in one way, as the sign [is related] to what is signified, and in this way a true sentence is related to the intellect. For the sentence insofar as it is true is a sign of the truth in the intellect.76 75

76

Cf. Nicholas of Paris, In Perih. (ed. Hansen and Mora-Márquez, 21:1–5): ‘Nam in libro Praedicamentorum voces significant res exsistentes, quoniam non ens neque est genus neque est in genere, in hoc libro significant intellectus, id est intelligibile large, id est imaginabile, sive sit ens sive non ens, quia ex nomine significante ens et non ens indifferenter potest constitui enuntiatio.’ Cf. Albert the Great, In Isag. (ed. Santos Noya, 10:29– 33): ‘Adhuc dictio sive vox significativa et articulata non ex hoc quod refertur ad rem designatam, sed ex hoc quod est in intellectu quaerentis ignotum scire per notum, habet quod quaedam ipsius est nomen, quaedam autem verbum.’ Cf. above, Section 1.1. Martin of Dacia, Super Perih. (ed. Roos, 250:24–251:5): ‘Dicendum sicut vult Aristoteles vi. Ethicorum, quod verum est bonum intellectus. Nam veritas est dispositio intellectus,

74

chapter 2

Most authors from the second half of the century reject this position and claim that at least some assertions, namely the ones about external facts, are immediately verified by real compositions and divisions. Accordingly, with respect to the immediate signification of concepts they object that if words signified concepts, assertions such as ‘a man runs’ would always be false, because the concept of a man is never running.77 Similarly, assertions such as ‘Socrates is’ would always be true, because the concept of Socrates always exists,78 and assertions such as ‘man is an animal’ or ‘Socrates is a man’ would always be false, because the concept of man and the concept of animal are always different.79 Consequently, only assertions about second intentions or fictional entities would be susceptible of truth and falsity. The Incertus sf, for instance, gives the following argument in order to defend the immediate signification of external things: The word signifies the thing x whose concept is provoked in the listener’s mind when she hears the word. But the listener of a word of first imposition understands an external thing. For, if when someone uttered ‘a man runs’ the listener understood the

77

78

79

secundum quam assimilatur rei. Cum ergo verum sit bonum intellectus, nihil erit verum nisi in comparatione ad intellectum. Sed aliquid comparatur ad intellectum tripliciter: uno modo si signum ad signatum, et sic oratio vera comparatur ad intellectum. Nam oratio, in eo quod vera, signum est veritatis in intellectu.’ Cf. Martin of Dacia, Super Perih. (ed. Roos, 245:25–246:3): ‘Verum complexum est dispositio intellectus sub qua refert unum ad aliud componendo vel separat unum ab alio dividendo, et talis intellectus sic se habet quod quando refert rem aliquam ad aliam, cum qua convenit, sicut hominem ad animal, vel quando separat rem a re, cum qua non convenit, sic est verus. Sed quando facit e converso, sic est falsus.’ Scotus bears witness to this position in his commentary [Duns Scotus, In Perih.1 (ed. Bonaventure, 48:3–8)]: ‘Item, veritas et falsitas tantum sunt in sermone ut in signo; ergo enuntiatio prolata illud significat in quo est veritas et falsitas. Illud est compositio et divisio intellectus, ut dicit Aristoteles…Ergo enuntiatio composita significat illud quod est in intellectu composito tantum. Ergo et enuntiationis partes significant ea quae sunt in intellectu simplici, cuiusmodi sunt species.’ For Martin’s position, see Mora-Márquez, ‘Martinus Dacus and Boethius Dacus’. See e.g. Duns Scotus, In Perih.1 (ed. Bonaventure, 50:9–11): ‘Et universaliter omnes propositiones essent falsae in quibus enuntiatur actus realis de aliquo subiecto, cuiusmodi est “homo currit” etc.’ See e.g. Duns Scotus, In Perih.1 (ed. Bonaventure, 50:18–20): ‘Tum quia omnis propositio esset vera in qua praedicatur hoc verbum “est” secundum adiacens, ut “Socrates est”, “Antichristus est”, quia species cuiuslibet subiecti, de qua enuntiamus “esse”, est.’ See e.g. Duns Scot, In Perih.1 (ed. Bonaventure, 50:6–9): ‘Tum quia omnis propositio affirmativa esset falsa in qua subiectum et praedicatum cognoscuntur per diversas species, ut illa “homo est animal”, cum alia species hominis per quam intelligitur, et alia animalis.’

SIGNIFICATION OF CONCEPTS AND THINGS

75

concept of man, the assertion could never be true because the composition of the concept of man and the act of running is impossible.80 However, Scotus hits the nail on the head by stressing that even if we accepted that the immediate significate of a word of first imposition is a concept, the fact that the concept is a likeness of external things would assure the reference to them in assertoric contexts, and thus the assertion’s truth-makers would be the ultimate, and not the immediate, significates of its terms: To this [objection] it must be said that insofar as several things are signified by the same [sign], one of which [i.e. the thought] is signified insofar as it is a sign of the other, if this [sign] is composed in a sentence with another [sign], this composition does not concern the signs [i.e. the thoughts] but the ultimate significates, which are not signs. And with the uttered assertion one does not signify a composition of species, but a composition of things, just as with the written sentence one does not signify a composition of utterances but a composition of things.81 Scotus thereby sketches a distinction between words that, qua signs, signify concepts and words that, qua tools for true descriptions of the world, signify things – a distinction that would likely be the reply given by Albert the Great and Martin of Dacia, along the lines of Boethius, to the detractors of their positions. As to this matter, Radulphus Brito puts forth a position that, once again, seems to relegate the notion of signification to its semiotic use. Brito claims that names do not signify things according to their intellectual existence, as was maintained by e.g. Albert the Great, because this would make it difficult to explain how real predication can be verified. On the other hand, neither do names signify things according to their external existence, because then intentional predication, e.g. ‘man is a species’, also could not be verified. Therefore, names signify things themselves regardless of their external or intellectual existence: 80 81

See text quoted above, p. 56, n. 43. Duns Scotus, In Perih.1 (ed. Bonaventure, 54:20–24): ‘…dicendum ad illud quod, quantumcumque per idem multa significantur quorum unum significatur in quantum est signum alterius, si illud in oratione componatur cum alio, non est compositio signorum sed significatorum ultimorum, quae non sunt signa. Et per enuntiationem prolatam non significatur compositio specierum sed rerum, sicut nec per orationem scriptam significatur compositio vocum sed rerum.’

76

chapter 2

And I say that some people want to claim that [the name] signifies the thing under its intellectual being, which I do not believe. However, I believe that [the name] signifies the essence in its indifference to one and the other [being], because some predicates are verified of a thing only with respect to its intellectual being, just as ‘genus’, ‘species’ and so on; but other [predicates] are verified of a thing with respect to its being in particular instances, just as ‘to drink’, ‘to eat’ and so on. If some name, e.g. ‘man’, or every name, signified the thing with respect to its intellectual being, then real predicates could not be verified of it… Likewise, if it signified the thing only with respect to its being in particular instances, intentional predicates could not be verified of it; but since both sorts of predicate are verified, it seems that [‘man’] signifies the thing indifferently.82 Consequently, names, qua signs, signify the ontologically neutral thing itself, and the assertions involving them would be verified on the basis of further conditions, whose consideration, as we shall see in Part 2, is not necessarily a logical matter per se. In Part 2 it will become evident that Brito’s notion of signification operates at a semiotic level that is also neutral as regards propositional verification. 2.2

Whether Words Lose Their Signification with the Destruction of Their Significate

The question whether an utterance (word/term/name) loses its signification with the destruction of the significate is raised both as a semiotic question regarding the significative character of words (dictiones) and as a question concerned with the truth-value of assertions with an empty term as a subject, namely when the question is a sub-problem of the sophisma83 ‘Whether “omnis 82

Radulphus Brito, In Met. (ed. Ebbesen, 111): ‘Et dico quod aliqui volunt dicere quod significat rem sub esse intellecto. Quod non credo, sed credo quod significat rem sub indifferentia ad utrumque, quia quaedam sunt praedicata quae verificantur de re solum pro esse intellecto, sicut genus et species et talia, aliqua autem verificantur de re pro esse quod habet in suppositis, sicut bibere, comedere et talia. Si nomen aliquod sicut “homo”, vel omne nomen, significaret rem pro esse intellecto, tunc de ipso non possent verificari praedicata realia…et similiter si significaret rem solum pro esse in suppositis, de ipsa non possent verificari praedicata intentionalia; cum ergo utraque praedicata verificentur, videtur quod significet rem cum indifferentia.’ 83 A sophisma is a statement whose sense is undetermined because it involves some logical or grammatical difficulty (e.g. the equivocation of its terms, its ambiguous syntactic construction,

SIGNIFICATION OF CONCEPTS AND THINGS

77

homo de necessitate est animal” is true when no man exists’ (henceforth ohnea).84 The question is also discussed from the latter perspective in

84

its paradoxical nature etc.), and hence needs to be solved. In the medieval scholarly environment, sophismata were solved at public disputational exercises, whose written counterparts form what is known today as the medieval sophismatic literature. Fundamental logical issues (e.g., the universal quantification of singletons and empty terms) were thoroughly discussed at sophismatic disputations, so as to render the medieval sophismatic literature an unavoidable source for historians of medieval logic. For a general account of sophismata and sophismatic literature, see F. Pironnet and J. Spruyt, ‘Sophismata’, Stanford Encyclopedia of Philosophy, http://plato.stanford.edu/entries/sophismata/; for an exhaustive catalogue of medieval sophis­ matic literature, see S. Ebbesen and F. Goubier, A Catalogue of 13th-Century Sophismata, 2. vols. (Paris: Vrin, 2010). For a list of the extant versions of this sophisma, see S. Ebbesen and F. Goubier, A Catalogue, pp. 21–61. This sophisma also discusses the nature of the modality of necessity, the nature of the composition entailed by the term ‘est’, and the nature of the distribution brought about by the universal quantifier ‘omnis’. I shall focus only on the problem of the relation between the signification of empty terms and the truth-value of the assertions that have them as subjects. For discussions about ‘omnis’ in commentaries on Aristotle’s Topics from the last quarter of the 13th century, see A.M. Mora-Márquez, ‘Boethius of Dacia and Radulphus Brito on the Universal Sign “Omnis”’, Logica Universalis (2015), doi 10.1007/s11787-014-0112-6. For 13th-century discussions about modality, quantification, predication and empty reference, see A. de Libera, ‘Faire de nécessité loi: Théories de la modalité dans le sophisma “Omnis homo de necessitate est animal” du codex parisinus 16135, ff. 11rb–12rb’, ahdlma 76 (2009), 179–233; id., ‘Omnis homo de necessitate est animal. Référence et modalité selon l’Anonymus Erfordensis Q. 328 (Pseudo-Robert Kilwardby)’, ahdlma 69 (2002), 201–237; id., La référence vide (Paris: puf, 2002); id., ‘Référence et quantification: Sur la théorie de la distributio au xiiie siècle’, in A. de Libera, A. ElamraniJamal and A. Galonnier (eds.), Langages et philosophie: Hommage à Jean Jolivet (Paris: Vrin, 1997), pp. 177–200; id., César et le phénix: Distinctiones et sophismata parisiens du xiiie siècle (Pisa: Centro di cultura medievale della Scuola Normale Superiore, 1991); A. de Libera and L. Gazziero, ‘Le Sophisma “Omnis homo de necessitate est animal” du Parisinus Latinus 16135, ff. 99rb–103vb’, ahdlma 75 (2008.1), 323–368; L. Cesalli, A. de Libera and F. Goubier, ‘Does Loving Every Mean Loving Every Every, even Non-existent Ones? Distribution and Universals in the Opus puerorum’, in Fink et al. (eds.), Logic and Language, pp. 305–336; Cesalli, Le réalisme propositionnel; S. Ebbesen, ‘The Present King of France wears Hypothetical Shoes with Categorical Laces: Twelfth-century writers on well-formedness’, Medioevo 7 (1981), 91–113 (repr. in Ebbesen, Collected Essays, vol. 2, pp. 15–30); id., ‘The Dead Man is Alive’, Synthese 40 (1979), 43–70; id., ‘The Chimaera’s Diary – edited by Sten Ebbesen’, in S. Knuuttila and J. Hintikka (eds.), The Logic of Being (Dordrecht: Reidel, 1986), pp. 115–143; F. Goubier, ‘Influences prédicatives et conséquences référentielles: un aspect de l’approche terministe de la première moitié du xiiie siècle’, cimagl 71 (2000), 37–70; A. Tabarroni, ‘Omnis phoenix est: quantification and existence in

78

chapter 2

commentaries on the Posterior Analytics,85 as well as in linguistic treatises, such as Roger Bacon’s De signis86 and Peter John Olivi’s Quaestiones logicales.87 In grammatical treatises, Ps.-Kilwardby,88 Michael of Marbais89 and John of Dacia90 raise the question from the semiotic perspective. Finally, the question is also raised in the commentaries on the Perihermeneias and on the Metaphysics, by e.g. Siger of Brabant (Metaph.),91 Radulphus Brito (Perih.)92 and Duns Scotus (Perih.),93 where the semiotic question is entangled with the problem of empty reference. As already mentioned, when the question is a sub-problem of the sophisma ohnea, what is at stake is the relation between signification and the truthvalue of assertions.94 It is noteworthy, though, that some dubitationes related a new sophismata collection (ms Clm 14522)’, in S. Read (ed.), Sophisms in Medieval Grammar and Logic (Dortrecht: Kluwer, 1993), pp. 185–201. 85 See e.g. the commentaries on the Posterior Analytics by Simon of Faversham, q. 21 [in J.L. Longeway, Simon of Faversham’s Questions on the Posterior Analytics: a Thirteenth-Century View of Science (Ithaca: Cornell University, 1977; PhD dissertation)] and by Radulphus Brito, In APo, q. I.38, ed. S. Ebbesen, ‘Talking about what is no more’, cimagl 55 (1987), 135–168, at pp. 161–168. 86 Roger Bacon, De signis iv.2 (ed. Fredborg et al., 127): ‘Utrum de necessitate remaneat dictio significativa postquam imposita est ad significandum, ita quod non poterit fieri nota significativa, ut aestimatur a vulgo insensato.’ 87 Peter John Olivi, ql, q. 3 (ed. Brown, 344): ‘Post hoc quaeritur ad veritatem propositionum necessariarum et affirmativarum sequatur actualis entitas subiecti et praedicati. Utpote, an sequatur “homo est animal, ergo est actu”.’ 88 Ps.-Kilwardby, In Prisc. mai. (ed. Fredborg et al., 71): ‘Utrum dictio, postquam significat aliquod ens, eo corrupto amittat suam significationem.’ 89 Michael of Marbais, Summa (ed. Kelly, 10): ‘Qualiter ratio significandi remanere potest in dictione ipsa re corrupta secundum existere.’ In: Summa de modis significandi, ed. L.G Kelly (Stuttgart: Frommann-Hoolzboog, 1995) (Grammatica Speculativa). 90 John of Dacia, Summa (ed. Otto, 183): ‘Utrum scilicet vox post suam impositionem possit destitui a sua significatione, et hoc est quaerere utrum vox significativa possit fieri non significativa.’ 91 Siger of Brabant, In Metaph. (ed. Maurer, 160): ‘Utrum nomen significet idem, re existente et corrupta’. In: Quaestiones in Metaphysicam, ed. A. Mauer (Louvain-la-neuve: Editions de l’Institut Supérieur de Philosophie, 1983) (Philosophes Médiévaux xxv). 92 Radulphus Brito, In Perih. (ed. Venice, 144): ‘Consequenter quaeritur utrum vox significet idem re existente et non existente.’ 93 Duns Scotus, In Perih.1 (ed. Bonaventure, 61): ‘Quaeritur utrum facta transmutatione circa rem quae significatur, fiat transmutatio in significatione vocis.’ 94 This problem is also tackled in sophismata that do not explicitly raise it as a sub-problem. The most important are: An. Liberanus, ohnea, ed. de Libera, in ahdlma 76 (2009), 179–233; An. Erfordensis, ohnea, ed. de Libera, in ahdlma 69 (2002), 201–237; Siger of

SIGNIFICATION OF CONCEPTS AND THINGS

79

to ohnea are found as early as the 1230s in the syncategorematic treatises (Syncategoremata95) by Nicholas of Paris, John Pagus, Peter of Spain and William of Sherwood.96 However, these early discussions do not revolve expli­ citly around the problem of empty reference, but rather around the problem of the composition imported by the copula ‘est’, the kind of necessity imported by the modal operator ‘necessario’ and the distribution imported by the universal quantifier ‘omnis’.

Brabant, ohnea, ed. Bazan [in: Ecrits de logique, de morale et de physique, ed. B. Bazán (Louvain: Publications Universitaires, 1974) (Philosophes Médiévaux xiv)], at pp. 43–52; An., ohnea, ms Wigornensi, Bibl. Cath. Q.13, sophisma 22, f. 74ra; An. Pragensis, ohnea, MS Praha, MK M.83, f. 45ra; An., ohnea, MS Cambridge, G&C 611/341, sophisma 7, f. 58va, ed. S. Ebbesen, in ‘By Necessity’, in V. Hirvonen, T.J. Holopainen and M. Tuominen (eds.), Mind and Modality. Studies in the History of Philosophy in Honour of Simo Knuuttila (Leiden: Brill, 2006), pp. 141–151; An., ohnea, MS Mainz 616, distinctio 21, f. 47va; An., ohnea, MS Brugge, sb 509, f. 106ra. 95 Syncategorematic treatises are medieval logical texts that discuss the logical features of syncategorematic terms. A syncategorematic term, in turn, is a term that does not have signification in itself, yet it affects the signification of the categorematic terms (i.e. the subject-term and the predicate-term that have signification in themselves) that it determines. Cf. Peter of Spain, Sync. (ed. De Rijk, 39) [in: Peter of Spain, Syncategoremata, ed. L.M. de Rijk (Leiden: Brill, 1992)]: ‘Dictiones sincategorematicae significant res aliquas. Sed non significant res subicibiles vel praedicabiles. Ergo significant res quae sunt dispositiones subicibilium vel praedicabilium… Sed dispositio item, sive res quae est dispositio, est duplex, quia est quaedam dispositio eius quod est subiectum vel eius quod est praedicatum…sicut pater dicitur ad filium et econverso, ita subiectum ad praedicatum et econverso… Alia est dispositio subiecti inquantum est subiectum vel praedicati inquantum est praedicatum…et illae non subiciuntur neque praedicantur, quia sunt ipsius subiecti in comparatione ad praedicatum et econverso. Et tales dispositiones significantur per dictiones sincategorematicas; dicunt enim comparationes sive habitudines subiecti inquantum subicibile et praedicati inquantum praedicabile.’ Among syncategorematic terms, we find e.g. quantifiers (‘omnis’, ‘aliquis’ and so on), the copula ‘est’, conjunctions and disjunctions, the markers of conditional and interrogative statements (e.g., ‘si’ and ‘utrum’), adversative terms (e.g. ‘etsi’) etc. For a comprehensive study of 13th-century theories of syncategorematic words, see F. Goubier, ‘Les syncatégorèmes au xiiie siècle’, Histoire Epistémologie Langage 25/ii (2003), 85–113. 96 Nicholas of Paris, Sync., ed. H.A.G. Braakhuis, De 13de Eeuwse Tractaten over Syncategorematische Termen, ii: Uitgave van Nicolaas van Parij’s Sincategoremata, 2 vols. (Meppel: Krips Repro, 1979); Johannes Pagus, Sync., ed. Braakhuis, De 13de Eeuwse Tractaten, vol. 1; William of Sherwood, Sync., ed. J.R. O’Donnell, ‘The Syncategoremata of William of Sherwood’, Mediaeval Studies 3 (1941), 46–93; Peter of Spain, Sync., ed. L.M. de Rijk (Leiden: Brill, 1992).

80

chapter 2

In the present chapter, I shall focus on the discussions found in Roger Bacon’s De signis iv.2, Peter John Olivi’s Quaestiones logicales, q. 2–3, Boethius of Dacia’s ohnea and Anonymus Alani’s ohnea. These texts present us with four different ways of relating the question about signification to ohnea, and therefore with four different positions regarding the relation between signification and empty reference: (i) Terms lose their signification with the destruction of things, therefore ohnea is neither true nor false when no man exists (Roger Bacon). (ii) Terms do not lose their signification with the destruction of things, nevertheless ohnea is false when no man exists (Boethius of Dacia; cf. Radulphus Brito). (iii) Terms do not lose their signification with the destruction of things, and ohnea is false in one sense and true in another, when no man exists, because ‘esse’ is equivocal to dispositional and actual being (Olivi). Finally, (iv) terms do not lose their signification with the destruction of things, and the assertion ohnea is always true, whether man exists or not (Anonymus Alani). I shall use Anonymus Alani (henceforth An. Al.) as the guiding thread of my analysis, inserting the other positions where relevant.

Anonymus Alani’s ‘Omnis Homo De Necessitate Est Animal’ (Paris bnf Lat. 16135, ff. 99rb–103vb) This sophisma is of great interest for our discussion; for, just as Scotus does in his question about the immediate signification of words, An. Al. also discusses at length the most influential treatments of ohnea that were put forth during the thirteenth century. Yet, contrary to Scotus, who does not introduce his own solution to the question he presents, the anonymous author introduces his own solution in radical opposition to the positions he presents and subsequently rejects. An. Al. discusses ohnea by dividing it into three problems (problemata): (a) whether the necessity of the composition requires the necessary existence of the subject and the predicate; (b) whether the assertion ‘ohnea’ is true or false; and (c) whether a term loses its significate with the destruction of the thing it signifies.97 An. Al. connects the three problems by starting his discussion with problem (a) and by introducing problems (b) and (c) in the places where they (sort of) naturally belong. Accordingly, he does not give a separate treatment to each problem (as e.g. Boethius of Dacia does), but rather intertwines the three discussions.

97

An. Al., ohnea (ed. de Libera and Gazziero, 332): ‘Circa istud sophisma plura fuerunt quaesita, sed tantum tria prosecuta. Primum est utrum necessitas compositionis requirat necessario existentiam extremorum. Secundum est de veritate et falsitate primae propositionis. Tertium utrum terminus cadat a suo significato, quando res desinit esse’.

SIGNIFICATION OF CONCEPTS AND THINGS

81

After having put forth his own solution of the sophisma, An. Al. closes his discussion with short summaries of his own solutions to problems (a)–(c).98 The discussion starts with the arguments proving directly that the necessity of the linguistic composition marked with ‘est’ requires the necessary existence of the subject and the predicate (in at least one individual, one should add).99 The problem amounts then to asking if for ‘ohnea’ to be true, at least one individual man must exist. A related question, which is often discussed separately in another group of sophismata, asks whether there have to be at least three individuals for the quantification with ‘omnis’ to make sense at all. But since this question is not directly related to the problems around the notion of signification, I shall leave it aside in this study.100 Several arguments support the position that indeed there has to be at least one individual man for the assertion to be true. The most important of these arguments goes as follows: (a) In an assertion, the composition is to the terms as form is to matter. (b) But a necessary form (i.e. necessary composition) cannot inform destructible matter (i.e. man and animal). (c) Therefore the necessity of the composition marked with ‘est’ entails the necessary existence of the terms ‘homo’ and ‘animal’.101 In other words, there cannot be necessary composition if it is not the case that the existence of at least one actual compound is necessary; however, no material compound has necessary existence. Therefore, if we hypothesise that no man exists, which is plausible since no material compound has necessary existence, then the modal assertion ohnea is false.102 It is worth noting that this is a metaphysical argument, which appeals to the contingence of material compounds: 98 99

See An. Al., ohnea (ed. de Libera and Gazziero, 360–368). By ‘existence of the subject and predicate’ I mean now and henceforth the existence of the significates of the subject and the predicate. 100 This problem is mainly discussed in the sophisma ‘Omnis phoenix est animal’. For a list of this sophisma, see Ebbesen and Goubier, A Catalogue, vol. 2, pp. 362–370. For The Phoenix Complex, see Cesalli et al., ‘Does loving every mean loving every every?’. 101 An. Al., ohnea (ed. de Libera and Gazziero, 334): ‘Item. Compositio sic se habet ad extrema sicut forma ad materiam; sed in materia corruptibili et non necessaria non potest esse forma necessaria; ergo in extremis corruptibilibus non potest esse compositio necessaria; ergo ad necessitatem compositionis requiritur necessitas extremorum’. The author puts forth 10 further arguments defending the same position; cf. An. Al., ohnea (ed. de Libera and Gazziero, 332–336). 102 As we shall see, Boethius of Dacia grants this argument.

82

chapter 2

Since the significates of the terms of an assertion about the material world can never be involved in a necessary real composition, no assertion about the material world is necessary. Let us now get rid of the universal quantifier ‘omnis’ and ask whether ‘homo est animal’ is true or false, when no man exists. For if it is true, then some sort of necessity seems to be involved in essential predication (i.e. the one involved in definitions of the sort ‘species + est + genus + specific difference’), which may be the one grounding the necessary and universally quantified assertion ‘ohnea’. The equivalence between essential predication, universal quantification and necessity of the composition is put forth explicitly by, for instance, the Anonymus Erfordensis, who claims that whatever inheres per se inheres in every individual and inheres by necessity: Whatever inheres in itself (per se), inheres in every individual and by necessity. But animal inheres in itself in man, because it is in the first mode of saying ‘in itself’, as a part of the definition inheres in itself in what is defined. Therefore, animal inheres in man, and by necessity. Therefore this proposition is necessarily true. It seems then that ‘ohnea’ is true.103 This equivalence leads us to the first and second senses of per se attribution that Aristotle puts forth in Posterior Analytics I.4: (i) When we attribute to a subject parts of its definition, e.g. ‘animal’ to ‘man’. (ii) When we attribute to a subject a pair of opposites that contains it, e.g. ‘even/odd’ to ‘natural number’.104 These two sorts of per se attribution, Aristotle claims, are universal and necessary.105 In other words, if the extension of A is by definition contained in the 103 An. Er., ohnea (ed. de Libera, 222): ‘Quod per se inest omni inest et de necessitate. Sed animal per se inest homini, quia in primo modo dicendi per se, ut pars definitionis per se inest definito; ergo animal inest homini et de necessitate; ergo propositio haec de necessitate est vera. Videtur igitur quod haec est vera “omnis homo de necessitate est animal”.’ 104 cf. al, APo I.4 (ed. Hamesse, 314:38–41): ‘Per se primo sunt quaecumque in ratione dicente quod quid est insunt…per se secundo sunt, quaecumque insunt ipsis…ut propria passio per se inest subiecto…Item per se sunt, quae non dicuntur de quodam alio subiecto, ut substantia. Item per se sunt, propter quod inest alicui aliquid, tamquam per causam. Hic habemus, quod quattuor sunt modi dicendi per se, quorum primus et secundus tantum ingrediuntur demonstrationem.’ I leave aside the question whether the last two senses are modes of per se predication. For Aristotle’s theory of predication, see A. Bäck, Aristotle’s Theory of Predication (Leiden: Brill, 2000), Ch. 6. 105 Cf. Arist., APo I.4.

SIGNIFICATION OF CONCEPTS AND THINGS

83

extension of B (e.g. the extension of ‘man’ in the extension of ‘animal’), or if A and B have by definition exactly the same extension (e.g. the extension of even + odd and the extension of natural numbers), then ‘A is B’ is both necessary and universal. Note, however, that the ‘per se, ergo universal, ergo necessary’-inference presupposes that A and B are not empty terms; for in the context of the Posterior Analytics the existence of the subject is presupposed in any demonstration of its attributes. It is also noteworthy that to presuppose the existence of a subject is not quite the same as to require its necessary existence. Yet, we see this requirement made in the arguments that block this inference because of the empty subject of ‘homo est animal’ (i.e. when it is hypothesised that no man exists). Two of the four arguments that prove ‘homo est animal’ to be false when no man exists go as follows: Argument 1 (a) The cause of a true essential assertion is its subject (cf. Posterior Analytics, I.4.73b10–25). (b) The cause of the truth of ‘homo est animal’ is man (applying a. to ‘homo est animal’). (c) But no man exists (by hypothesis). (d) Therefore ‘homo est animal’ has no subject that can bear the predicate and produce its truth. (e) Therefore the assertion is not true (applying a.). (f) Therefore it is false (principle of bivalence, PoB).106 Argument 2 (a) An assertion is true because it indicates a true mental composition or division (cf. Arist., Int. 1). (b) But the intellect composes and divides actual things, and not their concepts (otherwise every mental composition would be false, because all concepts are different from each other). (c) But no man exists (by hypothesis). 106 An. Al., ohnea (ed. de Libera and Gazziero, 337): ‘Item. Propositio per se vera habet causam suae veritatis in subiecto; sed nullo homine existente haec “homo est animal” non potest habere causam suae veritatis in subiecto, quia subiectum per hypothesim non est et in eo quod non est nihil esse potest; ergo ipsa non erit per se vera nullo homine existente.’

84

chapter 2

(d) Therefore there cannot be true mental composition of man and animal (applying b.). (e) Therefore ‘homo est animal’ is not true (applying a.). (f) Therefore it is false (applying the PoB). (g) Therefore ‘omnis homo est animal’ and ‘omnis homo de necessitate est animal’ are false (since the antecedent of the inference ‘per se, ergo universal, ergo necessary’ is false).107 Note that in both arguments, the steps e. to f. entail the use of the principle of bivalence. But in Perihermeneias 7, Aristotle explicitly says that ‘non-universal [assertions] about universals are not always one true, the other false; for it is possible to claim truly and at the same time that man is white and that man is not white’.108 This is equivalent to saying that indefinite assertions (e.g. ‘homo est animal’) do not comply with the PoB. Furthermore, it must be recalled that even in the case of the quantified assertion ‘omnis homo est animal’, the Aristotelian version of the principle of bivalence seems to presuppose that at least one individual of the universal term exists.109 Thus, it is at least problematic to assume the universal term to be empty and at the same time apply the PoB. As it seems, then, the hypothetical assumption that ‘homo’ is empty places the assertion in question in a paradoxical situation with regard to Aristotelian logic. On the other hand, those who grant ‘ohnea’ propose that whoever denies it, whether a man exists or not, puts forth a contradiction by claiming and denying the same thing of the same thing. The argument (Argument 3), which 107 An. Al., ohnea (ed. de Libera and Gazziero, 337): ‘Quia veritas est in oratione sicut in signo. Significat enim oratio veritatem quae est circa compositionem et divisionem intellectus componentis vel dividentis; intellectus autem non componit intentiones quas habet penes se de rebus extra ita quod dicat quod haec est illa; sic enim esset omnis compositio eius falsa; et etiam intelligens et iudicans non iudicat de ipsis intentionibus vel speciebus sed de rebus quarum sunt; componit ergo ad invicem ; res autem vere componere non potest nisi res ipsae sint; ergo ad hoc quod compositio designata per hanc “omnis homo est animal” sit vera, oportet quod omnis homo sit; cum ergo hominem esse non sit necessarium, non erit haec necessaria “omnis homo est animal”; ergo nec erit haec vera “omnis homo de necessitate est animal”.’ 108 Cf. Arist., Int. 7.17b29–32. 109 See Crivelli, Aristotle on Truth, p. 161. In fact, the only case where Aristotle considers an empty term in the whole Perihermeneias is in the a fortiori argument of the ‘goat-stag’, which he puts forth in Chapter 1, in order to show that if not even an empty term posits the non existence of its significate, a fortiori names in general do not posit the existence or non-existence of their significates either. Assertions, however, would seem to posit the existence of the subject about which something (i.e. an attribute) is stated.

SIGNIFICATION OF CONCEPTS AND THINGS

85

starts by denying ‘omnis homo est animal (ohea)’, can be reconstructed as follows: (a) In ‘ohea’, ‘homo’ has personal supposition (i.e. it stands for all its particular instances).110 (b) Every assertion whose subject has personal supposition involves three relations: (i) between the subject and the predicate, (ii) between the subject and its particular instances and (iii) between the predicate and the particular instances of the subject.111 (c) The affirmative assertion of this kind states the three relations. (d) Its negation denies the first and the third relation, while it states the second. (e) Therefore the negative assertion ‘non ohea’ posits that ‘homo’ stands for all its particular instances, but it denies that ‘est animal’ applies to all of them. (f) Therefore there is at least one particular instance of man, e.g. Plato, who does not bear a relation with being animal. 110 Supposition is a semantic feature of categorematic terms that was famously introduced in some 13th-century logical treatises on the properties of terms. The supposition of a cate­ gorematic term is, generally speaking, that for which the term stands when it is used in an assertoric context – e.g. for the whole of its extension, for sub-sets of its extension, for itself, for the concept with which it is associated etc. In the case at stake here, the term ‘homo’ has confused personal supposition. According to Peter of Spain [in Tractactus, ed. L.M. de Rijk (Assen: Van Gorcum, 1972)], a common term has personal supposition when it stands for its individuals (cf. Peter of Spain, Tractatus (ed. de Rijk, 82:10–12): ‘Personalis suppositio est acceptio termini communis pro suis inferioribus. Ut cum dicitur “homo currit”, iste terminus “homo” supponit pro suis inferioribus.’). A common term has confused personal supposition, when it stands for all its individuals, namely when it is affected by the universal quantifier ‘omnis’ (cf. Peter of Spain, Tractatus (ed. De Rijk, 82:28–83:3): ‘Confusa suppositio est acceptio termini communis pro pluribus mediante signo universali. Ut cum dicitur “omnis homo est animal”, iste terminus “homo” mediante signo universali tenetur pro pluribus, quia pro quolibet suo supposito.’) The literature on medieval theories of supposition is extensive. For three recent general accounts, see J. Ashworth, ‘Terministic Logic’, in Pasnau (ed.), The Cambridge History, pp. 146–158; S. Read, ‘Medieval Theories: Properties of Terms’, Standford Encyclopedia of Philosophy, http:// plato.stanford.edu/entries/medieval-terms/; and Catarina Dutilh Novaes, ‘Medieval Theories of Supposition’, in H. Lagerlund (ed.), Encyclopedia of Medieval Philosophy (Berlin: Springer, 2011), pp. 1229–1236. 111 These three relations obtain fundamentally because of the universal quantification of a common term that is put in relation to a predicate in an assertion. In other words, because of the universal quantification of a common term that is the subject of an assertion.

86

chapter 2

(g) Therefore ‘Plato non est animal’ is true. (h) Therefore ‘Plato non est homo’ is true (since ‘homo’ is defined as ‘animal rationale’, so that whatever is not an animal, is not a man either). (i) But ‘Plato est homo’ is true (applying d.). (j) Therefore ‘non ohea’ entails both ‘Plato est homo’ and ‘Plato est non homo’. (k) Therefore ‘non ohea’ is impossible. (l) Therefore ‘ohea’ is necessary. (m) Therefore ‘ohnea’ is true.112 Yet, also in this case the argument assumes that there are particular instances of ‘homo’, which is what the sophisma denies by hypothesis. Thus, both the argument proving the falsity of ‘ohnea’ and the one proving its truth are put on the spot precisely because of the assumption that no man exists. The paradoxical nature of the sophisma ohnea very much depends on the logical consequences both of the universal quantification of empty terms and of the modality of necessity in compositions where they are used, because evidently the logical properties of the syncategoremes ‘omnis’ and ‘necessario’ depend to a great extent on the non-emptiness of the categorematic terms that they affect. But what is it, after all, for a categorematic term (whether one calls it utterance, word, term or name) to be empty? Is it equivalent to not having signification? If so, the consequences for logic would be dramatic, because without a semiotic relation on which to support the logical properties of cate­ gorematic and syncategorematic terms, fundamental tenets of medieval logic would collapse. Then, even the most radical detractors of ‘ohnea’, such as Boethius of Dacia, accept that terms do not lose their signification with the destruction of things. Indeed, to the best of my knowledge the only extant pro112 An. Al., ohnea (ed. de Libera and Gazziero, 338): ‘Item. Quicumque negat eam, sive sit homo sive non, idem affirmat et negat; ergo sua negativa est impossibilis. Probatio antecedentis: terminus habens personalem suppositionem supponit suum significatum pro appellato vel appellatis…ergo per dictum tuum qui negat eam removet esse animal ab appellatis “hominis”; ergo cum ista negativa designet tres habitudines sicut et quaelibet negativa cuius subiectum habet personalem suppositionem, unam praedicati ad subiectum, aliam subiecti ad suum suppositum, tertiam praedicati ad suppositum subiecti, prima designabitur in ea negative, secunda affirmative et tertia negative… Semper in propositione autem affirmativa designantur esse tres dictae habitudines affirmativae; tu ergo in tua negativa qua explicite negas esse animal de homine affirmas implicite esse hominem de eius supposito et removes animal de eodem, quod idem est implicite ac si explicite diceres quod ipse est homo et non est animal… Ergo patet quod per tuam negativam dicis quod ipse est homo et non est homo. Et sic idem affirmas et negas.’

SIGNIFICATION OF CONCEPTS AND THINGS

87

ponent of something close to the opposite position is Roger Bacon, who is always a somewhat dissident voice in thirteenth-century discussions of logical and semantic nature. Roger Bacon’s De Signis iv.2 In his De signis, Roger Bacon raises the question whether a word remains significative after its imposition. He then goes on to introduce a difficult case, namely the one where the thing signified ceases to exist. Other than the usual authoritative argument taken from Boethius’ De divisione,113 Bacon introduces his own argument aiming to prove that a word actually loses its signification with the destruction of the thing it signifies. Since signification is a relation between an utterance and a thing, and every relation is destroyed with the destruction of one of its correlates, if the thing is destroyed the relation of signification is also destroyed: Likewise, once one of the extremes opposed in a relation is destroyed, even though the other one is not destroyed according to its substance, nonetheless the relation and the disposition of one to the other is destroyed, as when the substance of Socrates – who was Plato’s son – is destroyed; [for] even though Plato is alive, there is no paternity in Plato… Whence, with the destruction of the thing signified, even though the substance of the utterance remains, however it does not remain with respect to the notion of sign.114 113 Roger Bacon, De signis IV.2 (ed. Fredborg et al., 128): ‘Item, propter defectum rei et corruptionem hoc idem ostenditur post per Boethium, qui dicit in De divisionibus, quando vox imponitur alicui rei et illa res corrumpitur, ut non sit sive fiat non ens, iam fit vox non significativa.’ Cf. Boethius, De divisione (ed. Magee, 44:20–24): ‘Oportet autem maxime exercere hanc artem, ut ipse Aristoteles ait, contra sophisticas importunitates, si enim nulla subiecta sit res quam significat vox, designativa esse non dicitur, sin vero una res sit quam significat vox, dicitur simplex, quod si plures, multiplex et multa significans.’ 114 Roger Bacon, De signis iv.2 (ed. Fredborg et al., 128): ‘Item, corrupto uno extremorum relative oppositorum licet non corrumpatur alterum secundum substantiam eius, corrumpitur tamen relatio et habitudo unius ad aliud, ut corrupta substantia Socratis, qui fuit filius Platonis, Plato licet vivat, tamen paternitas non manet in Platone… Quare re significata corrupta licet substantia vocis maneat non tamen ratione signi.’ Bacon’s radical extensionalism entails, on one hand, the rejection of the notions of esse habitualis and esse essentiae that are used by other authors in order to provide the words’ significata with some ontological stability. It entails, on the other hand, the consideration of the copula ‘est’ as an operator that points to a relation of inherence between two real things. Bacon’s extensionalism could be related to his participation in some theological debates during the first half of the 13th century. One of these debates is concerned with the thesis

88

chapter 2

Bacon gives further support to his argument with an example that is reproduced in An. Al.’s ohnea115 – the sign of wine outside a shop does not have any signification when wine is sold out, and if it can still be interpreted as the sign of something, it would be of the image of wine in the mind of those who do not know yet that wine is sold out, namely because of a re-imposition of the sign on that image: Furthermore, we see this in other signs; for when there is no wine in the shop, and the buyers know this, they do not take the circular figure [that is a sign] of the wine to be a sign anymore, just as also in the case of other [signs]. Whence, with the destruction of the thing on which the imposition was made, the word does not continue to be significative. If someone says: ‘It actually happens that there is no wine in the shop, and nonetheless the buyers that see the circular figure ask for wine, because of the figure, therefore the figure is still a sign for them’, it must be said to this that for the ones who know that there is no wine here, the circular figure that was a real sign, and of real wine, loses for them, at this time, the account of a real sign. But the ones who do not know that there is no wine and that suppose that the shop has [it], since there is no [wine], when they see the figure, they suppose that the circular figure signifies wine and they renew the account of sign with respect to [the figure], so that in their minds they make the figure be a sign of the wine that is supposed and imagined by them, and thus the circular figure is made a new sign. But it loses the first account of signification that it had with respect to real wine. And it can happen in the same way in the case of utterances.116 (e.g. by Richard of Cornwall) that ‘homo’ is said univocally of Jesus and of his dead body during the three days before his resurrection. In the Compendium Bacon severely rejects this thesis and claims that Jesus’ dead body can only be called ‘homo’ after a reimposition of the name, which in passing makes it an equivocal name (cf. Roger Bacon, cst (ed. Maloney, 52:7–64:3) [in: Compendium of the Study of Theology (Leiden: Brill, 1988) (stgm 20)]). See also A. de Libera, ‘Roger Bacon et la référence vide. Sur quelques antécédents médiévaux du paradoxe de Meinong’, in J. Jolivet et al. (eds.), Lectionum varietates. Hommage à Paul Vignaux (1904–1987) (Paris: Vrin, 1991), pp. 85–120; and id., ‘Roger Bacon et le problème de l’appellatio univoca’, in Braakhuis et al. (eds.), English Logic and Semantics: From the end of the twelfth century to the time of Ockham and Burleigh. Acts of the 4th European Symposium on Medieval Logic and Semantics (Leiden/Nijmegen: Ingenium Publishers, 1981), pp. 193–221; for Bacon on equivocal words, see T.S. Maloney, ‘Roger Bacon on Equivocation’, Vivarium 22 (1984), 85–112. 115 See An. Al., ohnea (ed. de Libera and Gazziero, 350). 116 Roger Bacon, De signis iv.2 (ed. Fredborg et al., 128): ‘Item, hoc videmus in signis aliis, cum enim non vinum in taberna et emptores vini hoc , iam circulus vini non

SIGNIFICATION OF CONCEPTS AND THINGS

89

This strange example, where we are told that signs of non-existing things are re-imposed on an image of the thing in the mind of the interpreters, has an unfortunate consequence: We are forced to suppose that the buyers who do not know that there is no wine re-impose the sign without being aware of doing so (they do not know that there is no wine, so they cannot know that they need to re-impose the sign).117 Notwithstanding, Bacon adds that the same happens in the case of words, so that ‘Socrates’, which was imposed to signify a person who is now dead, lost its original signification when Socrates died and is now used not according to its original signification but according to a renewed signification. It is noteworthy, however, that contrary to the wine example, most people would be aware that Socrates is dead, so that in this case one could assume a voluntary re-imposition of ‘Socrates’ by its users. Bacon explains further the mechanisms of re-imposition by putting forth two modes of imposition: The first occurs when an utterance is given, at its original imposition, to a thing as its name, i.e. as in a baptism. The second, which takes place in the mind, occurs when ordinary speakers think of something (whether existing or not) about which they want to talk and re-impose words on it (words that according to the first mode of imposition were imposed on something else).118 True to the Franciscan focus on human will, Bacon explains that re-imposition is possible because the act of imposition is directed recipitur apud eos pro signo, et ita est in aliis. Quapropter corrupta re, cui facta est impositio, non remanebit vox significativa. Si dicatur: bene accidit quod vinum non sit in ta­berna, et tamen emptores videntes circulum quaerunt vinum propter circulum, ergo circulus adhuc est eis signum, quare similiter hoc dicendum est quod scientibus, quod non sit ibi vinum, circulus qui fuit signum verum et veri vini iam cadit apud eos et secundum veritatem a ratione signi. Sed ignorantes vinum non esse et putantes esse in taberna, cum non sit, videntes circulum aestimant circulum significare sibi vinum et renovant rationem signi in eo, ita quod apud imaginationem suam constituunt circulum esse signum vini aestimati ab eis et imaginati, et sic fit circulus signum novum. Sed cecidit a ratione prima significandi, quam habuit respectu vini veri. Et similiter potest esse de vocibus.’ The circulus vini example most likely comes from a theological context, since it is most often used by theologians in discussions about the nature of signs, notably in discussions about the semiotic character of sacraments; see Rosier-Catach, La parole efficace. 117 I am indebted to Chris Martin for pointing out to me the problem with this example. It seems to me, indeed, that there is a serious problem with a volontary notion of reimposition of which language-users are not aware, particularly since Bacon puts a lot of weight on the intentional character of linguistic imposition. See above, Section 2.1. 118 Roger Bacon, De signis iv.3 (ed. Fredborg et al., 130): ‘Ad horum ergo evidentiam potest quaeri de modo imponendi duplici an sit, ut tactum est, et dicendum quod sic. Unde unus est modus imponendi sub forma impositionis vocaliter expressa et assignata rei sicut imponuntur nomina infantibus et aliis rebus. vel non ens, de quo vult aliquid enuntiare, vel quod de alio vult enuntiare, et sic imponit nomen.’ 119 Roger Bacon, De signis iv.3 (ed. Fredborg et al., 130): ‘Quod autem hoc sit possibile manifestum est, quia positio est ad placitum, ergo secundum quod placet homini, potest in mente sua dare vocem rei vel exprimere impositionem vocaliter. Sic potest de omnibus aliis, de quibus cogitat et quae vult.’ 120 Roger Bacon, De signis iv.3 (ed. Fredborg et al., 131–132): ‘Si dicatur quod tunc avertemus impositiones istas, sed nullus quidem avertit, quando imponit, dicendum quod ubi non diligenter consideraret non bene averteret rationem et horam imponendi vocaliter expressa quia haec est prima et principalis et consueta in rebus et linguis… Et haec est prima causa latentiae… Tertia causa est quod non fiunt multum istae impositiones, nisi simul fiant enuntiationes de illis terminis, ita quod non est tempus sensibile nec minimum inter impositionem et enuntiationem…’. 121 This, of course, does not resolve the problem with the circulum vini, where there is not even awareness of the need of re-imposing the sign on the image of wine.

SIGNIFICATION OF CONCEPTS AND THINGS

91

Bacon also tells us that any assertion that follows an act of re-imposition is necessarily true or false, because it necessarily indicates something that either is or is not the case about the object of re-imposition. But if the assertion did not follow from an act of re-imposition, it would not indicate at all something that is or is not the case, because its terms would have no signification. In other words, when the terms of an assertion have no signification, there is no assertion at all: It must be said that, if there is imposition, it is necessary for such sentence to be true or false, and when we want to signify something true or something false, it is necessary that we formerly impose and transfer the utterance in order to signify.122 Hence, when no man exists, if the speaker did not re-impose the terms ‘homo’ and ‘animal’ on something else, the assertion ‘ohnea’ is neither true nor false – there is in fact no assertion at all.123 But, as a matter of fact, the notion of re-imposition makes sure that all our assertions indicate something that we state to be the case, because whenever we want to predicate something, we re-impose the subject on the thing we want to talk about. Note that this has the unfortunate consequence of us dealing with a different assertion at every act of re-imposition, which in passing seems to hinder any possible explanation of human communication succeeding. To the Aristotelian idea that assertions go from being true to being false because of an accidental change within the subject, Bacon replies that this only applies to the case of accidental predication. In the case of essential pre­ dication, though, if the significates of the terms change (e.g. when they go from existing to non-existing), the assertions at stake, before the change and after the change, are different assertions. Then, ‘Socrates is’ is one assertion, and a true one, when Socrates is alive, but it is another assertion, and a false one,

122 Roger Bacon, De signis iv.2 (ed. Fredborg et al., 129): ‘Et dicendum est quod si fiat impositio, necesse est quod talis oratio sit propositio et sit vera vel falsa, et quando volumus significare verum vel falsum, necessario prius imponimus et transsumimus vocem ad significandum.’ 123 Roger Bacon, De signis iv.2 (ed. Fredborg et al., 129): ‘Sed si non fieret impositio aliqua, certum est quod talis oratio esset propositio neque vera neque falsa immo esset non significativa; ab una enim parte non significativa fit tota oratio non significativa, et hoc patet.’

92

chapter 2

when Socrates is dead,124 without this going against the PoB, because in these assertions ‘Socrates’ is equivocal to a living man and to a dead man. Consequently, thanks to his idea of re-imposition, Bacon does not need different linguistic properties in order to explain the constitution of signs, the sense of sentences and the verification of assertions. In fact, his account puts forth one coherent notion of signification that, thanks to the idea of reimposition, can operate at the semiotic, the communicational and the verificational level. Note also that, strictly speaking, for Bacon there are no empty terms, because once a term is used in the context of an assertion it has necessarily been re-imposed on something. Rather, the weakness of his account is that this idea of re-imposition seriously compromises the very possibility of linguistic communication. Boethius of Dacia’s Sophisma ohnea Another solution to the sophisma is to deny that a term loses its signification with the destruction of the thing, but to reject ‘ohnea’ as false. Boethius of Dacia is perhaps the most significant extant witness to this position. Boethius divides his treatment of the sophisma ohnea into four problems, of which only two concern us: (i) whether ‘ohnea’ is true when no man exists and (ii) whether terms lose their significates with the destruction of things. As to the first problem, Boethius claims that the assertion ‘ohnea’ is false because of its modality; for men and animals are destructible things susceptible of change – they are essences that inform destructible matter – and hence men and animals cannot be the cause of any necessary truth.125 Recall that this 124 Roger Bacon, De signis iv.2 (ed. Fredborg et al., 129–130): ‘Si obiciatur quod Aristoteles in Praedicamentis quod oratio remanet vera vel falsa, et mutatur de vera vel falsa, ut “Socrates sedet”, haec est vera, dum sedet, eo modo surgente haec eadem est falsa, et ita similiter “Socrates est”… Sed si eadem est oratio, idem est subiectum, ergo idem est significatum per vocem subiecti… Dicendum est quod ibi loquitur de praedicatis accidentalibus, et quae possunt separari a subiecto existente, ita quod dictio “Socrates” remaneat significativa (significativum ed.) eius, quod prius significabat… Sed si accipiatur praedicatum quod inest per se et essentialiter ut “Homo est ”, “Homo est” et huiusmodi, non est possibile quod fiat mutatio eiusdem orationis a vero in falsum, sed erit diversa et diversa oratio, et ideo diversum subiectum et diversum significatum subiecti et accipietur aequivoce. Ut dum Socrates est, haec est vera “Socrates est” et “Socrates” significat vivum. Dum vero Socrates non est et dicatur haec oratio “Socrates est”, haec est falsa, sed non est oratio eadem quae prius…’ 125 Boethius of Dacia, Omnis homo etc. (ed. Ebbesen, 7): ‘Cum ergo homo et animal sint res transmutabiles et corruptibiles et nihil sit in eis incorruptibile nisi materia prima…secundum quam materiam non verificatur ista propositio “homo de necessitate est animal”,

SIGNIFICATION OF CONCEPTS AND THINGS

93

is the argument with which An. Al. starts his discussion of ohnea,126 an argument that, as already mentioned, is based on a metaphysical constraint: No event in the material world is necessary, therefore no necessary statement about the material world is true. Boethius also puts forth the following argument, which he grants when he gives his own solution to the problem:127 (a) There is linguistic composition, mental composition and real composition. (b) The truth of linguistic composition depends on the truth of mental composition (because the truth of what is posterior is caused by the truth of what is prior; cf. Arist., Int. 1). (c) Likewise the cause of the true mental composition is the real composition – ‘the support and the cause of any posterior truth’. (d) Therefore, if there is no real composition, neither the mental composition nor the linguistic composition can be true (from b. and c.). (e) But it is impossible that there is real composition of man and animal when no man exists. (f) Therefore, when no man exists, it is impossible for any linguistic composition involving man and animal to be true.128

sequitur quod ex hiis rebus homo et animal nulla causatur veritas intransmutabilis nec etiam necessaria.’ In: Sophisma Omnis homo de necessitate est animal, mss Brugge, Stadsbibliotheek 509, ff. 87v–91v; Firenze, Biblioteca Medicea Laurenziana Pl. 12 sin. 3, ff. 63ra–64rb. Unpublished edition by S. Ebbesen. I owe thanks to Sten Ebbesen for having given me access to his unpublished edition of Boethius of Dacia’ sophisma. 126 See above, pp. 81–82. 127 Boethius of Dacia, Omnis homo etc. (ed. Ebbesen, 8): ‘Rationes probantes esse falsam concedo.’ 128 Boethius of Dacia, Omnis homo etc. (ed. Ebbesen, 3): ‘Triplex invenitur compositio, scilicet compositio quae est in re ipsa, compositio quae est in intellectu, et compositio quae est apud sermonem, ita quod in istis semper veritas quae est in posteriori compositione est ex veritate prioris compositionis tamquam ex causa sua…quomodo erit compositio vera apud intellectum nisi consimilis sit compositio in re? Si enim sit in re divisio et apud intellectum compositio, erit falsa compositio intellectus. Ex hoc arguitur sic: Sicut se habet compositio sermonis ad compositionem intellectus, sic se habet compositio intellectus ad compositionem rerum; sed compositio sermonis non potest esse vera nisi sit vera compositio intellectus ex qua ipsa est; ergo nec potest esse vera compositio intellectus nisi sit talis compositio in re. Sed nullo homine existente animal homini non componitur; ergo nec intellectus nec sermo componens animal homini potest esse verus, cum deficiat compositio quae est in re, quae est fundamentum et causa cuiuslibet veritatis posterioris, scilicet intellectus et sermonis.’

94

chapter 2

Note that Boethius rejects ‘ohnea’ as false based on a general metaphysical assumption – that the material world is contingent. However, his logicosemantic argument only shows that it cannot be true and he is careful enough not to conclude from this to its falsity. Some pages later, Boethius claims that the objection, according to which after the destruction of things terms signify the same as they did before, puts forth a less than necessary argument, because even if their signification is the same irrespective of the existence or non-existence of the things, signification is not the cause of truth, but the cause of truth is that things are or are not arranged in a certain way. As regards signification, Boethius accepts that terms do not lose their signification with the destruction of things, because their signification does not depend on the existence of their content – a term signifies the thing itself, which is indifferent to existence and non-existence.129 Boethius’ argument to support this position is based on the principle that signification follows understanding,130 i.e. since it is possible to think of Socrates when he is dead, it is also possible to signify him: When the things are destroyed, it is not necessary that the terms lose their significates, because the signification of the utterance does not depend on the existence of the thing… Likewise, whatever is possible for the intellect when it understands, is possible for utterances as regards signification; for whatever can be understood can be signified, and in the same mode; but when the thing is destroyed, it is possible for the concept of the thing to be in the intellect, as it was before; for it is possible to understand the thing after its destruction; therefore [it is] also [possible] to signify it, just as before.131 Thus, a term is significative insofar as its content can be understood, and this irrespective of the existence or non-existence of this content. However, the opposite is the case regarding propositional verification, because the use of a word in an assertion entails the presupposition of its 129 Cf. Radulphus Brito in Section 2.1. 130 For this principle, see Section 2.1. 131 Boethius of Dacia, Omnis homo etc. (ed. Ebbesen, 20–21): ‘...corruptis rebus non oportet terminos cadere a suis significatis; et causa huius est quia significare vocis non pendet ab esse rei… Item, quod est possibile apud intellectum in intelligendo, possibile est apud voces significando; nam quod potest esse intellectum potest esse vocis significatum et illo modo; sed re corrupta possibilis est apud intellectum conceptio illius rei sicut prius; res enim post sui corruptionem possibilis est intelligi; ergo et per vocem significari sicut prius.’

SIGNIFICATION OF CONCEPTS AND THINGS

95

content’s existence – it is not possible to state a real composition or division if there is nothing real to compose with or divide from something else. Accordingly, Boethius rejects the idea that mental existence is enough for the verification of essential predication when things do not exist. Man only exists when the essence of man is individuated, and the mere understanding of this essence does not suffice for verifying essential predication: We argue against the solution in the second argument: If the essence of man and of animal remained after the destruction of every man and every animal, either they remain in matter or in the soul. Not in matter, because when the man is destroyed, the matter of man takes an opposite form… Not in the soul, because being in the soul is not sufficient for the essences of natural things to verify sentences, but it is necessary for them to be in matter.132 According to Boethius, then, essential assertions about material substances have existential import, and hence ‘homo est animal’ and ‘homo est homo’ amount to asserting ‘homo qui est, est animal’ and ‘homo qui est, est homo’. Consequently, the non-existence of the assertion’s subject prevents the possibility of its verification – and possibly also of its falsification – regardless of whether the assertion is essential or accidental. Note, finally, that Boethius also shows the tendency to relegate the notion of signification to the semiotic domain, leaving it next to no role to play in propositional verification. Peter John Olivi’s Quaestiones Logicales q.3 A last solution accepts that terms do not lose their signification with the destruction of things, but establishes a distinction because of the equivocation of ‘being’ to dispositional being (esse habituale)133 and to actual being (esse actualis). According to actual being, ‘ohnea’ is false, but according to dispositional being it is true. Peter John Olivi mentions such a distinction in his reply to the question ‘Whether the actual existence of the subjects and the predicates 132 Boethius of Dacia, Omnis homo etc. (ed. Ebbesen, 6): ‘Contra solutionem secundi argumenti arguitur. Si omni homine et omni animali corrupto maneret essentia hominis et animalis, aut ergo in materia aut in anima. Non in materia, quia corrupto homine materia hominis est sub forma opposita… Non in anima, quia ad hoc quod essentia rerum naturalium verificet orationes non sufficit esse in anima, sed oportet quod sit in materia.’ 133 ‘Dispositional being’ indicates something’s tendency or capacity to be in one way or another, regardless of the actualisation of this tendency or capacity.

96

chapter 2

follows from the truth of necessary and affirmative propositions’ (Quaestiones logicales, q. 3).134 Olivi introduces us to his solution to this question by putting forth the ‘common and reasonable’ position, according to which with ‘being’ we do not always intend to express actual existence. One example is when we use ‘est’ as a third adjacent (i.e. as the copula), in which case we usually intend to express an essential relation and coherence between subject and predicate: I reply: Some claimed that the truth of such propositions demands the presupposition of the actual existence of the subject [i.e. Boethius of Dacia]… But some others claim, more commonly and reasonably, that with the verb of being [i.e. ‘to be’], when it is taken only under the notion of the copula, we do not always intend to signify actual existence, but rather the identity of [the subject and the predicate] or their essential disposition or coherence.135 With ‘homo est animal’, we intend to say that the notion of animal is an essential part of the notion of man, that in some sense man and animal are the same thing, or that man and animal are essentially related to each other, rather than to say that the men who exist are animals.136 Note also in passing that, just as in Bacon’s case, Olivi’s treatment of the question highlights the speaker’s intention as a determining factor of the content of the assertion’s terms.137 Thereafter, Olivi relates that the sort of being indicated by ‘est’, when it is a third adjacent, is commonly called ‘dispositional being’ or ‘being of essence’.138 However, dispositional being and actual being do not posit two different ontological statuses, but only two different ways of understanding an essence; for

134 Peter John Olivi, ql (ed. Brown, 344): ‘Post hoc quaeritur: ad veritatem propositionum necessariarum et affirmativarum sequatur actualis entitas subiecti et praedicati? Utpote: An sequatur “Homo est animal, ergo est actu”?’ 135 Peter John Olivi, ql (ed. Brown, 344): ‘Respondeo: Quidam dixerunt quod ad veritatem huiusmodi propositionum praeexigitur praesuppositio actualis exsistentiae subiecti… Alii vero communius et rationabilius dicunt quod per verbum essendi quando sumitur sub sola ratione copulativae duorum non semper intendimus significare exsistentiam actualem, sed potius eorum identitatem sive essentialem habitudinem et cohaerentiam.’ 136 Cf. Peter John Olivi, ql (ed. Brown, 344–345): ‘Unde cum dicimus “Homo est animal” non intendimus dicere quod homo sit actu animal, sed solum quod animalis ratio sic cohaeret rationi hominis quod est pars eius, et aliquo modo sunt idem.’ 137 See Section 2.1 and Mora-Márquez, ‘Pragmatics in Peter John Olivi’. 138 Peter John Olivi, ql (ed. Brown, 345): ‘Et “esse” sic acceptum per modum solius copulae… vocatur a quibusdam esse habituale vel esse essentiae.’

SIGNIFICATION OF CONCEPTS AND THINGS

97

dispositional being is attributed to the essence only insofar as it is susceptible of being understood without its actual existence: Regarding the third [argument], the reply is twofold: First, dispositional being is different from actual being, however not really, as if there were diverse real beings in things that exist, but they are different only according to different notions or modes, because the thing has dispositional being insofar as it can be understood without the simultaneous understanding of its actual existence. Hence, the first notion or the first mode of being can be grasped by the intellect without the second one; for just as the notion of the essence is different from the notion of its actual existence, without this entailing any real composition or difference, in the same way dispositional being is different from actual being, because the first concerns only the notion of the essence and of its essential [properties] as it goes together with the notion of actual existence.139 This is to say that the intellect attributes dispositional being to an essence, when the essence is understood without its actual existence, without this entailing that there is a real difference and a real composition between actual being and dispositional being. The only ontological realm of essences is actual being, the dispositional one being only a mode of understanding them. Accordingly, since it is possible to consider the essence in this way, it is also possible to talk about it in this way, i.e. without existential import. Olivi goes on to introduce two modes of approaching actual existence: a nominal or essential mode and a verbal mode. The nominal mode is when we understand the definition of an actual being but separate that definition from its existence. The verbal mode is when with a verb we express explicitly that an essence has actual existence: Second, it is said that actual being can be taken in two ways: In one way in the nominal mode, i.e. in the mode of the quidity, just as when we 139 Peter John Olivi, ql (ed. Brown, 345): ‘Ad tertium dupliciter respondetur: Primo, quod esse habituale differt ab esse actuali, non quidem realiter, quasi sint diversa esse realia in rebus exsistentia, sed solum differunt secundum diversas rationes sive modos, quia esse habituale convenit rei prout est intelligibilis absque cointelligentia suae actualis exsistentiae. Et ideo prima ratio sive primus modus essendi potest ab intellectu accipi absque secunda. Sicut enim differt ratio essentiae a ratione suae actualis exsistentiae absque hoc quod inter se habeant aliquam differentiam vel compositionem realem, sic differt esse habituale ab esse actuali, quia primum esse respicit solam rationem essentiam et suorum essentialium in quantum se tenent cum ratione exsistentiae actualis.’

98

chapter 2

understand the notion of any actual being, even though it is not actual, or even though this knowledge does not include its being actual. The dispositional being of essential terms is taken in this way, because when someone knows the notion of man or of animal, he truly knows the notion of being and the notion of its essential being, but in the mode of the quidity – he knows what they express and which notions they involve – even though from this [notion] he does not know that they are actual or does not believe that they will be. In another way, in the verbal mode [being] is taken to be actual, insofar as [the notion] is known and said to be actually in things.140 Even in the case that a man actually exists, the name ‘man’ only indicates his essential features but not his actual existence and the mere knowledge of the essence that allows someone to understand the term ‘man’ does not involve the knowledge of its present or future existence. By contrast, the verb ‘est’ in the assertion ‘homo est’ – ‘a man exists’ – explicitly indicates the actual existence of an individual of man. Yet, according to the most common way in which we understand and use the assertion ‘man is animal’, we do not intend to express actual existence, as we do when we utter ‘a man is an actual animal’.141 Consequently, Olivi would claim that ‘man is animal’ is true, because most commonly with this assertion we intend to speak about the dispositional being of man, unless it is somehow evident that we intend to express its actual existence, in which case it would be false. Anonymus Alani’s Solution to the Sophisma ohnea In his discussion of the sophisma ohnea, An. Al. raises objections to several of the claims that support the arguments above – the ones that show either that 140 Peter John Olivi, ql (ed. Brown, 345–346): ‘Secundo, dicitur quod idem esse actuale potest dupliciter accipi: Uno modo per modum nominalem et per modum quidditatis, sicut fit cum intelligitur ratio cuiuscumque esse actualis, quamvis ipsum non sit actu, aut quam­ vis non cogitetur ipsum esse in actu. Et hoc modo sumitur esse habituale terminorum essentialium, quia cum quis cogitat rationem hominis aut animalis vere cogitat rationem entis et rationem sui esse sibi essentialis, per modum tamen quidditatis cogitando scilicet quod dicunt et quas rationes in se includunt, quamvis ex hac non sciantur nec fore credantur esse in actu. Alio modo sumitur esse actu per modum verbalem prout scilicet scitur vel dicitur actualiter esse in rebus.’ 141 Peter John Olivi, ql (ed. Brown, 346): ‘A primo autem modo sumendi non sequitur iste secundus modus. Et ideo secundum modum intelligendi et enuntiandi non est dicere “Homo est animal” quam dicere “homo est actu animal”.’ The editor supplies , but the sense of the passage requires .

SIGNIFICATION OF CONCEPTS AND THINGS

99

‘ohnea’ is false or that it is not true without qualification (simpliciter). An. Al.’s own solution is based on two assumptions: (i) Existence is not included in the signification – the semiotic value – of a word and (ii) there is a difference between being of essence and being of existence.142 The first assumption is supported by the first chapter of Aristotle’s Perihermeneias, where Aristotle tells us that names and verbs taken in themselves do not state anything, and by the fourth chapter of the same treatise, where Aristotle says that attributes outside assertions are names, because attributes by themselves do not state existence or non-existence.143 The second assumption, An. Al. tells us, is a corollary of the first one, because if words do not signify existence, they must signify only the essence.144 Words signify essences, because they signify the same thing that their definitions signify,145 but they do not signify existence, because when the intellect understands the essence of a thing, it does not necessarily understand its existence.146 An. Al. appeals to these two assumptions in his rejection of several positions that can be compared to the ones taken by Roger Bacon, Boethius of Dacia and Peter John Olivi. It is noteworthy, though, that none of the positions he rejects correspond exactly to the positions put forth by these authors. To the position that terms are equivocal to being and non-being, which we find in Bacon’s rejection of ‘Socrates est homo’ as true without qualification, 142 An. Al., ohnea (ed. de Libera and Gazziero, 344): ‘In his autem rationibus et in aliis sequentibus accipiam duas suppositiones. Prima est quod nomen de vi vocis non significat rem esse, cuius est ipsum nomen etiam si res sua sit. Secunda est quod esse existentiae et esse essentiae non sunt idem.’ 143 An. Al., ohnea (ed. de Libera and Gazziero, 344): ‘Et ne suppositiones videantur impossibiles possunt confirmari. Prima sic. Dicit Aristoteles primo Peri hermeneias capitulo primo: “… Nomina igitur et verba similia sunt intellectui sine compositione et divisione ut “homo” vel “album”, quando non additur aliquid; neque adhuc verum vel falsum est…”… Item. In capitulo de verbo probat quod verbum “secundum se dictum aliquid significet” et subiungit “sed si est vel non est nondum significat”, nec enim significat “est” “res esse vel non esse si ipsum ‘est’ purum dixeris: ipsum quoque nihil est etc.”.’ Cf. above, Section 1.1. 144 An. Al., ohnea (ed. de Libera and Gazziero, 345–346): ‘Secunda suppositio sequitur ex ista directe, quia si nomen non significat esse existentiae eius cuius est nomen, ut ostensum est, significat autem esse essentiae, sequitur quod esse essentiae et esse existentiae eius non sunt idem.’ An. Al. is silent about whether the difference between being of essence and being of existence amounts to two different ontological statuses of things. 145 Cf. Arist., Met. vii.5.1031a12–13; cf. Auctoritates Aristotelis 1, n. 163 (ed. Hamesse, 129:80). 146 An. Al., ohnea (ed. de Libera and Gazziero, 346): ‘Quod autem significet esse essentiae eius patet quia significat idem quod sua definitio… Item. Intellectus apprehendens rei essentiam non apprehendit necessario eius existentiam.’

100

chapter 2

An. Al. opposes that this goes against Aristotle in the Categories,147 where Aristotle allegedly claims that an assertion goes from being true to being false, because of an alteration within the subject. Consequently, an assertion and its contradictory (e.g. ‘Cesar exists’ and ‘Cesar does not exist’) could be true simultaneously if for instance ‘Cesar’ in each assertion happens to signify two different things (one that exists and another that does not exist).148 Recall, however, that for Bacon the Aristotelian tenet does not hold for assertions of the kind ‘X est’, which he takes to be essential and involving the equivocation of the subject when it stands for something existing at one time and non-existing at another. The reason is that assertions with equivocal subjects do not comply with the PoB, precisely because the subject is different in the assertion and in its negation, and hence An. Al.’s argument against Bacon’s position does not seem to be conclusive. The following objection, though, is sounder: What was said first, that the name ‘man’ signifies a living man, when a man exists, but a dead man, when no man exists – [things] of which ‘man’ is said equivocally – is of little or no value. In fact, [‘man’] always signifies a living man, because it is impossible to signify or to understand man without life, because soul and life are the same. But when a man does not exist or is dead, the name ‘man’ does not signify this.149 ‘Man’ signifies univocally the living man, whether existing or non-existing, because humanity includes life in its definition as an essential property (since to be animal is to have life and to be a man is to be a rational animal). In this way, the permanence of the subject in the essential predication ‘man is an animal’ is maintained, and this regardless of whether there are men or not. From the perspective of human communication, this is certainly preferable to

147 Cf. Arist., Cat. 5.4a21–b13. 148 An. Al., ohnea (ed. de Libera and Gazziero, 349): ‘Item. ...in se falsa est. Dicit enim Aristoteles quod “oratio manens una et eadem numero mutatur a veritate in falsitatem secundum rei mutationem”. Cum ergo haec “Caesar est” sit vera Caesare existente, illa re mutata manens eadem numero erit falsa. Non autem manet eadem numero nisi enuntiet idem numero et de eo eodem numero et, si ita sit, hoc nomen “Caesar” univoce supponet in una et in alia.’ 149 An. Al., ohnea (ed. de Libera and Gazziero, 351): ‘Quod autem dicebatur primo quod hoc nomen “homo” homine existente significat hominem vivum, nullo autem homine existente significat hominem mortuum, de quibus dicitur “homo” aequivoce, modicam \aut/ nullam habet proprietatem. Significat enim semper hominem vivum, quia impossibile est hominem significari vel intelligi sine vita, quia anima et vita sunt eadem. Si autem contingat hominem non esse vel mortuum esse, tamen hoc nomen non significat.’

SIGNIFICATION OF CONCEPTS AND THINGS

101

Bacon’s cumbersome position of having a different statement at every utterance of ‘man is an animal’. We saw that the position that words lose their signification with the destruction of the things they signify cannot be fairly attributed to Bacon without some qualifications. On the other hand, however, An. Al. rejects this position as something ‘similar to madness and beyond all madness’, because anyone who hears a word that signifies something non-existing, but which she nonetheless understands, will know the word has signification for her regardless of the existence or non-existence of its content.150 The following is suggested as the cornerstone of an argument that several authors from the second half of the thirteenth century put forth in favour of the permanence of the word’s signification when things do not exist: To signify is to provoke the formation of a thought, so that whatever can be understood can also be signified, and this from the point of view of both the speaker (who can speak about whatever she can understand) and the listener (for whom any word that evokes an understanding is significative). As Michael of Marbais explains, we can evoke the understanding of something that does not exist anymore thanks to its image (phantasma), which remains in the intellect even when things are not before the senses and which is the starting point of understanding. Thanks to this image and to the action of the agent intellect, concepts can be evoked as the intellect wishes, so that the relation of signification between a word and an essence is not destroyed with the destruction of external things: It must be known diligently that when the things are destroyed, the images remain, as is evident from De anima ii. Hence, when sense-objects are gone, the images (i.e. the phantasmata) remain… Whence, since the relation of understanding remains in the intellect when the thing is destroyed, thanks to the action of the agent intellect over the images, as it was said, the relation of signification can also remain in the word after the destruction of the thing according to its existence… Whoever listens to the word ‘Cesar’, grasps some real nature and has a real concept of the thing, as if Cesar really existed. And thus this argument makes it evident that the word remains, when the significate is destroyed according to its real existence.151 150 An. Al., ohnea (ed. de Libera and Gazziero, 342): ‘…sic videtur de ista positione quod sit similis dementiae et ultra omnes dementias, quia nec laicus, nec clericus, nec demens, nec sapiens in tantum egressus est quin nomine rei praeterita, quam cognovit, si ipsum audiat, moveatur in anima sua.’ 151 Michael of Marbais, Summa (ed. Kelly, 10–11): ‘...sciendum est diligenter quod rebus corruptis remanent phantasmata, ut patet in secundo De anima. Unde abeuntibus sensibilibus remanent imaginationes et phantasiae… Et ideo cum ratio intelligendi re corrupta in

102

chapter 2

Radulphus Brito introduces further qualifications to this idea. On the one hand, he agrees that the signification of a word is determined by what is understood through the word, which remains the same whether the thing exists or not, because the image (phantasma) of the thing remains the same whether the thing exists or not.152 However, this does not amount to saying that the essence remains the same when all its instantiations are destroyed. Instead, this only means that the essence qua object of understanding, and hence qua significate – the essence without any consideration of its actual existence – remains the same. However, the essence qua essence is destroyed when all its instantiations are destroyed, because its mental existence or its existence as an object of understanding is not a proper ontological status: The significate of the utterance is not the same when the thing exists and when it does not exist, because the significate of the utterance is the quidity of the thing, i.e. its essence. But [the essence] is not the same when the thing exists or when it does not exist, because when the thing does not exist the essence is destroyed…hence, when the thing is destroyed, the essence of the thing does not remain, as some people say; for it does not remain in the soul, because being in the soul is…[being] in a certain respect…nor it remains outside the soul…nevertheless, the significate remains the same, because when I say ‘the utterance signifies’, I mean the significate, not according to what it is absolutely, but insofar as it is a significate…153 intellectu remaneat per recursum intellectus agentis ad ipsa phantasmata ut dictum est, ipsa ratio significandi etiam remanere potest in dictione post corruptionem ipsius rei secundum eius existere…quicumque hanc vocem “Caesar” audit, aliquam veram naturam apprehendit et habet verum conceptum rei ac si ipse Caesar realiter existeret. Et sic patet manifeste per rationem quod dictio remanet, corrupto significato secundum eius reale existere.’ 152 Radulphus Brito, In Perih. (ed. Venice, 144–145): ‘Ad istam quaestionem dico duo, primo quod vox idem significat re existente et non existente; secundo dico quod quantum ad significatum vocis non est idem re existente et non existente. Primum declaratur sic, quia illud quod per vocem intelligitur per vocem significatur. Modo idem intelligitur per vocem sive res sit sive non sit; ergo idem significatur per vocem sive res sit sive non sit. Maior patet quia significare est intellectum constituere… Minor probatur, quia ubicumque manet eadem ratio intelligendi manet idem fantasma in fantasia sive sit res sive non sit; modo ex eodem fantasmate sumitur eadem ratio intelligendi; ergo eadem ratio manet sive res sit sive non sit.’ 153 Radulphus Brito, In Perih. (ed. Venice, 145): ‘Secundum declaratur sic secundum quod illud quod est significatum per vocem non sit idem re existente et non existente, quia illud quod est significatum per vocem est quidditas rei et essentia illa; autem non est

SIGNIFICATION OF CONCEPTS AND THINGS

103

It is then possible to make statements about things that do not exist anymore, but that are still known, and those statements make sense. However, this does not entail that an essence that only exists as an object of understanding is real; for essences are real only when they are individuated in some material compound. Now, as to the equivocation of ‘being’ with respect to dispositional being and actual being, An. Al. replies that ‘est’ indicates either actual being (when it is a second adjacent) or dispositional being (when it is a third adjacent), but never both at the same time in the same assertion. ‘Est’ as a third adjacent in essential predication always indicates a necessary relation of inherence between the terms – an essential disposition to occur together: Furthermore, some people establish a distinction concerning ‘ohnea’, because the verb ‘est’ can express dispositional being or actual being. If it expresses dispositional being, they say that it is true, if actual being, false… I believe that ‘est’ can attribute dispositional being and actual being, but never in the same proposition. Whenever it is predicated as a second adjacent it attributes actual being, but whenever it is predicated as a third adjacent with an essential predicate it attributes dispositional being… Ancient [authors] said that dispositional being expresses the disposition or inherence of the extremes, and not their existence; but to express actual being is to express their existence.154 Thus, while ‘Socrates is’ has an implicit claim of existence, ‘man is an animal’ has none, because the latter involves essential predication and, consequently, ‘is’ here only indicates that the notions of man and of animal are necessarily eadem re existente et non existente, quia re non existente corrumpitur rei essentia…unde re corrupta non manet essentia rei ut quidam dicunt; non enim manet in anima quia esse in anima est…secundum quid…nec extra animam manet…et tamen illud quod est significatum manet idem, quia quando dico “vox significat” dico significatum non secundum illud quod est absolute, sed ut significatum est…’. 154 An. Al., ohnea (ed. de Libera and Gazziero, 358–359): ‘Praeterea alii distinguunt istam “omnis homo de necessitate est animal”, eo quod hoc verbum “est” potest dicere esse habituale vel esse actuale. Si dicat esse habituale dicunt quod est vera, si actuale falsa… credo quod possit copulare esse habituale vel actuale, nunquam tamen in eadem propositione. Sed quandocumque praedicatur secundum adiacens copulat esse actuale, quandocumque autem praedicatur tertio adiacens cum praedicato essentiali copulat esse habituale… Sed esse habituale appellabant antiqui dicere habitudinem vel inhaerentiam extremorum, et non existentiam eorum. Dicere autem esse actuale est dicere eorum existentiam.’

104

chapter 2

related so as to always have the notion of animal implied by the notion of man, whether a man exists or not. The same distinction between ‘is’ as a second adjacent and as a third adjacent could apply to the claim that the truth of ‘man is animal’ requires the existence of at least one man, an objection that would be more incisive than the one An. Al. actually puts forth. As a matter of fact, An. Al. puts forth a weak objection that is based on his supposition that names do not assert existence, so that ‘homo’ in ‘homo est animal’ does not assert the existence of its significate: Some people who deny [‘ohnea’] say that the name has the same signification, whether the thing exists or not; but they say that the name is imposed only on things that are present, just as it was formerly said. Hence, the sense of ‘man is animal’ is always ‘the man, which exists, is animal’. Hence, when no man exists, [the assertion] is false, not because of the principal composition, but because of its implication that the man exists, which is false.155 It is noteworthy, however, that Aristotle’s claim in Perihermeneias 1 and 3 that names and attributes do not state existence – indeed they do not state anything at all – only holds for names and attributes outside of an assertoric context, and hence An. Al.’s first supposition does not apply to words that are the terms of an assertion. Boethius of Dacia, for instance, would agree that ‘man’ tout court does not assert the existence of man, even though he claims that the truth of ‘man is animal’ does assert the existence of at least one man, because the claim of existence lies in the statement involved by the composition of the terms.156 After having raised all these objections,157 An. Al. presents his own solution to the sophisma ohnea: ‘ohnea’ is true without qualification and its proof by 155 An. Al., ohnea (ed. de Libera and Gazziero, 351): ‘Alii autem negantes eam (i.e. ohnea) dicunt quod nomen significat, sive res sit sive non sit, et idem. Dicunt tamen quod nomen imponitur tantum praesentibus, sicut dixerunt priores. Unde semper est sensus huius ‘homo est animal’ iste ‘homo, qui est, est ’. Unde nullo homine existente haec erit falsa, non propter principalem compositionem, sed propter implicationem, scilicet quia implicat hominem esse, quod falsum est.’ 156 I am inclined to think that Aristotle would agree with Boethius of Dacia on this point; see Crivelli, Aristotle on Truth, Ch. 5.2, for a defense of this position. For the opposite position, see P. Thom, ‘On the Pervasiveness of Being’, in V. Caston and D.W. Graham (eds.), Presocratic Philosophy: Essays in Honour of Alexander Mourelatos (Aldershot: Ashgate, 2002), pp. 293–301, at pp. 298–300. 157 And many more that are not relevant for my purposes.

SIGNIFICATION OF CONCEPTS AND THINGS

105

means of Argument 3158 is granted.159 Recall, though, that Argument 3 presupposes the existence of at least one particular instance of ‘man’, which is what the sophisma denies by hypothesis. A sounder argument would have been to say that essential predication, e.g. in definitions, is an epistemic device that states a coherence between notions rather than a blatant statement about an actual event (even though a good deal of knowledge of the external world is indeed needed in order to arrive at the formulation of a definition). However, although An. Al. flirts here and there with this idea, he never formulates it in a clear and open manner. *** The medieval quaestiones we have analysed in the last two sections illustrate well the extent to which the thirteenth-century notion of signification is used in order to resolve a number of linguistic problems that, albeit related, are fundamentally different, i.e. the problem of the efficacy of human communication, the problem of instituting linguistic signs and the problem of propositional verification. These three linguistic concerns came down to the thirteenth century via some of the most influential authoritative texts of the period, namely Aristotle’s Organon and Boethius’ logical corpus. In these texts, these concerns 158 See above pp. 85–86. 159 When An. Al. grants Argument 3, he rejects in passing the multiplicity of forms, and claims that, contrary to the species that adds a difference to the genus, individuals and their species have exactly the same form. Since names are imposed thanks to the understanding of a form, the names of e.g. human individuals – proper names – are imposed thanks to the understanding of the form of man, so that ‘Socrates’ and ‘man’ indicate exactly the same essence – the essence of man – although in different ways. Consequently, to predicate ‘man’ of ‘Socrates’ is to predicate the same thing of the same thing, so that ‘Socrates is a man’ is as necessary as ‘homo est homo’ and even more necessary than ‘homo est animal’; see An. Al., ohnea (ed. de Libera and Gazziero, 361–362): ‘Et iterum individuum non addit formam aliquam supra speciem vel differentiam nec habet aliam formam a forma speciei et hoc est quod scribitur quod “species est totum esse individuorum”; genus autem non est totum esse speciei, quia species addit differentiam et esse supra genus, quae differentia non est in genere nisi potestate solum. Nomen autem imponitur a forma, ergo cum individuum non habeat aliam formam a forma speciei, ut dictum est, ab eadem forma et ad eandem imponitur nomen speciei et individui; idem ergo significant hoc nomen “Sor” et hoc nomen “homo” modo solum alio, qui modus non est accidentalis… ergo praedicare hominem de Sorte est praedicare idem de se, et qui negat talem praedicationem idem affirmat et negat, sicut visum fuit prius. Haec ergo “Sor est homo” non solum est necessaria, immo magis necessaria quam ista “homo est animal”, et aeque necessaria ut ista “homo est homo”.’

106

chapter 2

were already entangled with each other and closely associated with the notion of signification. It is no surprise, then, that most authors from the thirteenth century amalgamate these linguistic problems and jump from one to the other in their treatment of these two quaestiones directly concerned with the signification of words. Notwithstanding, in dealing with the issue of whether terms lose their significate with the destruction of things, Boethius of Dacia, like Radulphus Brito, dares to introduce a notion of signification that accounts for the semiotic content of a word, so that other notions that are not signification but depend on it, can explain further linguistic concerns. More precisely, in the last two sections we have seen how Brito and Boethius of Dacia exclude the notion of signification from the problem of propositional verification. The attempt to relegate the notion of signification to the semiotic level also surfaces in the following and last discussion. As we shall see, towards the end of the century masters of Arts including Boethius of Dacia and Radulphus Brito no longer directly use the notion of signification in their deductions of grammatical categories (partes orationis) and accounts of grammaticality, i.e. the central tasks of the medieval commentator on Priscian’s Institutiones grammaticae.

PART 2 Signification in Logic and in Grammar



chapter 3

Names and Verbs in Priscian and in Aristotle Priscianus Caesariensis’ Institutiones grammaticae (ca. 500 ad)1 was undoubtedly the main source for the study of theoretical grammar in the late Middle Ages.2 His enquiry shares with Boethius’ commentaries on and translations of Aristotle’s logical treatises crucial objects of analysis, such as utterance (vox), name (nomen), verb (verbum) and sentence (oratio),3 and uses the explanatory power of signification. Nonetheless, the role this notion plays in the grammatical analyses of names, verbs and sentences differs in important respects from the role it plays in the same analyses in the Aristotelian logical tradition. We shall see how these divergences were also at the origin of important discussions about signification in the thirteenth-century faculties of Arts. 3.1

Priscian on the Constitution of Parts of Speech and Sentences

Priscian begins his explanation of the constituent parts of speech – words of a certain grammatical category – by defining the utterance (vox), i.e. the material 1 Priscian, Institutiones grammaticae (= ig), ed. M. Hertz (Leipzig: Teubner, 1855–1859) (gl ii–iii). For Priscian’s life and work, see the first three chapters of M. Baratin et al. (eds.), Priscien. Transmission et refondation de la grammaire de l’antiquité aux modernes (Turnhout: Brepols, 2009) (Studia Artistarum 21), pp. ix–xviii and 1–34. For a discussion of Priscian’s philosophical sources, see S. Ebbesen, ‘Priscian and the Philosophers’, in Baratin et al. (eds.), Priscien, pp. 85–108. 2 For the reception of Priscian’s grammar in the 12th and 13th centuries, see Baratin et al. (eds.), Priscian, Ch. 6.2. 3 Priscian’s Institutiones was traditionally divided into two parts – Priscianus maior (books i– xvi) and Priscianus minor (Books xvii and xviii). Priscianus maior deals with Latin morphology and follows the traditional order of Latin grammars in that it introduces and discusses the notions of utterance (vox), letter (littera), syllable (syllaba), part of speech (dictio) and sentence (oratio) in a progressive way, it proceeds to the division of the parts of speech into eight parts, and it gives extensive accounts of each of these parts. In contrast, in De constructione or Priscianus minor, which is a Latin adaptation of books one to three of Apollonius Dyscolus’ On Syntax, Priscian introduces the notion of grammaticality and illustrates in detail the possible constructions of different parts of speech that bring about a grammatical sentence (oratio perfecta). The structure of Priscian’s Institutiones is basically that of Donatus’ Ars Maior; see V. Law, The History of Linguistics in Europe. From Plato to 1600 (Cambridge: cup, 2003), p. 88. For a thorough description of the literary genre and the composition of the Institutiones, see M. de Nonno, ‘Ars Prisciani Caesariensis: problemi di tipologia e di composizione’, in Baratin et al. (eds.), Priscien, pp. 249–278. © koninklijke brill nv, leiden, ���� | doi 10.1163/9789004300132_005

110

chapter 3

part of oral expressions. Philosophers (the stoics4), Priscian explains, define an utterance as a ‘very subtle air that is hit or the proper sense-object of the ear, i.e. whose proper feature is to affect the ear’.5 This definition involves two aspects: The first – the substantial aspect – states what an utterance is in itself (i.e. air that is hit). The second – the notional aspect – states its proper feature (i.e. its capacity to affect the organ of the sense of hearing).6 Priscian’s focus on the proper feature of utterances – their capacity to affect the ear – can be explained by its relevance when one is to account for the utterances’ place of privilege as vehicles of linguistic communication: Utterances are of interest because they transmit information (which depends on their ability to affect the ears) and not because of their material constitution. Priscian further provides us with a twofold, but not mutually exclusive, division of utterances into articulate and inarticulate, and literate and illiterate. An articulate utterance entails an affection in the mind of the speaker and its content is comprehensible (supposedly by a listener). A literate utterance is one that can be written. Inarticulate and illiterate utterances are defined negatively as not complying with these two conditions.7 Hence, an articulate and literate 4 Priscian’s grammar relies heavily on the Greek grammatical tradition, and most notably on Apollonius Dyscolus’ On Syntax [Apollonius Dyscolus, De constructione libri quattuor, ed. G. Uhlig (Leipzig: Teubner, 1910) (gg ii.2); for a translation, see The Syntax of Apollonius Dyscolus. Translated, and with commentary by F.W. Householder (Amsterdam: John Benjamins, 1981) (Studies in the History of Linguistics 23)]. Apollonius’ stoic-oriented grammar endowed Priscian’s with a strong focus on semantic notions that is quite unique in late-ancient Latin grammar. Thus, Priscian’s grammar carries stoic notions into the thirteenth-century linguistic discussions. 5 Priscian, ig i.1 (ed. Hertz, 5:1–2): ‘Philosophi definiunt vocem esse aerem tenuissimum ictum vel suum sensibile aurium, id est quod proprie auribus accidit.’ 6 Priscian, ig i.1 (ed. Hertz, 5:2–4): ‘Et est prior definitio a substantia sumpta, altera vero a notione…hoc est ab accidentibus. Accidit enim voci auditus, quantum in ipsa est.’ This twofold definition can be traced back to the stoic Diogenes of Babylon; see Diogenes Laertius, Vitae philosophorum vii.55, 2 vols., ed. H.S. Long (Oxford: Clarendon Press, 1964). J.B. Gourinat tells us that the stoic division of definitions is related to the distinction between a pre-notion (prolepseis) and a notion (ennoia), in that the pre-notion, just as the substantial definition, identifies the sense-object that is the starting-point of an enquiry, whereas the notion, just as the accidental definition, provides the proper feature with which the enquiry in question is concerned; cf. J.B. Gourinat, La dialectique des stoïciens (Paris: Vrin, 2000), p. 53. In this case, the substantial definition would provide Priscian’s enquiry with the sense-object that is its starting point, and the notional (or accidental) definition, would provide it with the proper feature that is relevant in a linguistic enquiry, namely the capacity of being heard. 7 Priscian, ig i.i.1 (ed. Hertz, 5:6–9): ‘Articulata est, quae coartata, hoc est copulata cum aliquo sensu mentis eius, qui loquitur, profertur. Inarticulata est contraria, quae a nullo affectu profiscitur mentis. Literata est, quae scribi potest, illiterata, quae scribi non potest.’

Names And Verbs In Priscian And In Aristotle

111

utterance (e.g. ‘arma virumque cano’) can both be written down and understood by a listener.8 A word (dictio) qua part of speech (which is an articulate and literate utterance) is the minimal part of a sentence as regards its intelligible content – the minimal part that contributes to the intelligibility of the sentence as a significative expression.9 Finally, a sentence (oratio) is defined as a grammatical arrangement (congrua ordinatio) of parts of speech presenting us with a complete sense.10 At this point, Priscian stresses that sentences and their parts are presented in remarkably different ways by the dialectici (supposedly neoplatonic and peripatetic philosophers) and the stoics, thereby introducing the notion of signification as the distinctive criterion of his division of the parts of speech. Different approaches to the signification of sentences and their parts thus explain the varied treatments they receive in the different linguistic traditions.

The Role of the Notion of Signification in the Division and Order of the Parts of Speech We have already observed how Aristotle regards only names and attributes as necessary for the production of assertions susceptible of truth or falsity.11 Priscian briefly describes a similar position that he ascribes to the dialectici: According to logicians, there are two parts of the sentence, the name and the verb, because only the union of these [parts] produces a complete

8

There are also articulate illiterate utterances (e.g. human cries of pain), literate inarticulate utterances (e.g. ‘coax’) and inarticulate illiterate utterances (e.g. a rattling); cf. Priscian, ig i.i.1–2 (ed. Hertz, 5:9–6:2). Note that, contrary to Boethius who follows a peripatetic tradition, in Priscian’s stoic background articulate utterances are not the ones that can be written. For Priscian, for an utterance to be articulate it merely has to indicate an affection of the soul, whereas in the peripatetic tradition to be articulate is to be made of parts – the very reason why articulate utterances can be written with letters. 9 Priscian, ig ii.iii.14 (ed. Hertz, 53:8–9): ‘Dictio est pars minima orationis constructae, id est in ordine compositae: pars autem quantum ad totum intelligendum, id est ad totius sensus intellectum…’. Note that Priscian talks here of the dictio as the minimal part of a sentence, but evidently he means that the dictio qua part of speech is the minimal part of a sentence. 10 Priscian, ig ii.iv.15 (ed. Hertz, 53:27–28): ‘Oratio est ordinatio dictionum congrua, sententiam perfectam demonstrans.’ 11 See above, Part 1, Section 1.1.

112

chapter 3

sentence, and they called the other parts ‘syncategoremes’, i.e. consignificative items.12 Stoic logic, however, considers five parts – proper name (nomen), common noun (appellatio), verb (verbum), pronoun or article (pronomen sive articulus) and conjunction (coniunctio).13 After noticing the discrepancies in the number of parts of speech in these two philosophical traditions, Priscian concludes that in order to rightly deduce the number of parts, one ought to look at the different significations of the words that produce a complete sentence: Whence, the parts of speech cannot be distinguished from one another, unless we consider the proper feature of each [part]’s signification.14 Signification is thus introduced as the distinctive criterion of his division of the parts of speech. To be sure, Apollonius had already advanced this approach in his On syntax, as his own modification of the stoic division of the parts of speech. Let us examine this in some detail. Stoic dialectics involves two aspects – the study of significative utterances and the study of their significates.15 However, at least in early stoic dialectics, these two aspects depend on each other, so it is not possible to separate the syntactic analysis from the enquiry into the content of language and its ability to express something true or false.16

12 Priscian, ig ii.iv.15 (ed. Hertz, 54:5–7): ‘Partes igitur orationis sunt secundum dialecticos duae, nomen et verbum, quia hae solae etiam per se coniunctae plenam faciunt orationem, alias autem partes “syncategoremata”, hoc est consignificantia, appellabant.’ It is uncertain to whom exactly Priscian refers when he talks about the dialectici, but we find this analysis of the statement into onomata and rhêmata as early as in Plato’s Sophist and in Aristotle’s Perihermeneias. 13 Cf. Priscian, ig ii.iv.16 (ed. Hertz, 54:8–9). 14 Priscian, ig ii.iv.18 (ed. Hertz, 55:4–5): ‘Igitur non aliter possunt discerni a se partes orationis, nisi uniuscuiusque proprietates significationum attendamus.’ 15 Cf. e.g. Diogenes Laertius, vp vii.43–44. For introductory studies to stoic philosophy, see J. Barnes and J.B. Gourinat, Lire les stoïciens (Paris: puf, 2009); Long and Sedley, The Hellenistic Philosophers. Vol. 1 (Cambridge: cup, 1987); and J. Christensen, An Essay on the Unity of Stoic Philosophy (Copenhagen: Museum Tusculanum Press, 2012); for an exhaustive study of stoic dialectics, see Gourinat, La dialectique des stoïciens. 16 The detachment of grammar from dialectics seems to be at least as old as Diogenes of Babylon; see Gourinat, La dialectique des stoïciens, p. 107.

Names And Verbs In Priscian And In Aristotle

113

Speech (λόγος) is significative insofar as it indicates a content that can be analysed into: (i) the incorporeal thing that is potentially indicated by a sentence – a sayable (λεκτόν or πρα᷑γμα17); (ii) this very same incorporeal thing insofar as it is actually signified by a sentence – a significate (σημαινόμενον) and (iii) the corporeal thing that is the bearer of a name or of a sayable (τυγχάνον).18 Complete sayables – the ones indicated by complete sentences (e.g. assertions, questions and commands) – are compound, i.e. they have a subject and a predicate.19 Such sayables, in turn, somehow reproduce the ontological structure of the world, which is ultimately constituted of material bodies and their qualities, and their actions and undergoing of actions.20 The proper name, whose content indicates the proper feature of a body, and the common name, whose content indicates a feature shared by other bodies, usually play the role of the subject in complete sentences. On the other hand, verbs indicate incomplete sayables – predicates – that are related to actions and undergoing of actions.21 Finally, articles or relative pronouns and conjunctions have, respectively, the functions of narrowing the extension of common names and of attaching the parts of speech.22 Now, the completeness of sentences is independent of their truth or falsity, which somehow explains how it is possible to focus only on the syntactic part of the stoic analysis, leaving aside the semantic considerations. This detachment 17

18

19 20

21

22

It is called a λεκτόν insofar as it is a sayable and a πρα᷑γμα insofar as it is a state of affairs, although not necessarily an existing state of affairs; see Gourinat, La dialectique des stoïciens, p. 117. See Gourinat, La dialectique des stoïciens, ch. 1 and M. Baratin and F. Desbordes, L’analyse linguistique dans l’antiquité classique. 1. Les théories (Paris: Editions Klincksieck, 1981), pp. 26–34. Cf. e.g. Diogenes Laertius, vp vii.64. Actions and undergoing of actions are not considered corporeal things, but incorporeal things, and hence an incomplete sayable can be considered a sort of mental counterpart of an action or undergoing of an action, without considering its being actually performed by any existing body; see Gourinat, La dialectique des stoïciens, p. 118–119. Note that while it is possible to have incomplete sayables that are predicates, the same does not seem to be the case for nominal parts, which seem to occur always as parts of complete sayables; M. Frede [in his article ‘The Origins of Traditional Grammar’, in Essays in Ancient Philosophy (Minessota: Minessota Archive Editions, 1987), at p. 347] claims that nominal parts are never taken as incomplete sayables. This fact would reflect the emphasis that stoic philosophy puts on actions as opposed to the emphasis that Aristotelian philosophy puts on substances. The question remains whether there are nominal incomplete sayables or whether their absence results from an opposition between Aristotelian causality – things being the cause of other things – and stoic causality – bodies being the cause of actions. Cf. Diogenes Laertius, vp vii.58.

114

chapter 3

of grammar from stoic dialectics is already fully achieved in Apollonius’ On Syntax. Apollonius passes to Priscian a stoic-oriented hierarchy of the parts of speech that depends on the extent to which each part contributes to the completeness of a sentence. When giving his account of the constitution of syllables, parts of speech and sentences, Apollonius underscores the importance of the order in which the different parts of speech are introduced;23 for just as some letters are pronounced by themselves (vowels) and some need to be attached to vowels in order to be pronounced (consonants), in the same way some words are significative by themselves (e.g. names, verbs and pronouns) and some are not (e.g. articles, prepositions and conjunctions). Based on this observation, he introduces a first division of words into words significative by themselves and consignificative words.24 Among the significative parts of speech, the name is prior to the verb because it indicates a thing that is performing or undergoing an action and that is thus naturally prior to this performing or undergoing.25 Pronouns follow 23 Apollonius, On syntax 13: ‘But…it’s important to compare and justify the ordering of the parts of speech. Perhaps someone, flaunting his own ignorance of such matters, may foolishly urge that there’s no need to bother with such investigations, suggesting that these matters have been arranged arbitrarily and by chance. But such people can also propose that, in general, nothing is ordered and there are no errors of ordering; but this would be quite idiotic.’ (tr. Householder, 23). [κατεπει᷑γον μέντοι ἐστὶν ἕνεκα τω᷑ ν προεκκειμένων λόγων ἀντιπαραβάλλεσθαι τὴν τάξιν τω᷑ ν μερω᷑ ν του᷑ λόγου. ἴσως τινὲς ἀπειρότερον ἀναστρέφοντες περὶ τὰ τοιαυ᷑τα τὰς ἰδίας ἀπειρίας παραμυθου᷑νται, ὡς οὐ δέον περὶ τὰς τοιαύτας ζητήσεις καταγίνεσθαι, ὑπολαβόντες τὰ τοιαυ᷑τα κατὰ τύχην τεθεματίσθαι. (ed. Uhlig, 16:4–8)]. 24 Apollonius, On syntax 12: ‘Furthermore, just as some phonemes are vowels, which are complete sounds even in isolation, and other consonants, whose pronunciation is not possible without vowels, so also one can consider words of two kinds. Some words are, like vowels, independently speakable, e.g. verbs, nouns, pronouns, adverbs, when they can be applied to actions in the situational context… Other words resemble consonants, and just as they require vowels, so these require the presence of some of the aforesaid parts of the speech; this is the case with prepositions, articles and conjunctions.’ (tr. Householder, 22). [Ἔτι ὃν τρόπον τω᷑ ν στοιχείων ἃ μέν ἐστιν φωνήεντα, ἃ καθ’ ἑαυτὰ φωνὴν ἀποτελει᷑, ἃ δὲ σύμφωνα, ἅπερ ἄνευ τω᷑ ν φωνηέντων οὐκ ἔχει ῥητὴν τὴν ἐκφώνησιν, τὸν αὐτὸν τρόπον ἔστιν ἐπινοη᷑ σαι κἀπὶ τω᷑ ν λέξεων. αἱ μὲν γὰρ αὐτω᷑ ν τρόπον τινὰ τω᷑ ν φωνηέντων ῥηταί εἰσι, καθάπερ ἐπὶ τω᷑ ν ῥημάτων ἔστιν ἐπινοη᷑ σαι, ὀνομάτων, ἀντωνυμιω᷑ ν, ἐπιρρημάτων, ἅπερ ἐπιλέγεται ται᷑ς γινομέναις ἐνεργείαις…αἱ δὲ ὡσπερεὶ σύμφωνα ἀναμένουσι τὰ φωνήεντα, τουτέστιν τὰ προκατειλεγμένα τω᷑ ν μερω᷑ ν του᷑ λόγου, οὐ δυνάμεναι κατ’ ἰδίαν ῥηταὶ εἰν� αι, καθάπερ ἐπὶ τω᷑ ν προθέσεων, τω᷑ ν ἄρθρων, τω᷑ ν συνδέσμων· (ed. Uhlig, 13:1–14:1)]. 25 Apollonius, On syntax 16: ‘The noun necessarily precedes the verb, since influencing and being influenced are properties of physical things, and things are what nouns apply to,

Names And Verbs In Priscian And In Aristotle

115

verbs because they ‘were invented for the purpose of accompanying verbs’26 (of the first and second persons, that is); for names indicate third persons, and therefore they can only be attached to the third person of a verb. Thereafter follow the non-significative parts of speech in the order article, preposition and conjunction. On similar grounds, Priscian introduces the parts of speech in the order: name, verb and participle, pronoun, preposition, adverb and conjunction. This order, just as in Apollonius, takes into account the extent to which the content of a part of speech contributes to the completeness of the sentence. Thus, since things and their properties, and their actions and undergoing of actions, contribute to the greatest extent to the sense of a sentence, names and verbs are introduced in the first and second place. Thereafter come pronouns that, Priscian tells us, can take the place of proper names and definite persons.27 Finally come adverbs, prepositions and conjunctions, which are not significative by themselves but contribute to the signification of the parts of speech to which they are attached and, thereby, to the sense of the whole sentence. Let us now take a look at the particular cases of the name and the verb. Names and Verbs and Their Construction according to Priscian When giving the specific feature of the signification of names, Priscian claims that:

and to things belong the special features of verbs, namely doing and experiencing.’ (tr. Householder, 25). [Καὶ του᷑ ῥήματος δὲ ἀναγκαίως πρόκειται τὸ ὄνομα, ἐπεὶ τὸ διατιθέναι καὶ τὸ διατίθεσθαι σώματος ἴδιον, τοι᷑ς δὲ σώμασιν ἐπίκειται ἡ θέσις τω᷑ ν ὀνομάτων, ἐξ ὡ� ν ἡ ἰδιότης του᷑ ῥήματος, λέγω τὴν ἐνέργειαν καὶ τὸ πάθος. (ed. Uhlig, 18:5–8)]. 26 Apollonius, On syntax 19 (tr. Householder, 25): ‘Someone could, I think, quite reasonably object: “why on earth shouldn’t the next place after the noun be filled by the one wordtype which may replace it syntactically, i.e. the pronoun, since it, too, by substitution for a noun, may combine with a verb to make the sentence complete?” On this matter the clearest evidence is this, that pronouns were invented for the purpose of accompanying verbs.’ [Οὐκ ἀλόγως δὲ ἐκείνῳ οἰμ� αί τις ἐπιστήσει, τί δή ποτε οὐχὶ μετὰ τὴν του᷑ ὀνόματος θέσιν τὸ ἀντὶ τούτου μόριον παραλαμβανόμενον ὑπετάγη, λέγω δὴ τὴν ἀντωνυμίαν, εἴγε κατὰ ἀμοιβὴν του᷑ ὀνόματος πάλιν σὺν τῳ� ῥήματι συνέχει τὸν λόγον. Περὶ οὑ� ἂν προφανὴς ἀπόδειξις γένοιτο ἥδε, ὡς ἕνεκα τη᷑ ς τω᷑ ν ῥημάτων συνόδου ἐπενοήθησαν αἱ ἀντωνυμίαι. (ed. Uhlig, 20:1–6)]. 27 It is noteworthy that interrogative pronouns, e.g. ‘who’, are not considered pronouns but names; for, even if they share the same declination with other pronouns, a word is not a part of speech because of its declination, but because of the specific feature of its signification; cf. Priscian, ig ii.iv.18–19 (ed. Hertz, 55:13–28).

116

chapter 3

It is proper to the name to signify a substance and a quality.28 Here ‘substance’ and ‘quality’ are not to be taken as Aristotelian categories, but as a bearer and its essential or accidental features. Some pages later, Priscian defines names in this way: The name is a part of speech that assigns an individual or common qua­ lity to any of the things or of the bodies that are subjects [of such a quality].29 Although this definition follows in the footsteps of the Apollonian tradition,30 Priscian introduces a terminology that is not attested in Apollonius; for, instead of mentioning bodies, Priscian talks about substances (substantiae), which he then paraphrases as ‘the things or the bodies that are subjects of qualities’, where ‘things’, as opposed to bodies, supposedly stands for incorporeal things.31 Thus, when Priscian says that names signify qualities, he means qualities inhering in substances, reading here ‘substance’ as the bearer where accidental or essential qualities inhere.32 28 Priscian, ig ii.iv.18 (ed. Hertz, 55:6–7): ‘Proprium est nominis substantiam et qualitatem significare.’ 29 Priscian, ig ii.v.22 (ed. Hertz, 55:29–56:1): ‘Nomen est pars orationis, quae unicuique subiectorum corporum seu rerum communem vel propriam qualitatem distribuit.’ 30 Apollonius’ definition is transmitted in scholia: ‘ὄνομά ἐστι μέρος λόγου πτωτικὸν ἑκάστῳ τω᷑ ν ὑποκειμένων σωμάτων ἢ πραγμάτων κοινὴν ἢ ἰδίαν ποιότητα ἀπονέμον κοινὴν μὲν οἰο� ν ἄνθρωπος ἵππος ἰδικὴν δὲ οἰο� ν γραμματικὴ Σωκράτης’. (gg i.3, 258:29–32) See A. Luhtala, Grammar and Philosophy in Late Antiquity. A study of Priscian’s sources (Amsterdam: John Benjamins, 2005), p. 89. 31 See also ig ii.v.26 (ed. Hertz, 59:10–13) [where Priscian divides names into corporeal and incorporeal]: ‘Sunt enim quaedam corporalia in appellativis, ut “homo”, sunt etiam in propriis, ut “Terentius”, alia incorporalia in appellativis, ut “virtus”, in propriis, ut “Pudicitia”.’ Note that this paraphrase is coherent with Apollonius’ definition. 32 Priscian’s use of the terms ‘substance’ and ‘quality’ is not entirely coherent. In fact, in his description of common names [Priscian, ig ii.v.24 (ed. Hertz, 58:14–18): ‘…appellativum naturaliter commune est multorum, quos eadem substantia sive qualitas vel quantitas generalis specialisve iungit: generalis, ut “animal”, “corpus”, “virtus”; specialis, ut “homo”, “lapis”, “grammaticus”, “albus”, “niger”, “grandis”, “brevis”.’], Priscian introduces a difference between substance, quality and quantity that is supposed to account for the difference between generic nouns and specific terms (e.g. ‘man’, ‘white’ and ‘big’). But this use of ‘substance’, ‘quality’ and ‘quantity’ seems to be at odds with his former description of names according to which both ‘man’ and ‘white’ indicate substances with qualities. Moreover, he tells us some paragraphs later that adjectives are a sort of common name that indicate a specific

Names And Verbs In Priscian And In Aristotle

117

The verb, in turn, is defined as ‘a part of speech with tenses (temporibus) and modes, without case, significant of action and undergoing of action’.33 Contrary to Aristotle’s attributes, which are specifically that which is said of a subject, without any further precision of a grammatical category, Priscian’s verb is already an achieved grammatical category that is the result of a very complex development of the stoic treatment of predicates. In Priscian’s account, the features (accidentia) of verbs include their genus vel significatio (active or passive depending on whether the verb indicates an action or the undergoing of an action) as well as their tense, mode and species (depending on whether the verb is of first imposition or derivative).34 Finally, concerning the construction of the parts of speech, Priscian gives a place of privilege to names and verbs. This place of privilege is ensured by a test that we also find in Apollonius: If the name or the verb is removed from a sentence, the sentence is incomplete.35

33 34

35

quality or quantity, and that are attached to other names, generic, specific or proper, that signify substances [Priscian, ig ii.v.25 (ed. Hertz, 58:19–22): ‘Haec enim quoque, quae a qualitate vel quantitate sumuntur speciali, id est adiectiva, naturaliter communia sunt multorum; adiectiva autem ideo vocantur, quod aliis appellativis, quae substantiam significant, vel etiam propriis adici solent ad manifestandam eorum qualitatem vel quantitatem’.] Thus he separates substantive nouns (e.g. ‘man’, ‘virtue’ and ‘body’), which indicate generic substances, qualities and quantities respectively, from adjectives (e.g. ‘white’ and ‘big’), which indicate specific qualities and quantities respectively. As Marc Baratin has rightly pointed out, two senses of substance and quality are certainly at stake here. On the one hand, substance1 seems to be equivalent to the bearer of a quality in the sense of quality1, i.e. any feature by virtue of which a qualified thing x can be named by either proper or common names (all the sub-classes of common names included). On the other hand, substance2 seems to fall under the notion of quality1, along with quantity and quality2. Thus, general substance2, general quality2 and general quantity are supposed to account respectively for the difference between generic nouns such as ‘animal’, ‘virtue’ and ‘body’, which are actually examples of substance, quality and quantity in Aristotle’s Categories. Specific quantity and quality2, in turn, are supposed to separate adjectives from other types of common names and to establish some sort of division inside the adjectives themselves. Thus, specific quality accounts for the difference between adjectives such as ‘white’, ‘grammarian’ and ‘prudent’, in opposition to adjectives such as ‘big’, which indicate a specific quantity. For a similar analysis, see M. Baratin, La naissance de la syntax à Rome (Paris: Editions de minuit, 1989), pp. 387–407. Cf. Priscian, ig viii.i.1 (ed. Hertz, 368:1–2): ‘Verbum est pars orationis cum temporibus et modis, sine casu, agendi vel patiendi significativum.’ For a description of the most important stages of this development, see M. Baratin, ‘Sources philosophiques de Priscian: classement stoïcien des prédicats’, in Baratin et al. (eds.), Priscien, pp. 139–150. Cf. Priscian, ig xvii.ii.12 (ed. Hertz, 116:5–11): ‘Sicut igitur apta ordinatione perfecta redditur oratio, sic ordinatione apta traditae sunt a doctissimis artium scriptoribus partes orationis,

118

chapter 3

Hence, just as in Aristotle, names and verbs seem to be the only necessary parts for the production of a complete sentence. However, this would be a misleading parallel because in Aristotle’s analysis a name and an attribute are the elements necessary to produce an assertion susceptible of truth or falsity, and thus Aristotle’s attributes involve a fundamental assertive aspect that is not entailed by Priscian’s verbs. Since the possibility of producing assertions susceptible of truth or falsity is the essential role of Aristotle’s names and attributes, Aristotle’s analysis focuses on the features that allow this possibility; features of which the univocity of names and attributes and the assertiveness of attributes are of the greatest importance. Priscian’s names and verbs, by contrast, are the minimal conditions for a statement to make sense, but this can be achieved even if the name is equivocal or if the verb is non-assertive, e.g. when it is a subjunctive of wish. While in Aristotle’s Perihermeneias the signification of concepts by names and attributes is fundamental to explaining how they bring about assertions that can be granted as true or rejected as false, in Priscian’s Institutiones the signification of words is fundamental to the division and further definition of the parts of speech. However, the role of this signification in his explanation of grammaticality is far from crucial. 3.2

Names and Verbs in Aristotle’s Perihermeneias

In Chapters 2 and 3 of the Perihermeneias, Aristotle provides further conditions for a name and an attribute to bring about an assertion that can be granted as true or rejected as false. Aristotle defines names as follows: A name is an utterance that is significant by convention, without [signification of] time, of which no part is significant [if taken] separately.36 In this definition, the expressions ‘without indication of time’ and ‘of which no part is significant if taken separately’ refer to the distinctive features that separate names from attributes and sentences (λόγοι), respectively. cum primo loco nomen, secundo verbum posuerunt, quippe cum nulla oratio sine iis completur, quod licet ostendere a constructione, quae continet paene omnes partes orationis. A qua si tollas nomen aut verbum, imperfecta fit oratio; sin autem cetera substrahas omnia, non necesse est orationem deficere…’. Cf. Apollonius, On Syntax (ed. Uhlig, 16:12–17:1). 36 Arist., Int. 2.16a19–21: ‘Ὄνομα μὲν οὐ�ν ἐστὶ φωνὴ σημαντικὴ κατὰ συνθήκην ἄνευ χρόνου, ἡ�ς μηδὲν μέρος ἐστὶ σημαντικὸν κεχωρισμένον.’

Names And Verbs In Priscian And In Aristotle

119

The feature ‘of which no part is significant if taken separately’, i.e. a feature that names share with attributes,37 intends to show that names and attributes, as opposed to sentences, are semantically simple and hence not susceptible of truth and falsity. This feature is explained by means of an a fortiori argument: If even a compound name is semantically simple, a fortiori simple names must also be semantically simple. Aristotle considers the case of the proper name ‘Κάλλιππος’, in order to show that its signification is not the composition of the signification of its parts: In fact, in ‘Κάλλιππος’, ‘ἵππος’ indicates nothing by itself, as it does in the phrase ‘καλὸς ἵππος’.38 This name, which would be the crasis of ‘καλός’ (fair) and ‘ἵππος’ (horse), is in fact a proper name for human beings. Hence, evidently what is indicated by ‘ἵππος’ does not contribute directly to the signification of the word ‘Κάλλιππος’; for, when someone makes an assertion about ‘Κάλλιππος’, she only has a man – Κάλλιππος – in mind and nothing related to horses or to fairness. By contrast, she would have this in mind when using ‘καλός’ and ‘ἵππος’ separately, as in the sentence ‘καλὸς ἵππος’. Therefore, since the part of a compound name, as ‘Κάλλιππος’, does not contribute to its signification, a fortiori the parts of simple names do not contribute to their signification either – names and attributes in general are semantically simple and therefore not susceptible of truth and falsity.39 On the other hand, the lack of indication of time in names is better understood in contrast to the indication of time in attributes. In Chapter 3, Aristotle defines attributes as follows: An attribute is that which in addition signifies time, of which no part signifies if it is separated; and it is the sign of things that are said about something else.40 He goes on to explain the two features that separate attributes from names, namely that in addition they indicate time and are signs of things said about something else. The opposition between the lack of indication of time in 37 Cf. Arist., Int. 3.16b5–6. 38 Arist., Int. 2.16a21–26: ‘ἐν γὰρ τῳ� Κάλλιππος τὸ ιππος οὐδὲν καθ’ αὑτὸ σημαίνει, ὥσπερ ἐν τῳ� λόγῳ τῳ� καλὸς ἵππος.’ 39 For compound names, see Whitaker, ‘Aristotle’s De interpretatione’, pp. 37–45. 40 Arist., Int. 3.16b6–7: ‘Ῥη᷑ μα δέ ἐστι τὸ προσσημαι᷑νον χρόνον, οὑ� μέρος οὐδὲν σημαίνει χωρίς· ἔστι δὲ τω᷑ ν καθ’ ἑτέρου λεγομένων σημει᷑ον.’

120

chapter 3

names and the indication of time in attributes suggests that in the context of an assertion, the utterance of a name amounts to a timeless expression of its content. For example, if someone says ‘health is good’, the utterance of ‘health’ (ὑγίεια) within the assertion involves no indication of the time at which health occurs. Now, if someone says ‘Socrates is healthy’, the utterance of the attribute ‘is healthy’ (ὑγιαίνει) expresses not only that health is the case for Socrates but also that health is the case for Socrates at the present time. However, when they are uttered in a non-assertoric context, there is no difference between names and attributes.41 In other words, the non-assertoric utterance of ‘ὑγιαίνει’ indicates a sort of being, just as names do, but does not express that e.g. being healthy is the case for some bearer at a certain time, which is exactly what an utterance qua attribute does. Here again, an a fortiori argument supports this claim: For neither ‘to be’ nor ‘not to be’ is the sign of a state of affairs (πράγματος), nor if you say ‘being’ alone; for it is nothing, but it signifies in addition a composition that cannot be understood without the things that are composed.42 Just as in the case of ‘goat-stag’, which provided an a fortiori argument for the fact that the mere utterance of a name does not state the existence or nonexistence of its content, the case of the word ‘being’ provides the following argument: If not even the utterance of ‘being’ indicates that something is actually the case, a fortiori the utterance of other attributes does not indicate that something is or is not the case. In other words, no simple utterance in a nonassertoric context can properly be said to be an attribute – attributes are assertive, i.e. they entail a composition with a subject of attribution that produces a statement that is susceptible of truth or falsity (i.e. an assertion). To sum up, the attribute is a significative part of an assertion that, in addition to its indication of a way of being, expresses both that this way of being is the case for something else and that it is so at a certain time. By contrast, the name, within or outside the assertion, always expresses a way of being, timelessly and neutrally with respect to existence or the actual possession of attributes. Thus, right after having put forth his definitions of names and attributes, Aristotle defines the assertoric sentence (λόγος ἀποφαντικός) as a sentence that 41 Cf. Arist., Int. 3.16b19–26. 42 Arist., Int. 3.16b22–25: ‘οὐ γὰρ τὸ εἰν� αι ἢ μὴ εἰν� αι σημει᷑όν ἐστι του᷑ πράγματος, οὐδ’ ἐὰν τὸ ὂν εἴπῃς ψιλόν. αὐτὸ μὲν γὰρ οὐδέν ἐστιν, προσσημαίνει δὲ σύνθεσίν τινα, ἣν ἄνευ τω᷑ ν συγκειμένων οὐκ ἔστι νοη᷑ σαι.’

Names And Verbs In Priscian And In Aristotle

121

is susceptible of receiving a truth-value – true or false – because it is composed of a name and an attribute.43 Thereafter, he defines the affirmation as an assertoric sentence stating that something holds of something else (ἀπόφανσις τινὸς κατὰ τινός) and the negation as an assertoric sentence stating that something is separated from something else (ἀπόφανσις τινὸς ἀπὸ τινός).44 § It is clear that Aristotle’s definitions of names and attributes greatly differ from Priscian’s definitions, because the two authorities approach these terms with fundamentally different agendas in mind: While Aristotle intends to set forth the features that allow words to bring about assertions that are susceptible of truth and falsity, Priscian intends to determine the features that allow words to be divided into different grammatical categories. Consequently, the roles that Aristotle’s ‘σημαίνειν’ and Priscian’s ‘significare’ play in their discussions of names, verbs and sentences diverge in fundamental aspects. This does not come as a surprise if we recall that the main purpose of a grammatical treatise in Priscian’s time was to provide a tool for the textual reconstitution and literary analysis of classical literature.45 In this regard, poetry has a place of privilege and thus such a task quite evidently demands an assessment of the grammaticality and sense of the sentences that are to be analysed and reconstituted, but has little to do with their truth and falsity. Thirteenth-century scholars had to read and comment on Aristotle’s logical works and on Priscian’s grammar, and they had to deal with crucial terms with quite different senses coming from ancient traditions with different approaches to language. The medieval task was arduous, because of the lack of access to the Greek sources and the subsequent ahistorical approach to authorities. For medieval scholars this approach was unavoidable given their lack of access to the historical aspects of the development of ancient linguistic ideas. The interaction of these ancient sources – Aristotle’s logic and Priscian’s grammar – also 43

44 45

Cf. Arist., Int. 5.17a16–20. Note that being the compound of a name and an attribute is a necessary but not a sufficient condition for an assertion to be susceptible of being true or false. Cases of assertions that are the compound of a name and an attribute without being susceptible of receiving a truth-value are indeterminate assertions, such as ‘man is white’ or contingent statements in the future tense, such as ‘there will be a sea-battle’. Cf. Arist., Int. ch. 7 and 9. Cf. Arist., Int. 6.17a25–26. See de Nonno, ‘Ars prisciani caesariensis’, pp. 260–268.

122

chapter 3

yielded scholarly discussions that greatly contributed to the development of medieval linguistics in general and of the notion of signification in particular. In the last section of this study, we shall see the discussions concerned with the different uses of the notion of signification in logic and grammar. These discussions shall give further evidence of the progressive transition in thirteenth-century linguistic treatises from a somewhat attached and conflictive reading of authoritative sources to a more detached reading in which the notion of signification finds a narrower field of application – that of semiotics.

chapter 4

The Role of the Significate (significatum) in Grammar and in Logic Thirteenth-century commentators on Priscian’s Institutiones wonder about his remark that the logician considers only names and verbs as parts of speech,1 as well as about the different definitions of names and verbs that we find in Priscian’s treatise and in Aristotle’s Perihermeneias. The medieval way of tackling the issue is to read the different significationes that serve to define Priscian’s parts of speech as modalities of signification and not as a diversity of semiotic contents. However, there are still substantial divergences regarding the way to use this modal approach to signification in grammatical enquiries and regarding the role that the significate of the word ought to play in grammatical and logical accounts – e.g. in the definitions of parts of speech in general, and of names and verbs in particular. 4.1

The Pre-Modist Tradition2

The early thirteenth-century analysis of the word qua part of speech is based on the idea that linguistic signs3 – words (dictiones) – are a hylomorphic 1 Note that this question is also raised by Boethius, who says [In Perih.2 (ed. Meiser, 14:9–15:5)]: ‘Exsistit hic quaedam quaestio cur duo tantum nomen et verbum se determinare promittat, cum plures partes orationis esse videantur… Quamquam duae propriae partes orationis esse dicendae sint, nomen scilicet atque verbum. Haec enim per sese utraque significant, coniunctiones autem vel praepositiones nihil omnino nisi cum aliis iunctae designant; participia verbo cognata sunt, vel quod a gerundivo modo veniant vel quod tempus propria significatione contineant; interiectiones vero atque pronomina necnon adverbia in nominis loco ponenda sunt, idcirco quod aliquid significant definitum, ubi nulla est vel passionis significatio vel actionis.’ 2 The most important studies on the medieval grammatical tradition are L.G. Kelly, The Mirror of Grammar: Theology, philosophy and the modistae (Amsterdam: John Benjamins, 2002); C.H. Kneepkens, ‘The Priscianic Tradition’, in S. Ebbesen (ed.), Sprachteorien in Spätantiken und Mittelalter, vol. 3 of P. Schmitter (ed.), Geschichte der Sprachtheorie (Tübingen: Gunter Narr Verlag, 1995), pp. 239–264; J. Pinborg, Die Entwicklung der Sprachtheorie im Mittelalter (Münster: Aschendorff, 1967); Rosier-Catach, ‘Grammar’; id., ‘Modisme, pré-modisme, proto-modisme: vers une définition modulaire’, in S. Ebbesen and R. Friedmann (eds.), Medieval Analyses, pp. 45–81. 3 For an exhaustive study of the development of 13th-century semiotics, see C. Marmo, La semiotica del xiii secolo (Milano: Studi Bompiani, 2010); id., Semiotica e linguagio. © koninklijke brill nv, leiden, ���� | doi 10.1163/9789004300132_006

124

chapter 4

composition of utterance (vox), as the material part, and significate (significatum), as the formal part, a composition that follows from the imposition of the utterance on a concept – its first articulation.4 A word is a grammatical part of speech (pars sermonis) – name, verb, adverb etc. – because of its second articulation, i.e. its modes of signifying or its consignificates. Already in the twelfth century, Peter Helias5 wonders why names are defined as words that signify qualified substances, even though some names, e.g. ‘a walk’, clearly do not signify a substance at all. His way to tackle this problem is to claim that, although words are generally instituted on a significate with the purpose of communicating someone’s own will, words qua parts of speech are instituted with further distinct purposes, so that different parts of speech have different causes of invention (i.e. of institution or imposition): The common cause of the invention of every word is for a man to have a way to make public his own will to someone else. But we shall show the 4 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 84–85): ‘Dicendum quod omnes dictiones univocantur in eo quod omnes significant ex impositione, et ex hoc quod omnes constant ex voce pro materia et ex intellectu pro forma… Ad secundum dicendum quod grammaticus non considerat dictionem principaliter nisi propter orationem; et propter hoc definit eam per orationem, cuius est pars. Potest tamen definiri dictio per suum significatum dicendo: dictio est vox primo instituta ad significandum, vel dictio est vox significativa ad placitum.’ For the hylomorphic composition of the dictio and its Aristotelian origin, see Pinborg, Entwicklung; and Marmo, Semiotica e linguaggio; for the Porphyrian origin of the notion of imposition, see C.H. Keepkens, ‘A Note on Articulatio and University Grammar’, in Fink et al. (eds.), Logic and Language, pp. 221–238; Kelly, The Mirror, ch. 1; and Rosier-Catach, La parole, ch. 4.1. For imposition, see also above, Part 1, Section 2.1. 5 Early 13th-century commentaries on Priscian’s Institutiones are greatly indebted, both in structure and in content, to Peter Helias’ Summa super Priscianum (ca. 1139). The two main contemporary sources of Peter’s Summa are an early commentary on Priscian that comes from the school of William of Champeaux (also called the Glosule) and the commentary on Priscian by William of Conches (also called the Glosae). Helias is also a direct source of the early 13th-century modal interpretation of Priscian’s definitions of the parts of speech; see Leo Reilly’s introduction to his edition of Petrus Helias’ Summa super Priscianum [Petrus Helias, Summa super Priscianum. 2 Vols., ed. L. Reilly (Toronto: pims, 1993) (Studies and Texts 113)]; K.M. Fredborg, ‘Notes on the Glosule and its reception by William of Conches and Peter Helias’, in I. Rosier-Catach (ed.), Arts du langage et théologie aux confins des xie–xiie siècles. Textes, maîtres, débats (Turnhout: Brepols, 2011) (Studia artistarum 21), pp. 453–484; and R.W. Hunt, ‘Studies on Priscian in the Eleventh and Twelfth Centuries I: Petrus Helias and his Predecessors’, Medieval and Renaissance Studies 1 (1943), 194–231; for a recent collection of essays on the 11th and 12th-centuries schools, see Rosier-Catach (ed.), Arts du langage et théologie.

the Significate in Grammar and in Logic

125

proper cause of the invention of each part of speech and of its accidents when dealing with each of them.6 Accordingly, the words’ modes of signifying make them a specific part of speech: Do you know why all names are said to be one and the same part of speech? This is because this part of speech is separated from the other [parts] by virtue of its mode of signifying.7 In addition to the significate – which is given to a word to make public a thought – a word falls under a particular grammatical category because of the way in which it indicates this significate; a way that follows from a further particular intention at the moment of the word’s institution. Thus, we find in Peter Helias the suggestion of an intentional cognitive approach to the significates, which is at the origin of the different modes of signifying that determine the different parts of speech. This modal approach to the parts of speech is also found in Helias’ interpretation of Priscian’s claim that names signify substance with quality: Names signify something in the way of a qualified substance; for, although some names signify something that is actually a qualified substance, not every name does so. The signification of substance with quality is thus understood as a modality – the indication of a significate in a certain way: Therefore, every name signifies substance and quality. Not that every name signifies something that is a substance, but because every name signifies either something that is a substance or something in the way of a substance with respect to a common or proper nature… We judge that, at present, this reading of the signification of names or verbs is to be preferred, as it suits best the proper feature of grammar.8 6 Peter Helias, Summa (ed. Reilly, 177:51–54): ‘…communis causa inventionis omnium dictionum est ut haberet homo quomodo propriam voluntatem alteri manifestaret. Propriam vero causam inventionis cuiusque partis orationis eisque accidentium tractando de singulis demonstrabimus.’ 7 Peter Helias, Summa (ed. Reilly, 181:38–40): ‘Scis quare omnia nomina dicantur esse una et eadem pars orationis, ita quod unumquodque illorum est illa pars? Ideo quod haec pars orationis distinguitur ab aliis secundum suum modum significandi.’ 8 Peter Helias, Summa (ed. Reilly, 196:36–46): ‘Omne igitur nomen significat substantiam et qualitatem, non quod omne nomen significet id quod est substantia, sed quia omne nomen vel id quod est substantia significat vel aliquid modo substantiae in natura communi vel

126

chapter 4

According to Helias, then, in the context of Priscian’s definitions of parts of speech, ‘signification’ is more properly read as a modality, i.e. as a mode of signifying (modus significandi). Helias’ modal reading will be adopted by most thirteenth-century interpreters of Priscian’s Institutiones. For e.g. Ps-Kilwardby the claim that names signify substance with quality means that they signify a qualified substance, where neither substance nor quality are to be understood as Aristotelian categories but as modes of signifying something as if it were a substance or a quality: It is evident that to signify the substance is to signify something in the way of what is standing and permanent, finite or infinite, whether it is a being or a non-being…and the quality that is signified by the name is not the category, as it was opposed, but a mode of understanding the bearer with respect to a common or proper nature, or common or proper as it were…9 The grammarian deals with things insofar as they are susceptible of being signified in a certain way, and not insofar as they are of a certain sort. Thus, the actual essential or accidental forms that make something a particular sort of thing, as well as its existence or non-existence, are irrelevant for the grammatical analysis of names. Therefore, in Priscian’s definition of names, substance and quality are not categories of being or of predication, but modes of understanding – and hence of signifying – a bearer (suppositum) as if it were the holder of common or proper features. Here again, we find in the phrase ‘mode of understanding’ the suggestion of a cognitive approach to the significate that is at least concomitant to, if not the cause of, the mode of signifying that makes a word a particular part of speech. Ps.-Kilwardby goes a step further and tells us that the modes of signifying are not only the principles of the division of the parts of speech but also the principles of the grammaticality of linguistic constructions: The general mode of signifying is the first principle, while the accidental mode of signifying is the proximate principle, from the point of view of propria… Hanc vero de significatione nominum seu verborum sentenciam ad praesens praeferendam utpote artis gramaticae proprietati accomodatam iudicavimus.’ 9 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 116): ‘…planum est quod significare substantiam est significare quid per modum stantis seu permanentis, finite aut infinite, sive sit ens sive non-ens… Qualitas autem quae significatur per nomen non est res praedicabilis, sicut opponebatur, sed modus intelligendi suppositum in natura communi vel propria aut quasi communi aut quasi propria…’.

the Significate in Grammar and in Logic

127

the construction. Hence, one must say that in general the unity of proportion or of likeness of the general mode of signifying (or of consignifying) is the precise cause of grammaticality.10 The name signifies the content on which it is conventionally imposed – a concept – as if it were a qualified substance. The verb, in turn, signifies its content as if it were an action or the undergoing of an action. Finally, the name and the verb can bring about a grammatical sentence because their modes of signifying are compatible with each other. By the time of Ps.-Kilwardby, then, the modes of signifying have become a central explanatory principle in thirteenth-century theoretical grammar. This important fact also involves a modal approach to signification, where certain intentionality is at the origin of both a particular cognitive approach to a significate (a mode of understanding) and the subsequent grammatical approach to it (the mode of signifying). However, this does not happen without raising a number of problems, namely about the possibility that logic – which is concerned with truth or falsity – and grammar – which is concerned with grammaticality – share some of their central explanatory principles, with signification under the spotlight. Ps.-Kilwardby indeed raises some questions about the different approaches to language in grammar and logic and decides on a number of aspects where the grammarian’s enquiry departs from the logician’s. First, grammar and logic have different principles (or at least they should have) because neither is subordinated to nor part of the other. They are in fact mutually independent disciplines with mutually independent concerns and principles. The principles of grammar are the words’ modes of signifying and of consignifying – the causes of grammaticality – while those of logic are common intentions – presumably the causes of truth and falsity. Hence, modes of signifying and common intentions ought to be independent from each other: The principles of grammar are not arranged with respect to the principles of logic, because the principles of grammar are the general or specific modes of signifying and consignifying of words. The principles through which logic comes forth, and that it considers, are the common intentions 10 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 93–94): ‘Unde modus significandi generalis est primum principium, modus significandi accidentalis sive consignificata est proximum principium a parte constructionis. Unde generaliter dicendum quod unitas in modo significandi generali aut consignificandi proportionis vel similitudinis est causa praecisa congruitatis.’

128

chapter 4

grounded on things, just as the universal, the particular, the genus, the species, the cause, the caused, and so on. Now, the modes of signifying and consignifying things and the common intentions of things are not arranged with respect to each other, but they are rather separated, since they have different causes.11 On the other hand, the concerns of each discipline also bring about different analyses of the arrangement of parts of speech. In grammar, the arrangement of parts of speech is considered with respect to grammaticality and its cause – the general modes of signifying. In logic it is considered with respect to truth and falsity and their cause – the significates (significata specialia): ‘Articulate utterance’ is said in two ways: in itself or with respect to something else. [Articulate utterances] in the first sense are the word and the sentence; in the second sense the letter and the syllable. [The articulate utterance] in itself is threefold: (i) the articulation with respect to grammaticality, which depends on the general modes of signifying; (ii) [the articulation] with respect to truth, and this one depends on the specific significates; (iii) [the articulation] with respect to ornament…12 Accordingly, each discipline puts forth a different number of parts of speech. For the grammarian every word that contributes in a distinct manner to the grammaticality of a sentence is a part of speech, whereas for the logician only 11 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 25–26): ‘...principia grammaticae non habent ordinem ad principia logicae, quia principia grammaticae sunt modi significandi vel consignificandi, generales vel speciales dictionum. Principia per quae procedit logica et quae considerat sunt communes intentiones fundatae in rebus, sicut sunt universale, particulare, genus, species, causa, causatum et sic de aliis. Modi autem significandi res aut consignificandi et communes rerum intentiones non habent ordinem sed potius disparationem, cum a diversis causentur.’ 12 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 39): ‘...vox articulata dicitur duobus modis: uno modo in se, alio modo in altero. Primo modo dictio et oratio, secundo modo littera et syllaba. In se tripliciter: una est articulatio respectu congrui quae est quantum ad modos significandi generales, alia respectu veri et hoc est quantum ad significata specialia, alia respectu ornatus…’. See also Kneepkens, ‘A note on articulatio’. Cf. e.g. Nicholas of Paris, In Perih. (ed. Hansen and Mora-Márquez, 35:24–36:3 and 38:25–39:7). For the same position, see also AnOx, In Perih. (f. 35rb). For significatio generalis and significatio specialis, see C.H. Kneepkens, ‘Significatio generalis and significatio specialis: Notes on Nicholas of Paris’ Contribution to Early Thirteenth-Century Linguistic Thought’, in Ebbesen et al. (eds.), Medieval Analyses, pp. 17–43. For Nicholas’ incomplete commentary on Priscian minor, see Kneepkens, ‘A note on articulatio’, p. 230.

the Significate in Grammar and in Logic

129

the subject (de quo aliquid dicitur) and the predicate (quod de alio dicitur) are considered parts: To the first [argument] it must be said that the grammarian says that a word is what by virtue of its signification produces a sentence. But the logician says that words are only that about which something is said and that which is said about something else. Hence, [the logician] says that there are two parts of speech…13 In fact, in order to produce a sentence that is susceptible of truth or falsity, only a subject and a predicate are required. Consequently, the significate plays different roles in each discipline. In logic, it is the cause of truth and falsity, insofar as the significate is the bearer of some internal features, such as actual existence, as well as identity and diversity. In grammar, it is the cause of grammaticality and ungrammaticality, insofar as the significate is the bearer of certain external relations that determine e.g. the different grammatical cases (nominative, accusative etc.).14 Hence, there are more than two parts of speech according to the grammarian; for the words’ significates can have external relations that determine grammatical features in more than two ways: It must be said that there are more than two parts of speech, grammatically speaking, because there are several modes of signifying and understanding whereby there are more than two parts of speech. To the first [argument], it must be said that the parts of speech are not there only in order to signify an intelligible thing, whether it is or is not an entity, but they are there in order to signify in different ways the same or diverse intelligible things, or in order to signify in the same way several things that are intelligible also in different ways.15 13 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 85): ‘Ad primum dicendum quod grammaticus dicit illud esse dictionem quod quantum ad significationem constituit orationem. Sed logicus dicit illud solum esse dictionem de quo aliquid dicitur vel quod de alio dicitur. Unde dicit duas esse partes orationis…’. Cf. Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 89): ‘Ad secundum dicendum quod logicus considerat orationem ratione suae significationis, et quia trahit illam a suis partibus, quae sunt nomen et verbum, ideo ante definitionem orationis ponit definitionem nominis et verbi… Sed grammaticus considerat orationem proprie in quantum est finis ordinis dictionum per constructionem et in quantum est instrumentum loquendi.’ Note that in the first passage I paraphrase ‘significationem’ as ‘sense’. 14 Cf. Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 95). 15 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 104): ‘Dicendum quod plures sunt partes orationis quam duae grammatice loquendo, quia plures sunt modi significandi vel

130

chapter 4

Thus, the different external features of the significate explain the different cognitive approaches to it; different approaches that in turn give support to the different modalities of signification that are explanatory in grammar. Thereby, it is also clear why names and verbs qua parts of a grammatical sentence and qua subject and predicate are not quite the same. For in the first case a word is a name or a verb because its significate has a mode of signifying allowing it to produce a grammatical sentence, whereas in the second case a word is a name or a verb because its significate has a different set of features allowing it to produce an assertion that can be true or false. It is noteworthy, however, that both the modes of signifying and the logical features, such as actual existence and identity, are features of the significate. Thus, through two distinct sets of features, the same significate plays a role both in the explanation of grammaticality and in the explanation of truth and falsity – two accounts that are central to grammar and logic, respectively. Let us see this in greater detail. Ps.-Kilwardby raises and discusses the question whether the significate actually contributes to the grammaticality of sentences.16 The question is interesting; for, given that there is a modal approach to signification in medieval grammar, it would also be plausible to consider the different modalities of signification (rather than the significate) as the scientific object of the grammatical enquiry.17 In Ps.-Kilwardby’s question, the argument pro states that a name demands an inflected name because of the quality it signifies (according to Priscian’s definition of names, that is). Therefore the significate of the name, namely its quality, determines its construction with e.g. the dative or the genitive.18 The argument contra states that truth and grammaticality are different features i­ntelligendi, propter quos sunt plures partes orationis repertae quam duae. Ad primum ergo dicendum, quod partes orationis non sunt repertae ad significandum intelligibile solum sive sit ens sive non, sed sunt repertae ad diversimode significandum idem intelligibile aut diversum aut ad eodem modo significandum diversa intelligibilia et diversimode.’ 16 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 95): ‘Sed tunc quaeritur, utrum ipsum significatum faciat ad congruitatem.’ 17 Indeed, this position was clearly taken in the modist tradition. See below: The modist tradition. 18 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 95): ‘Videtur quod sic, quia nomen exigit obliquum per naturam qualitatis dependentis. Sed illa qualitas significata est. Ergo etc. Maior patet in hiis orationibus “capa Socratis”, “pater Verbi est ubique maritus”.’ Note that Latin is an inflected language, and hence some grammatical constructions demand particular inflexions of the words involved. For instance, in an expression of possession, such as ‘Peter’s dog’, the possessor’s name – ‘Peter’ – will always be inflected in the genitive, in

the Significate in Grammar and in Logic

131

(passiones) of sentences and consequently they have different causes, and since the significate is the cause of truth and falsity, it cannot be the cause of grammaticality.19 Ps.-Kilwardby goes about the problem by establishing two possible perspectives on the significate of a word. On the one hand, one can focus on the significate as something of a determinate sort of being (e.g. a man, a horse, an apple, white or just), which has actual existence or not, and which is identical to or different from another sort of being. In other words, one can focus on the properties according to which something can be affirmed or denied of the significate. According to these properties, the significate is the cause of truth and falsity: It must be said that the word’s significate can be seen from two perspectives: in one way as being or non-being and further under those differences that are identity and diversity, and thus it is the principle of what is true or false, because [something is true or false] because the thing is or is not etc.20 On the other hand, the significate can be seen insofar as it is externally related to other things, and according to these relations it is the cause of grammaticality: In another way, [the significate] can be considered insofar as it depends on something else, such as to whom it belongs, for whom it is intended, and so on for the other relations of the cases – and thus it is the principle of what is grammatical or ungrammatical…21 Consider e.g. the word ‘apple’, whose significate is the concept [apple]. One could see this concept as having an internal composition that includes (a) the this way: ‘Canis Petri’; and in a sentence like ‘the flowers for Peter’, ‘Peter’ will be inflected in the dative, in this way: ‘Flores Petro’. 19 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 95): ‘Contra: Verum et incongruum sunt diversae passiones, ergo habent diversa principia. Sed veritas habet ortum a significato dictionis, ergo congruitas vel eius oppositum non habebit ortum a significato.’ 20 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 95): ‘Dicendum quod significatum dictionis potest dupliciter considerari: uno modo ut ens vel non-ens et ulterius sub hiis differentiis quae sunt idem et diversum; et sic est principium veri vel falsi, quia ab eo quod res est vel non est etc.’ 21 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 95): ‘Alio modo potest considerari in quantum dependens est ad aliud, ut cuius est vel cui aliquid acquiritur, et sic de aliis habitudinibus casuum; et sic est principium congrui vel incongrui…’.

132

chapter 4

essential or accidental features of apples themselves and (b) the external relations of apples to other things. ‘Apple’ is the potential term of a true assertion by (a), but it is a potential part of speech by (b). In other words, different features of the significate account for the truth and falsity of assertions involving it – which belongs to the logical domain – and for the grammaticality or ungrammaticality of sentences involving it – which belongs to the grammatical domain. Hence, to the argument according to which the name is of a certain grammatical nature because of the quality it signifies, Ps.-Kilwardby adds: To the first [argument] it must be said that such quality can be referred to the bearer in two ways: In one way, as to that in which it is, and this is [the quality] of a being or non-being and it forms a unity with that in which it is, and in this way [the quality] is neither a principle of construction, nor of the demand [of a case]. In another way, as to the term to which it is related, and thus it is a principle of construction and it demands cases as the term of its dependence.22 So, ‘quality’ can be understood in two ways: First, as an inhering essential or accidental form, i.e. the sort of quality that can be predicated of a bearer with which it constitutes some sort of ontological unity, so as to be the principle of true or false attribution. Second, as an external relation, i.e. the sort of quality that determines a relation to external things, so as to be the principle of the grammatical construction of the name with other nominal inflexions. Consequently, the same sense of ‘quality’ is not at stake in logical and in grammatical principles,23 to the effect that different features of the significate determine truth or falsity, on the one hand, and grammaticality, on the other.

22 Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 96): ‘Ad primum dicendum quod illa qualitas potest dupliciter referri ad suppositum, et sic est entis vel non entis et facit unum cum eo cuius est; et sic non est principium construendi nec exigentiae. Alio modo ad terminum ad quem est; et sic est principium construendi et exigit obliquos tamquam terminum suae dependentiae.’ 23 It is noteworthy that this approach is not taken directly from Priscian; for, as we already saw, in Priscian’s grammar signification is only explanatory as regards the deduction of the parts of speech, and not as regards the construction of grammatical sentences. The sources of this approach go back to 12th-century grammar – to Peter Helias and his immediate sources. See Rosier-Catach, ‘Grammar’, pp. 196–200; and C.H. Kneepkens, ‘Grammar and Semantics in the Twelfth Century: Petrus Helias and Gilbert de la Porrée on the Substantive Verb’, in M. Kardaun and J. Spruyt (eds.), The Winged Chariot: Collected Essays on Plato and Platonism in Honour of L.M. de Rijk (Leiden: Brill, 2000), pp. 237–275.

the Significate in Grammar and in Logic

133

Ps.-Kilwardby’s question is directly related to another question about the different definitions of names and verbs in logic and grammar; a question that is mostly raised in commentaries on the Perihermeneias from the first half of the thirteenth century and that echoes Boethius’ concern in his own commentary about the different approaches to names in the two disciplines.24 We find the question in the two recensions of the commentary on the Perihermeneias by Nicholas of Paris, as well as in AnOx’s and Robert Kilwardby’s commentaries. Like Ps.-Kilwardby, Nicholas of Paris tackles the question by saying that logic deals with truth and falsity, whereas grammar deals with grammaticality and ungrammaticality: To the first [question] it must be said that regarding speech the grammarian considers the grammatical and ungrammatical sentence [and] the logician [considers] truth and falsity. Therefore, since an arrangement is a placement of parts (in fact, order and place are placements of parts) – for a sentence is said [to be] grammatically arranged, not because it is arranged in relation to other [sentences], but because its parts are placed and arranged in the due manner – whence the grammarian defines the name and the verb insofar as they are parts [of speech], for insofar as they are parts [of speech] they take an arrangement. But truth and falsity come from the thing signified; for [they come about] ‘because the thing is or is not etc.’. Hence the logician defines the name and the verb in comparison to the significate.25 Hence, grammaticality follows from a coherent arrangement of words, which does not depend on the significates of a sentence’s parts. The truth or falsity of an assertion, by contrast, depends on the significates of its parts, because its truth or falsity comes from the truth or falsity of the compound thought it 24

25

Cf. Boethius, In Perih.2 (ed. Meiser, 32:17–22): ‘Vox enim quae nihil designat, ut est garalus, licet eam grammatici figuram uocis intuentes nomen esse contendant, tamen eam nomen philosophia non putabit, nisi sit posita ut designare animi aliquam conceptionem eoque modo rerum aliquid possit.’ Nicholas of Paris, In Perih. (ed. Hansen and Mora-Márquez, 38:15–24): ‘Ad primum dicendum quod grammaticus circa sermonem considerat congruam vel incongruam orationem, logicus veritatem vel falsitatem. Quoniam ergo ordinatio est positio partium (est enim positio partium situs et ordo), dicitur enim oratio congrue ordinata non quia ipsa sit ad alia ordinata, sedpartes eius sunt debito modo ordinatae et positae, ideo grammaticus definit nomen et verbum [et hoc est] inquantum partes, nam inquantum partes suscipiunt ordinem. Veritas autem et falsitas est a parte rei significatae; nam “in eo quod res est vel non est etc.” Ideo logicus definit nomen et verbum per comparationem ad significatum.’

134

chapter 4

indicates. In other words, the grammaticality of a sentence is not determined by comparison with another compound item from which its grammaticality is taken, whereas the truth of an assertion is determined by its comparison with the truth of the compound thought it indicates. To explain this, Nicholas puts forth a double signification of words – their general signification (significatio generalis) and their significate (significatio specialis); a distinction he applies to the case of names. A name is a grammatical part of speech because of its general signification – substance with quality –, which is the cause of its construction in grammatical sentences. But logic’s concern with names regards their significates (significata specialia); for names contribute to the truth or falsity of assertions through their relation to the things they signify: To the other [question], it must be said that the signification of the name is double: general, whereby every name is said [to be] a name – because it signifies substance with quality; and this general signification is the principle of construction. Whence, the grammarian puts this one in the definition of the name. There is also the specific signification, on which the name was arbitrarily imposed, and the logician considers this one, because truth and falsity come about thanks to this [signification]; but [Aristotle] could not put it in the definition of the name, because it is not unique but diverse in diverse names.26 Thus, all names, qua parts of speech, signify qualified substances, but each name qua subject of a true or false assertion signifies a particular significate that involves a determinate sort of being (e.g. the significate of ‘man’ involves the notion of being a man) and that is different in different names. Along the same lines, AnOx claims that names and verbs qua parts of speech signify substance with quality and action and undergoing of action; for these are the principles of grammatical constructions. However, contrary to Nicholas, AnOx is explicit about the modal reading of signification in these definitions: The grammarian says that names signify a substance with a quality – a qualified 26

Nicholas of Paris, In Perih. (ed. Hansen and Mora-Márquez, 38:25–39:7): ‘Ad aliud dicendum quod duplex est significatio nominis, scilicet generalis, qua nomen dicitur esse nomen, scilicet quia significat substantiam cum qualitate; et haec significatio generalis est principium construendi. Ideo hanc ponit grammaticus in definitione nominis. Est etiam significatio specialis, ad quam impositum est nomen ad placitum, et hanc consi­ derat logicus, quia penes hanc consistit veritas et falsitas. Sed hanc non potuit ponere in definitione nominis, quia non est una sed diversa in diversis nominibus.’

the Significate in Grammar and in Logic

135

substance – because substances are what perform and undergo actions. However, this does not entail that the significate of a name is necessarily a substance, but only that it is signified as a substance – as something that performs or undergoes an action: To the other [question] that was raised, it must be said that for the grammarian the name signifies substance with quality, and the verb [signifies] action or undergoing of action; but for the grammarian the name consi­ dered in itself does not always signify a substance. That is, the name in itself does not signify a substance, for instance ‘whiteness’ [does not signify a substance], but a quality. However, when someone says: ‘The whiteness fades’, ‘whiteness’ is a substance as regards the arrangement. Similarly, the verb in itself does not signify always an action or the undergoing of an action…27 The anonymous master ends his discussion by putting forth the distinction between general signification and significate that we find in Nicholas. The name enters into a grammatical construction because of its general signification – substance with quality – but it is a principle of truth or falsity because of its significate: The other question can be resolved by saying that the signification of the name is double: general and specific. [The name] is the principle of the 27 AnOx, In Perih. (f. 35rb): ‘Ad aliud quod postea quaeritur, dicendum est quod nomen apud grammaticum significat substantiam cum qualitate, et verbum agere vel pati; sed nomen apud grammaticum secundum se consideratum non significat substantiam universaliter. Hoc est, nomen non significat secundum se substantiam, ut “albedo”, sed qualitatem: Cum autem dicitur “albedo disgregat”, respectu huius praedicati “disgregare” substantia est in ordine. Similiter verbum secundum se universaliter non significat agere vel pati…’. Cf. Peter Helias, Summa (ed. Reilly, 859:78–860:100): ‘Sed nota quod unaquaeque ars habet modum suum loquendi. Aliter enim dialecticus loquitur in arte sua, et aliter grammaticus. Dicit enim dialecticus quod non omne nomen significat substantiam quia hoc nomen “albedo” non significat substantiam, sed qualitatem. Grammaticus vero dicit quod omne nomen significat substantiam et qualitates nec est ibi contrarietas… Grammaticus vero dicit quod omne nomen significat substantiam cum qualitate, non quod omne quod nomen significat sit substantia…et ita omne nomen dicitur significare substantiam quia omne nomen significat modo substantiae, id est, significat rem ut de ea aliquid dicitur et sine tempore, et ita significat substantiam. Unde Boecius in Commento super librum Peri ermeneias: “Omne nomen aut significat substantiam aut tamquam substantiam”.’ Cf. Peter Helias, Summa (ed. Reilly, 861:22–864:75).

136

chapter 4

grammatical or ungrammatical sentence with respect to the general signification, but the principle of the true or false sentence with respect to the specific [signification].28 Robert Kilwardby, in turn, adds that logic takes names and verbs qua constructible from the grammarian, but not qua subject and predicate; for the fact that names are constructible because of their signification of substance with quality does not explain why they are subjects of predication: It must be said that the different intentions of the authors account for the diversity of the definitions. Therefore, in grammar [the name] is defined through its essential parts insofar as it is constructible, because [it is constructible] through substance and quality; but here [i.e. in the Perihermeneias] [it is defined] through its essential parts insofar as it is susceptible of being subject… And since in this science [i.e. the Perihermeneias] [Aristotle] intends [to determine] about certain kinds of sentences – the assertion primarily and in itself – whose parts are the subject and the predicate, and [since] all subjects fall under one difference – the name – and all predicates fall under another – the verb – for this reason here the division of parts concerns these two differences, but it is not so in grammar.29 Hence, Kilwardby rightly stresses that in the context of the Perihermeneias, ‘name’ and ‘verb’ are to be understood as ‘subject’ and ‘predicate’ – the essential components of the assertion that is susceptible of truth and falsity because it properly mirrors a compound thought that is susceptible of truth and falsity.

28 AnOx, In Perih. (f. 35rb): ‘Ad aliud, potest solvi dicendo quod in nomine est duplex significatio: Generalis et specialis. Et quantum ad generalem significationem est principium orationis congruae vel incongruae, quantum ad specialem est principium verae orationis vel falsae.’ 29 Robert Kilwardby, Not. sup. Perih. (P 68va; M 47rb; V 4v): ‘Set dicendum quod diversa intentio auctorum fecit diversitatem definitionum. Definitur ergo in grammaticis per partes suae essentiae in quantum est constructibile, quia per substantiam et qualitatem; hic autem per partes suae essentiae in quantum est subicibile… Et quia in scientia ista intendit de quadam specie orationis, scilicet de enuntiatione primo et per se, cuius partes sunt subiectum et praedicatum, et omnia illa quae subiciuntur cadunt in differentiam unam, et est nomen, et illa quae praedicantur in differentiam aliam, et est verbum, propter hoc est hic divisio partium tantum in hiis duabus differentiis; non sic autem in grammaticis.’

the Significate in Grammar and in Logic

137

In the second half of the thirteenth century, Martin of Dacia takes over this idea and says that grammar and logic discuss names and verbs from different perspectives: grammar insofar as they are the principles of grammatical constructions because of their modes of signifying; logic insofar as they are principles of truth and falsity because they are susceptible of being subjects and predicates.30 The same goes for Radulphus Brito, who puts forth a position not considerably different from the positions of the first half of the century: The logician is not concerned with the name insofar as it is an aggregate of utterance, significate and mode of signifying, but as such [the name] concerns the grammarian… And for this reason the grammarian defines the name as a part of speech, because he considers it in itself insofar as it is constructible through its mode of signifying; mode of signifying that makes the parts be parts. But the logician considers the same name, as it is an extreme about which something that signifies such notion can be verified; hence the logician says that [the name] is a significative utterance. Therefore, the utterance and the notion of signification are essential parts of the name, as it concerns the logician.31 Therefore, for the logician, a word is a name if it singles out a subject of predication, and this is why Aristotle claims in the Perihermeneias that inflected names are not names; for they fail to meet this requirement and hence cannot produce assertions that are susceptible of being true or false. However, not only are inflected names not names in logic, but also some words that are not names in grammar are names in logic. For instance, in the assertion ‘to eat 30

31

Martin of Dacia, In Perih. (ed. Roos, 254:20–24): ‘Ad primam quaestionem dicendum est, quod ad logicum pertinet et cetera. Cuius ratio est: nam ad logicum pertinet determinare de omni eo, cui applicata est ratio logica. Sed nomen et verbum sunt huiusmodi. Potest enim eis applicari ratio subicibilis et praedicabilis.’ Martin of Dacia, In Perih. (ed. Roos, 254:30–255:3): ‘Nam grammaticus considerat nomen et verbum, prout habent modum significandi, qui est principium constructionis congruae vel incongruae.’ Radulphus Brito, In Perih. (ed. Venice, 146a–147a): ‘…nomen non pertinet ad logicum ut aggregat in se vocem, significatum et modum significandi, sed sic pertinet ad grammaticum… Et propter hoc dicit grammaticus nomen definiendo: “nomen est pars orationis”, quia ipsum per se considerat prout est constructibile (communicabile inc.) per modum significandi, qui modus significandi facit partes esse partes. Logicus autem considerat ipsum nomen ut est extremum de quo potest verificari aliquid talem rationem significans, ideo logicus dicit quod est vox significativa. Tunc ergo vox est de essentia nominis ut pertinet ad logicum et ratio significandi.’

138

chapter 4

healthy food is good’, the expression ‘to eat healthy food’ is a name – it is the expression that singles out the subject of predication. In grammar, however, ‘to eat healthy food’ is in fact an incomplete sentence. Along these lines, Ps.-Johannes Pagus tells us that, in logic, only names in the nominative can give a bearer to the action that is signified by the verb, so that logic looks at names only insofar as they indicate the bearers of actions and undergoing of actions, because as such they can produce assertions susceptible of truth and falsity: Third, it must be noted that the inflected name is not a name for the logician. And you must note that the inflected name, just as the name in the nominative, is a name with respect to the general significate, because just as the name in the nominative, also the inflected name signifies substance with quality. Similarly as regards the specific significate, because the specific significate is the same in the inflected name and in the name in the nominative… But they differ with respect to the specific mode of signifying… Hence, the logician considers the name insofar as it renders a bearer for the verb, because then truth or falsity are caused in the sentence. Hence, since the inflected name cannot render a bearer for the verb, and consequently cannot cause truth or falsity in the sentence, for this reason the logician does not say that the inflected name is a name, because the logician only considers the parts of a sentence that can cause truth or falsity.32 Interestingly, Ps.-Pagus also claims that the significate, the general signification and the general modes of signifying are the same in a name in the nominative and in its inflexions, but that their specific modes of signifying are different. Thus, the principle of truth or falsity is not the significate (significatum speciale), but the specific mode of signifying that enables a name in the nominative to single out a subject of predication. Ps.-Pagus’ use of a ‘specific mode of 32 Ps.-Pagus, Scriptum super librum Perihermeneias (MS Padova Bibl. Univ. 1589 (= P), ff. 82r–93r, f. 70vb): ‘Tertio notandum quod nomen obliquum non est nomen quoad logicum. Et debetis notare quod nomen obliquum est nomen sicut rectum quoad significatum generale, quoniam sicut rectum significat substantiam cum qualitate, et obliquum; similiter quantum ad significatum speciale, quoniam idem est significatum speciale in recto et obliquo… Sed differunt quantum ad modum significandi specialem… Unde cum logicus consideret nomen inquantum reddit suppositum verbo, quoniam tunc causatur veritas vel falsitas in oratione, cum igitur obliquum non possit reddere suppositum verbo, et per consequens non causare veritatem vel falsitatem in oratione, ideo logicus non dicit obliquum esse nomen, quoniam logicus considerat partes orationis solum quae possunt causare veritatem vel falsitatem in oratione.’

the Significate in Grammar and in Logic

139

signifying’ as a logical principle of truth and falsity stands out as exceptional with respect to comparable commentaries. For, on the one hand, it diverges from grammars where the specific mode of signifying accounts for certain subdivisions within a part of speech (e.g. the internal division of names into proper names and common names); on the other hand, it diverges from authors such as Robert Kilwardby, Ps.-Kilwardby and AnOx, for whom the principle of truth and falsity is the significate itself and not a mode of signifying. Note, however, that Ps.-Pagus’ way of explaining the difference between onomata and rhêmata in Aristotle’s Perihermeneias and nomina and verba in Priscian’s Institutiones is strikingly interesting; for quite rightly he notices that in the Aristotelian definitions, at stake is not so much what the onomata and rhêmata signify but how they signify it.33 The interest in the different approaches to names in grammar and logic disappears in the second half of the thirteenth century. Some questions in this sense are raised in the commentaries on the Perihermeneias by Martin of Dacia, Peter of Auvergne and Simon of Faversham, but only tackled superficially,34 and the problem is ignored by Duns Scotus in his own commentaries. However, scholars from the last quarter of the thirteenth century will discuss, and oppose, most aspects of the positions put forth during the first half of the century, subsequently introducing new perspectives on the subject matter of logic and on the role the word’s significate plays in both grammar and logic. 4.2

The Modist Tradition35

Modistic theoretical grammar implements the earlier idea of a correspondence between modes of understanding and modes of signifying; an idea that accounts for both the distinction of parts of speech and the grammaticality of sentences. But it also adds a further correspondence between modes of signifying and understanding, and modes of being (properties of external things), in 33 34

35

See also Kneepkens, ‘Significatio generalis etc.’ These authors basically limit themselves to rephrase Aristotle’s own argument that the inflected name joined to a verb does not bring about an assertion. Cf. e.g. Simon of Faversham, Quaestiones super Perihermeneias q. 8, in Opera logica i, ed. P. Mazarella (Padova: cedam, 1957), at pp. 161–163. The pioneer study about modism is Jan Pinborg’s Entwicklung. This study was followed by Irène Rosier’s La grammaire spéculative des modistes (Lille: Presses Universitaires de France, 1983), by C. Marmo’s Semiotica e linguaggio; and by L.G. Kelly’s The mirror.

140

chapter 4

order to provide theoretical grammar with the sufficient external ground to make it a science concerned with reality (scientia realis). Accordingly, masters from the second half of the thirteenth century (we shall focus on Martin of Dacia, Boethius of Dacia and Radulphus Brito36) are called modists after their claim of a correspondence between modes of signifying, modes of understanding and modes of being, which explains the division of parts of speech and the grammaticality of sentences. Modistic theoretical grammar, however, has a fuzzy process of development.37 In a first stage (here represented by Martin of Dacia), a word (an utterance imposed on a significate, i.e. on a concept or thing understood [res intellecta]38) is a specific part of speech – e.g. a noun – because of its general mode of signifying.39 The immediate cause of this mode of signifying is a corresponding mode of understanding the significate. Its ultimate cause is a corresponding mode of being that is a feature of external things because it is in external things as in a subject. For Martin of Dacia, the modes of signifying, understanding and being are notionally identical but materially different because they have different bearers. The mode of understanding is a property of the thing understood (res intellecta) qua understood as in a subject. The mode of signifying is a property of the thing understood qua significate as in a subject. The modes of signifying and of understanding are thus properties of the thing understood qua significate and qua understood, respectively, and they correspond to and have real 36

37

38

39

Boethius of Dacia, Modi significandi sive quaestiones super Priscianum maiorem, ed. J. Pinborg and H. Roos (Copenhagen: gad, 1969) (CPhD iv); Radulphus Brito, Quaestiones super Priscianum minorem. Grammatica speculativa, ed. J. Pinborg and H.W. Enders (Stuttgart: Frommann-Holzboog, 1980) (Sprachtheorie und Logik des Mittelalters 3.1–2). It is unclear how this development took place in the time frame between Martin of Dacia and Radulphus Brito, due to the uncertain datation of the grammatical treatises by Martin and Boethius of Dacia. It is certain, however, that there are substantial disagreements between Martin of Dacia and Radulphus Brito that can only be explained as the result of a series of ruptures that took place in the span of 30 years that separate them. See also Rosier-Catach, ‘Grammar’ and ‘Pré-modisme’. Martin of Dacia, Modi significandi, (ed. Roos, 5:10–12): ‘Ulterius intellectus volens alii conceptum suum significare, rei intellectae vocem imponit, ut eius conceptus scilicet res intellecta per vocem tamquam per signum exprimatur.’ In: Martini de Dacia opera, ed. H. Roos (Copenhagen: gad, 1961) (CPhD ii). For res significata, see I. Rosier-Catach, ‘Res significata et modus significandi: Les implications d’une distinction médiévale’, in Ebbesen (ed.), Sprachtheorien, pp. 135–168. Martin of Dacia, Modi significandi (ed. Roos, 9:22–24): ‛…omnis pars orationis est pars per suum modum significandi essentialem generalem.’

the Significate in Grammar and in Logic

141

ground in properties that are in external things as in a subject – their modes of being.40 In another stage (here represented by Boethius of Dacia), the significate of a word is often presented as the thing itself (res ipsa), which is the content of a concept: Regarding this, one must understand that in his speculation about the things themselves and the modes of being that are their properties, as well as about the modes of understanding, the [artist] is a philosopher. But when he joins the thing itself to an utterance, thus making it the significate of the utterance, making the modes of being be modes of understanding and the modes of understanding be modes of signifying, he starts to be a grammarian…41 Boethius of Dacia, however, makes it explicit that the modes of being are not necessarily real properties of the actual significate of a word. For, on the one 40

41

Martin of Dacia, Modi significandi, (ed. Roos, 4:1–7:12): ‘…sciendum quod modi significandi accepti sunt a modis intelligendi sicut a causa immediata. Quidquid enim contingit intelligere, contingit et significare. Et a modis essendi accepti sunt sicut a causa mediata, quia mediantibus modis intelligendi. Modi autem essendi sunt proprietates rei secundum quod res est extra intellectum. Modi autem intelligendi sunt eaedem proprietates rei secundum quod res est in intellectu et ut eaedem proprietates cum re sunt intellecta. Modi autem significandi eaedem proprietates sunt in numero secundum quod res est significata per vocem. Et ad hoc intelligendum notandum est quod res extra intellectum multas habet proprietates. Habet enim se per modum habitus et quietis et per modum agentis et patientis… Et per istas proprietates distinguuntur res ab invicem. Omnes autem istae proprietates rei extra intellectum existentes dicuntur modi essendi…et post copulationem sive impositionem vocis ipsa res dicitur res significata, et omnes proprietates rei quae prius dicebantur modi essendi rei extra et modi intelligendi rei intellectae, iam dicuntur modi significandi…modi essendi et modi intelligendi et modi significandi sunt idem penitus quod patet ex dictis, differunt tamen accidentaliter…modi essendi sunt in re extra sicut in subiecto; modi autem intelligendi in re intellecta sicut in subiecto et per consequens in intellectu sicut cognitum in cognoscente; modi autem significandi sunt in re significata sicut in subiecto et in voce sicut in signo.’ Cf. Boethius of Dacia, Modi significandi (ed. Pinborg et al., 64:67–65:84). Boethius of Dacia, Modi significandi (ed. Pinborg et al., 7:53–58): ‘Circa hoc intelligendum quod quamdiu ipse est in speculatione ipsarum rerum et modorum essendi, qui eis appropiantur, et modorum intelligendi, ipse est philosophus. Cum autem ipsam rem voci copulat faciendo ipsam vocis significatum, et modos essendi modos intelligendi faciendo, et modos intelligendi modos significandi vocis, iam incipit esse grammaticus…’. (My italics) See also S. Ebbesen, ‘Boethius of Dacia: science is a serious game’, Theoria 66 (2000), 145–158 (repr. in Ebbesen, Collected Essays, vol. 2, pp. 153–162).

142

chapter 4

hand, the significate – the thing itself – can have a mode of being without its word actually having the corresponding mode of signifying; on the other hand, a word can have a mode of signifying without its significate actually having the corresponding mode of being. Although the different grammatical modalities are grounded in some modes of being of things in the external world, they are not necessarily grounded in real properties of the significate of a given word. Hence, for Boethius the modes of being, understanding and signifying are at best similar, but not identical, as for Martin of Dacia: To this, it must be said that the modes of being, understanding and signifying are not completely the same, because then, immediately with the mode of being of a thing, there would be the [corresponding] mode of signifying in the word of that thing, which is false. However, the mode of signifying is taken from the likeness with a mode of understanding, and the mode of understanding from the likeness with a mode of being. Hence, it is not necessary that things, one of which is taken from its likeness with another, are completely the same.42 Accordingly, for Boethius of Dacia the modes of signifying are properties of the word and not of the significate (they are in the word as in a subject), because to signify in one way or another is a property of the word as a grammatical item. However, to be signified in one way or another is a property of the thing insofar as it is the significate of the word, and hence to be signified in one way or another is in the thing qua significate as in a subject: It must be said that the modes of signifying are in the word as in a subject, because to be signified in one way or another is to be signified with one mode of signifying or another. To be signified with one mode of signifying or another is in the thing signified as in a subject. But there is a thing signified only insofar as it is in relation with some word. Whence, the modes of signifying are in the word as in a subject.43 42

43

Boethius of Dacia, Modi significandi (ed. Pinborg et al., 81:22–28): ‘Dicendum ad hoc, quod modi essendi et intelligendi et significandi non sunt idem penitus, quia tunc, statim cum esset modus essendi rei, statim esset modus significandi in dictione illius rei, quod falsum est. Tamen modus significandi accipitur ad similitudinem modi intelligendi et modus intelligendi ad similitudinem modi essendi. Unde non oportet quod illa sunt idem penitus, quorum unum accipitur ad similitudinem alterius.’ Boethius of Dacia, Modi significandi (ed. Pinborg et al., 85:35–42): ‘Dicendum quod modi significandi sunt in dictione sicut in subiecto, quia sic significari vel aliter est tali modo significandi significari vel alio. Tali modo significandi significari vel alio est in re significata

the Significate in Grammar and in Logic

143

Consequently, from Martin to Boethius there is an important shift from the consideration of the significate as the bearer of modes of understanding and signifying to the consideration of the word as the bearer of modes of signifying that result from different cognitive approaches to the thing itself. Boethius of Dacia can thus explain how it is that different parts of speech have the same significate without such a significate being at the same time the bearer of opposite features: The same concept of the mind can be the significate of any part of speech. In fact, whatever can be conceived by the mind can be signified by any part of speech, as long as the specific mode of signifying of the part is not incompatible with it; and that concept of the mind, when it falls under the specific mode of signifying of the name produces the significate of the name; and when it falls under the specific mode of signifying of the verb produces the significate of the verb, and so on; which is evident, because when someone says: ‘pain, to be in pain, the one in pain, painfully and ouch’, they all signify the same.44 In the last stage (here represented by Radulphus Brito), the word is analysed into utterance, significate (the thing itself45) and relation of signification.46

44

45

46

tamquam in subiecto. Res autem significata non est nisi alicuius dictionis. Ideo modi significandi sunt in dictione sicut in subiecto.’ Boethius of Dacia, Modi significandi (ed. Pinborg et al., 55:60–66): ‘…idem conceptus mentis potest esse significatum cuiuslibet partis orationis. Quicquid enim a mente concipi potest, hoc potest per quamlibet partem orationis significari, dummodo modus significandi specificus partium illi non repugnet; et ille mentis conceptus cadens sub modo significandi specifico nominis facit significatum nominis, et cadens sub modo specifico verbi facit significatum verbi et sic de aliis ut patet dicendo sic “dolor, doleo, dolens, dolenter et heu”, quae omnia idem significant.’ Radulphus Brito, Sup. Prisc. min. (ed. Pinborg, 99): ‘Sed voces imponuntur ad significandum rem sub ratione propria illius rei, sicut haec vox “homo” significat hominem sub ratione propria hominis et sic de aliis, et ideo quilibet artifex in sua scientia secundum quod indigebat rebus imposuit voces ad significandum illas res.’ A relational consideration of the significate appears already explicitly in John of Dacia’s analysis of the word into utterance, significate (res significata) and relation of signification (ratio significandi). In his Summa [John of Dacia, Summa grammatica, in Johannis Daci Opera, ed. A. Otto (Copenhagen: gad, 1955) (CPhD i–ii)] John defines the ratio significandi as that whereby the word is related to its significate [John of Dacia, Summa (ed. Otto, 195:33–196:5)]: ‘Ad hanc quaestionem dicendum est quod ratio significandi est illud quo refertur vox significativa ad significatum. Vox autem est quod profertur. Omne autem id quod profertur motivum est sensus auditus. Sed ratio significandi non movet auditum

144

chapter 4

The introduction of the relation of signification into the analysis of the word leads to a more refined analysis of the part of speech that follows from the introduction of a further distinction between active and passive modes of signifying: One is the active and another the passive mode of signifying. The mode of being of the thing insofar as it is consignificated by the utterance is called ‘passive mode of signifying’. But the relation of consignifying, whereby the utterance consignifies that mode of being, is called ‘active mode of signifying’. For the utterance is in itself in the category of quality, and whence the utterance in itself is related neither to a significate nor to a property of the thing, if it were not for some relation of signifying or consignifying that was given to it.47 External things and their properties can be considered either as entities in themselves or as the terms of relations of co-understanding and of consignification. External things and their accidental properties are res extra and modi

47

sed intellectum. Vox autem ipsa movet auditum. Ergo ratio significandi non profertur simul cum ipsa voce.’ John analyses the word into utterance, significate and relation of signification, because the mere utterance of a word does not produce the understanding of its significate, unless the listener establishes a link – a relation of signification – between the utterance affecting her ears and the concept. The union of only utterance and significate cannot explain the communicative aspect of words – how a listener comes to evoke the exact concept that the speaker intends to communicate to her – and therefore the relation of signification that both speaker and listener establish in an act of communication becomes an essential component of the word. Now, the relation of signification is not something really added to the utterance, but it is rather a relation of reason [John of Dacia, Summa (ed. Otto, 197:21–36)]: ‘Ad istam quaestionem dico quod ratio significandi non est aliquid additum reale essentiae vocis. Ad cuius intellectum est notandum quod vox per essentiam suam non refertur ad significatum, quia vox per essentiam suam est de praedicamento qualitatis. Vox igitur per aliquod additum refertur ad significatum, sed illud additum est ratio significandi. Haec autem ratio significandi non profertur realiter cum ipsa voce, ut prehabitum est. Ergo illa ratio significandi solum est aliquid additum essentiae vocis secundum rationem, unde huiusmodi additio nihil est vocis nisi secundum quod cadit sub usu intellectus accipientis significationem vocis.’ Radulphus Brito, Sup. Prisc. Min. (ed. Pinborg, 152): ‘…quidam est modus significandi activus et quidam passivus. Modus significandi passivus dicitur modus essendi rei ut consignificatur per vocem. Sed modus significandi activus dicitur ratio consignificandi per quam vox consignificat illum modum essendi. Nam vox de se est in genere qualitatis, et ideo vox de se non refertur ad significatum nec ad proprietatem rei, nisi per aliquam rationem significandi et consignificandi sibi traditam.’

the Significate in Grammar and in Logic

145

essendi, respectively. As the terms of the relations of co-understanding, these modi essendi are passive modes of understanding. Similarly, as the terms of the relations of consignification, they are passive modes of signifying. Modes of being and passive modes of understanding and of signifying are in this way numerically the same but notionally different. On the other hand, the relation of consignification between the utterance and the properties of the thing and the relation of co-understanding between the intellect and the properties of the thing are active modes of signifying and of understanding, respectively. The active modes of signifying and of understanding are both notionally and numerically different; for the former are in the utterance as in a subject and the latter are in the intellect as in a subject.48 Finally, the passive and active modes of signifying and the passive and active modes of understanding are notionally identical but numerically different.49 Thus, the principles of the division of parts of speech and of grammatical constructions – the active modes of signifying – are properties of the parts of speech themselves and not of their significates, although they are immediately grounded in properties of things in the external world. To sum up, the modal analysis of signification that is already suggested in some grammatical treatises from the first half of the century (e.g. Ps.-Kilwardby’s) reaches a high level of sophistication during the second half of the century, with the three following points of development being of particular importance: (a) The significate of a word becomes the thing itself (instead of a concept) without the accidental features of existence and non-existence. 48

49

Radulphus Brito, Sup. Prisc. min. (ed. Pinborg, 152–156): ‘…modus significandi activus et passivus sunt idem formaliter licet differant materialiter…modus significandi passivus et modus intelligendi passivus et modus essendi sint idem materialiter. Secundo quod non sint idem formaliter. Tertio quod modus significandi et intelligendi activi non sint idem cum modis essendi…modus significandi passivus est ipsamet proprietas rei ut consignificata est per vocem. Et modus intelligendi passivus est ipsa proprietas rei ut apprehensa est ab intellectu. Et modus essendi est ipsa proprietas rei extra absolute sumpta… Ergo modi significandi et intelligendi passivi quantum ad id quod est ibi materiale, scilicet secundum ipsam proprietatem, sunt idem cum modo essendi… Sed modus significandi activus est relatio per quam haec vox habet rationem consignificandi respectu modi essendi, et modus intelligendi activus est ratio cointelligendi per quam intellectus refertur ad rei proprietatem. Et ergo modus significandi et intelligendi active dictus non est idem cum modis essendi.’ Radulphus Brito, Sup. Prisc. min. (ed. Pinborg, 157): ‘Dicendum ad quaestionem primo quod modi significandi et intelligendi activi differunt secundum essentiales rationes… Primum probatur: quia quorum unum est forma intellectus et aliud forma partis orationis illa differunt, quia sunt formae diversorum… Tamen credo quod modus significandi activus et passivus non differunt formaliter sed differunt materialiter.’

146

chapter 4

(b) The word is analysed into utterance, significate and relation of signification (instead of into utterance and significate). (c) The active modes of signifying become properties of the words (instead of properties of the significate), which explain both the grammatical division of parts of speech and the grammaticality of sentences. As regards the question of the significate’s role in grammar and logic, the rupture between the pre-modist tradition and modist authors from the last quarter of the century, such as Radulphus Brito, is substantial. Martin of Dacia already mildly opposes earlier authors with respect to the question of the role of signification in logic and grammar. In his commentary on the Perihermeneias,50 Martin raises the question whether the logician is concerned with truth and falsity. He then puts forth the already familiar argument that just as the grammarian is concerned with grammaticality and ungrammaticality, the logician is concerned with truth and falsity; for the logician analyses the subject and the predicate, which are the causes of truth and falsity: Just as the grammarian is concerned with grammaticality and ungrammaticality, in the same way the logician is concerned with truth and falsity; and the reason is evident. Likewise, whoever considers the subject and the predicate, whose composition or division is the basis of what is true and what is false, also considers what is true and what is false; but the logician considers the subject and the predicate; therefore etc.51 The argument against it states that it is possible to analyse the ‘methods’ of knowledge (literally ‘the way of knowing’ – modus sciendi) provided by logic without an analysis of truth and falsity: The logician ought to consider only the methods of knowledge. Hence, the logician can stand without the things without which the method of knowledge can stand. But the method of knowledge can stand without 50 51

Martin of Dacia, In Perih. (ed. Roos, 248:14–15): ‘Quaestio est utrum pertineat ad logicum considerare veritatem et falsitatem.’ Martin of Dacia, In Perih. (ed. Roos, 248:29–35): ‛…sicut se habet congruum et incongruum ad grammaticum, sic verum et falsum ad logicum. Et patet ratio. Item. Cuius est considerare subiectum et praedicatum, in quorum compositione vel divisione consistit verum et falsum, eius est considerare verum vel falsum. Sed logici est considerare subiectum et praedicatum. Ergo et cetera.’

the Significate in Grammar and in Logic

147

any truth in the specific terms. Therefore, the logician could also stand without truth and falsity.52 Martin’s introduction to the question evinces an important shift as regards the subject matter of logic between the first half of the century and the last quarter of the century. While in the first half of the century it is common to claim that logic deals with truth and falsity (largely taken, so as to include syllogistic validity), most authors from the last quarter of the century maintain that logic deals with logical ‘methods’ (i.e. demonstration, definition etc.) and their principles (i.e. logical notions or logical intentions).53 This is the position of e.g. Radulphus Brito, who tells us that: Logic must be divided according to the division of its main subject. The subject of logic is the method of knowledge according to which [logic] is an instrument of knowledge…and since the syllogism is the method of knowledge to which every method of knowledge is subordinated, hence logic itself is divided according to the division of this main method of knowledge – the syllogism – which is said to be the subject of logic as a whole.54 Logic is concerned with the syllogism insofar as it is the main logical instrument used by other logical methods of knowledge production (i.e. definition, division, demonstration etc.). For Brito, an analysis of the syllogism demands an evaluation of its capacity to produce probable or certain knowledge,55 52

53 54

55

Martin of Dacia, In Perih. (ed. Roos, 248:24–28): ‘Item: logici tantummodo considerare modum sciendi. Igitur sine quo potest stare modus sciendi, sine illo potest stare logicus. Sed modus sciendi potest stare sine omni veritate in terminis specialibus. Ergo et logicus poterit stare sine veritate et falsitate.’ For the different logical traditions of the thirteenth century, see J. Brumberg-Chaumont (ed.), L’Organon. Radulphus Brito, In Isag. (ed. Venice, 6): ‘…logica dividenda est secundum divisionem sui subiecti principalis. Subiectum autem logicae est modus sciendi secundum quod est instrumentum sciendi…et quia omnis modus sciendi ad quem ordinantur omnes alii modi sciendi est sillogismus, ideo secundum divisionem istius modi sciendi principalis qui est sillogismus qui dicitur esse subiectum in tota logica et ipsa logica dividatur.’ Radulphus Brito, In Isag. (ed. Venice, 6–7): ‘Sillogismus autem potest considerari dupli­ citer vel in se et absolute, et sic de ipso est scientia libri Priorum, vel secundum quod habet esse in partibus subiectis. Et hoc dicit quia partes sillogismi subiectiv sunt per constructionem eius ad materiam specialem. Modo ista materia specialis est duplex, quia vel materia necessaria vel materia probabilis. Et ideo duae sunt partes sillogismi

148

chapter 4

which in turn demands an analysis of the items that constitute a syllogism capable of producing these sorts of knowledge, i.e. the premises. Since gene­ rally speaking a premise is an assertion, the analysis of the assertion in the Perihermeneias would be made from the perspective of its possibility to become the premise of a dialectical or demonstrative syllogism, which suppo­ sedly could happen without any consideration of its truth or falsity. Brito’s move is indeed legitimate if one considers that an analysis of the syllogism could presuppose its premises to be probable or certain without being concerned with the actual determination of their truth-value. The situation, which is actually quite common in mathematical logic, would go like this: Suppose that you have two premises that have a certain alethic modality (ne­cessary or probable), or even a determinate truth-value (true), and proceed to determine what can validly be inferred from them. In such an analysis, truth-determination is not, stricto sensu, part of the logician’s job. This, however, does not prevent other logical analyses of the syllogism from being also legitimately concerned with general accounts of propositional verification. As to this matter, Martin takes a position that seems to go in the direction of the latter possibility; for, according to him, logic is indeed concerned with how the three terms of a syllogism contribute to the truth or falsity of its premises, as well as to its validity: To this question, I say three things: First, the logician ought to consider what is true and false as regards his own terms. The proof of this is that whoever must consider the terms, must also consider their truth and falsity. But only the logician ought to consider the terms that are proper to logic. For he is the only one concerned with the consideration of the three terms that form the syllogism. Hence he is the only one concerned with the truth and falsity that results from the arrangement of these terms with each other.56

56

subiectivae, quia si sillogismus contrahitur ad materiam necessariam sicut sillogismus demonstrativus de quo tanquam de subiecto accipiendo subiectum pro obiecto determinatur in libro Posteriorum. Si autem contrahitur ad materiam probabilem, sic est sillogismus dialecticus de quo determinatur in libro Topicorum.’ Note that although Nicholas of Paris tells us that logic is concerned with truth and falsity, he agrees with Brito that the Perihermeneias deals with assertions that are to become premises of dialectical or demonstrative syllogisms. For this, see my case study in Ebbesen et al., History of Philosophy, pp. 131–147. Martin of Dacia, In Perih. (ed. Roos, 249:1–9): ‘Ad istam quaestionem dico tria: primum est, quod logicus in terminis propriis proprie habet considerare verum et falsum. Declaratio huius est: nam cuius est considerare terminos, eius est considerare veritatem

the Significate in Grammar and in Logic

149

Moreover, logic is also concerned with what is true and what is false in general, i.e. with the general principles of truth and falsity: Second, I say that, other than his proper terms, the logician ought to consider what is true and what is false according to the common notion of what is true and what is false. This is evident from Aristotle’s Topics, where he says that logic has the way towards the principles of all methods. But this is not through proper principles. Therefore, it is through common principles.57 Accordingly, logic is concerned with the general causes of truth and falsity, knowledge of which is fundamental to determining truth in other disciplines (e.g. in physics, psychology etc.). Furthermore, these general causes are supposedly related to the assertion’s accurate representation of the mental acts of composition and division; for Martin replies in the affirmative to the question whether truth and falsity come about through composition and division: To the first [objection], I say that truth and falsity come about in every composition and division, as long as there is affirmation or negation.58 This is to say that, generally speaking, an assertion that accurately represents a mental act of composition or division (i.e. by being the compound of a name and an attribute, which are not equivocal) is susceptible of truth or falsity.59 Nevertheless, logic is not concerned with the truth and falsity of assertions within particular disciplines (e.g. physics); for the truth and falsity of such assertions come directly from the objects that are proper to those disciplines and that are unknown to the logician qua logician. The mathematician, the

57

58

59

et falsitatem in eis. Sed solius logici est considerare terminos proprios logices. Nam ad ipsum solum spectat considerare tres terminos, ex quibus constituitur syllogismus, et ideo ad ipsum solum spectat considerare veritatem et falsitatem, quae resultant ex collectione illorum terminorum ad invicem.’ Martin of Dacia, In Perih. (ed. Roos, 249:13–18): ‘Secundo dico quod logicus extra terminos proprios habet considerare verum et falsum secundum communem rationem veri et falsi. Declaratio huius patet per Aristotelem in Topicis, qui dicit quod logica ad omnia methodorum principia viam habet. Sed hoc non est per principia propria. Ergo per principia communia.’ Martin of Dacia, In Perih. (ed. Roos, 250:21–23): ‘Dico ergo ad primam quod circa omnem compositionem vel divisionem est veritas et falsitas, dummodo ibi fit assertio vel deassertio.’ Cf. above, Part 1, Section 1.1.

150

chapter 4

physicist and the metaphysician, for instance, analyse mathematical, physical and metaphysical objects, respectively, in order to determine the truth or falsity of their assertions. However, this analysis does not belong to the logician per se: Third, I say that the logician ought not consider what is true and what is false according to their proper notions outside his proper terms. The proof is that truth and falsity are caused by things as their proper and immediate principles. But the logician ought not consider the things themselves, nor their principles… Hence, it must be known that when someone asks whether a sentence is grammatical or ungrammatical, this concerns the grammarian; whether it is true or false, this concerns the logician. Likewise, if someone asks whether the line has magnitude, this regards the mathematician. And thus it is evident that any artist ought to consider truth and falsity as regards his own terms.60 60

Martin of Dacia, In Perih. (ed. Roos, 249:18–29): ‘Tertio dico quod logicus extra proprios terminos non habet considerare verum et falsum secundum proprias rationes eorum. Cuius declaratio est: nam veritas et falsitas causantur ex rebus tamquam ex propriis et immediatis principiis. Sed logici non est considerare res ipsas nec principia rerum… Unde sciendum est quod si quaeratur utrum oratio sit congrua vel incongrua, hoc pertinet ad grammaticum; si vera vel falsa, ad logicum. Similiter si quaeratur utrum linea habeat magnitudinem, spectat ad mathematicum. Et sic patet quod in propriis terminis quilibet artifex proprie habet considerare veritatem et falsitatem.’ See also Boethius of Dacia [Boethius of Dacia, Topica, ed. N.J. Green-Pedersen (Copenhagen: gad, 1976) (CPhD vi)], for whom the immediate causes of logic are not things but their properties – their modes of being (cf. Boethius of Dacia, Topica (ed. Green-Pedersen, 3:14–4:24)). Furthermore, things themselves are the only causes of true and false assertions (cf. Boethius of Dacia, Topica (ed. Green-Pedersen, 6:85–87:108)). And finally, dialectics – the part of logic that deals with dialectical argumention – is not concerned with the truth that is involved in composition and division, because the origin of truth is in the things themselves and these things do not fall under the scope of the dialectician [Boethius of Dacia, Topica (ed. Green-Pedersen, 8:135–149)]: ‘Et quia in praesenti intentio est de dialectica, ideo scire debes primo, quod illa, per quae argumentatur dialecticus, non sunt causa conclusionis; ideo non faciunt scientiam, sed opinionem. Licet enim de necessitate sequitur “Socrates est homo, ergo est animal”, tamen habitudo speciei ad genus et intentio communis non est causa huius necessitatis, sed identitas rerum essentialis. Similiter licet necessario sequitur “Socrates est albus, ergo non est niger”, tamen communis intentio et ipsa habitudo contrarii ad contrarium non est causa huius, sed ipsa incompossibilitas rerum, quae per terminos significantur, quae multum alia est ab ipsa habitudine, quam dialecticos considerat. Secundo debes scire quod dialecticus ipsam veritatem non considerat. Et causa huius est, quia ipsa veritas ortum habet ex ipsis naturis rerum, quas dialecticus

the Significate in Grammar and in Logic

151

Radulphus Brito’s position as to the role of the significate both in logic and in grammar is clearly along this line of reasoning. In his commentary on Priscian minor, he raises the questions whether the grammarian considers the significate in itself,61 whether the mode of signifying presupposes the significate62 and whether the significate is the cause of the construction (i.e. of sentences).63 To the first question, Brito replies that in the case of the institution of words (in the case of grammatica positiva), the grammarian has to consider things, not according to their real existence, but insofar as they are the object of the act of linguistic imposition: For the solution to this question, we must distinguish three sorts of grammar – impositive grammar, practical grammar and normative grammar. Impositive grammar is the one [dealing] with the imposition of utterances on the specific significates and it teaches the essence of the name, i.e. of appellations. Then, I say that with respect to impositive grammar the grammarian ought to consider things, however not according to their real existence, but according to their being significates – insofar as things are the significate of words.64 The part of grammar that explains how significative words are instituted by the imposition of utterances on things – impositive grammar – must consider things, not insofar as they exist in external reality, but insofar as they can be known in themselves (without their existence or non-existence) and be, thereby, an object of imposition and signification. secundum quod dialecticus non considerat… Sed (i.e. dialecticus) docet modum, quam artifex specialis debet materiae speciali applicare ad inquisitionem veritatis.’ 61 Radulphus Brito, In Prisc. min. (ed. Pinborg, 136): ‘Utrum grammaticus consideret significatum speciale per se.’ 62 Radulphus Brito, In Prisc. min. (ed. Pinborg, 138): ‘Utrum modus significandi praesupponat significatum speciale.’ 63 Radulphus Brito, In Prisc. min. (ed. Pinborg, 140): ‘Utrum significatum sit causa constructionis.’ 64 Radulphus Brito, In Prisc. min. (ed. Pinborg, 137): ‘Ad solutionem istius quaestionis distinguendum est de triplici grammatica, scilicet positiva, usualis et regularis. Positiva est quae est de impositione vocum ad significata specialia et docet quid nominis sive vocabulorum… Tunc dico quod quantum ad positivam grammaticam grammaticus habet considerare res, non tamen secundum esse reale sed secundum esse significatum, scilicet in quantum res sunt significatum dictionis.’ See Rosier-Catach, ‘Grammar’, at p. 205, for the introduction of this distinction in Dominicus Gundissalinus’ De divisione philosophiae. Recall that for Brito the words’ significates are the things themselves without consideration of their existence or non-existence.

152

chapter 4

However, in the case of theoretical grammar, the grammarian is not concerned with the things themselves: But if one asks about theoretical grammar, which deals with causes and principles, then I say that [the grammarian] ought to consider things, but neither per se nor principally. And the reason is that the grammarian ought to consider that without which there cannot be knowledge of the modes of signifying. But the modes of signifying are not known without the properties of things. Therefore, the grammarian ought to consider the properties of things, and the thing, yet not principally but as a support, just as the logician ought not consider the things per se, but insofar as he grounds in them the second intentions.65 The grammarian has to look at the properties of things – the modes of being – in which the modes of signifying are ultimately grounded, and thus things enter into the analysis of theoretical grammar only insofar as they are the bearers of modes of being. More importantly, and contrary to what was the case in the first half of the century, the logician considers the things themselves only insofar as they give support to the principles of logic – second intentions. Accordingly, the significate is notionally presupposed by the part of speech, because it is a part of the word that bears the modes of signifying, but the significate is not the immediate cause of grammatical constructions: To which it must be said that every mode of signifying presupposes the significate, also if the significate and the mode of signifying were the same thing. And the reason is that the significate is related to the mode of signifying, just as the word is related to the part of speech; but the part of speech always presupposes the word; therefore the mode of signifying always presupposes the significate. The major [premise] is evident, because the mode of signifying is the form of the part [of speech], just as the significate is the form of the word; and thus the form of the word presupposes the form of the part of speech. However, note that this 65

Radulphus Brito, In Prisc. min. (ed. Pinborg, 137): ‘Si vero quaerat de grammatica speculativa quae est per causas et principia, tunc dico quod habet considerare res quamvis non per se nec principaliter. Et huius ratio est, quia illa habet grammaticus considerare sine quibus cognitio modorum significandi non potest haberi; sed modi significandi sine proprietatibus rerum non cognoscuntur; ergo grammaticus habet considerare proprietates rerum et res, non tamen ex principali sed ex adiuncto, sicut nec logicus habet considerare res per se sed prout fundat super ipsas intentiones secundas.’

the Significate in Grammar and in Logic

153

anteriority is not necessarily temporal, but natural, because it must be in art, just as it is in nature.66 The significate is notionally presupposed by the modes of signifying; for without significate there is no word, and without word there is no part of speech. However, strictly speaking, a part of speech is the conjunction of a word and a mode of signifying, so that the significate does not enter immediately into the analysis of the part of speech, and thus does not fall per se under the grammarian’s consideration. Likewise, things are notionally presupposed by second intentions, without being their immediate cause: But it must be understood that, just as the logician does not consider things, unless by accident – as the common intentions are grounded in them – neither does the grammarian.67 Just as in the case of grammatical enquiries, the analysis of the significate per se no longer concerns the logical enquiry. Furthermore, while the principles of grammar – the modes of signifying – are the immediate causes of grammatical sentences, and accordingly the grammarian is directly concerned with grammaticality, the principles of logic – second intentions – are not the cause of truth and falsity, and accordingly the logician qua logician is not concerned with truth or falsity: To the other [position]: ‘Just as the logician deals with the truth of the propositions etc.’, I say that it is not similar, because the cause of the truth of some proposition is not a second intention, but the cause of the grammaticality of any construction is the proportion of modes of signifying. And hence, although logic cannot judge the truth of a specific sentence 66

67

Radulphus Brito, In Prisc. min. (ed. Pinborg, 138–139): ‘Ad quod dicendum quod omnis modus significandi praesupponit significatum, etiam si significatum et modus significandi fuerint idem. Et huius ratio est, quia sicut se habet dictio ad partem orationis, ita significatum ad modum significandi. Sed pars orationis semper praesupponit dictionem, ergo modus significandi semper praesupponit significatum. Maior patet quia sicut significatum est forma dictionis, sic modus significandi est forma partis; et sic forma dictionis praesupponitur formae partis. Tamen nota quod haec praecedentia non est de necessitate in tempore, sed secundum naturam, quia sicut est in natura, sic debet esse in arte.’ Radulphus Brito, In Prisc. min. (ed. Pinborg, 104): ‘Sed est intelligendum quod sicut logicus non considerat res nisi per accidens, ut scilicet super eas fundantur intentiones secundae, sic etiam grammaticus…’.

154

chapter 4

through its second intentions, the grammarian, however, through his modes of signifying can judge grammaticality, because the truth of specific propositions is caused by the identity of the thing, just as in ‘man is animal’, and not because the one can be found in several things numerically different, and the other can be found in several things specifically different.68 The truth of an assertion, such as ‘some man is an animal’, depends on a real identity between a particular man and a particular animal in some real man. However, the fact that ‘man’ is a species or that ‘animal’ is a genus, or that ‘man’ is related to ‘animal’ as a species to its genus, or that ‘animal’ is a universal, and so on, plays no role in the verification of ‘some man is an animal’. Yet, these facts do play a role in the validity of syllogisms that have ‘some man is an animal’ as a premise, as in: ‘Every animal has a soul; some man is an animal; therefore some man has a soul’, which is valid precisely because of the relations that ‘man’, ‘having soul’ and ‘animal’ have with each other insofar as they are subject to second intentions. Second intentions are the principles of the logical methods that use the syllogism as their main instrument, and hence logic is concerned with these principles and not with propositional verification. Since according to Brito the active modes of signifying – the ones that make a word a particular grammatical item – are properties of the part of speech and not of the word qua bearer of a relation of signification, it is clear that semiotic considerations can and should be absent from the grammatical analysis of sentences. For this analysis is carried on through properties of the parts themselves and not through properties of their semiotic value. Admittedly, the active modes of signifying are ultimately grounded in the properties of things. However, they are not immediately grounded in them, nor are they identical to them. More importantly, semiotic considerations are also absent from the analysis of logical methods, such as demonstration, because such analysis depends on the relations between second intentions. Consequently, the verification of an assertion 68

Radulphus Brito, In Prisc. min. (ed. Pinborg, 129–130): ‘Ad aliam “sicut se habet logicus ad veritatem in propositionibus etc.”, dico quod non est simile quia causa veritatis alicuius propositionis non est aliqua secunda intentio, sed causa congruitatis cuiuslibet constructionis est proportio modorum significandi; et ideo licet logica per suas intentiones secundas non possit iudicare veritatem orationis specialis, tamen grammaticus per suos modos significandi potest iudicare congruitatem, quia veritas in propositionibus specialibus causatur ex parte identitatis rei sicut in ista “homo est animal” et non ex eo quod unum est reperibile in pluribus differentibus numero et aliud in pluribus differentibus specie. Sed congruitas causatur ex modis significandi.’

the Significate in Grammar and in Logic

155

regards the particular discipline that is concerned with the assertion and is not as such a logical concern. It is clear, then, that in the span of years separating Brito from Ps.-Kilwardby, the notion of signification goes from being directly used in both the explanations of grammaticality and truth or falsity to being used only in the part of grammar that is concerned with the institution of words, i.e. impositive grammar – a development coherent with the one we saw taking place in the two questions discussed in Part 1. This grammatical discussion further demonstrates the considerably narrower role the notion of signification plays in explaining both the construction of grammatical sentences and the mechanisms of propositional verification. Indeed, the determination of truth-values is no longer considered a logical matter per se, as the analysis of the methods of knowledge production can be effectuated without a direct consideration of the words’ values. § There are important points of divergence regarding the role of signification in grammar and logic between the grammatical and logical treatises from the first half of the thirteenth century and those from the last quarter of the century. In the years before the 1260s, a word results from the articulation of an utterance and a significate. Some properties of the significate are the sources of truth and falsity, and some others are the source of grammaticality and ungrammaticality. Towards the end of the century, the word is analysed into utterance, significate and relation of signification. For some authors this relation links an utterance and a thing understood, the modes of signifying being properties of the thing understood (e.g. in Martin of Dacia). Martin of Dacia also suggests that the significate is the general principle of truth and falsity, and thereby he still follows in the footsteps of his predecessors. Some later authors (e.g. Radulphus Brito and Boethius of Dacia) claim that the relation of signification links the utterance and the thing itself, the modes of signifying being a grammatical feature of the utterance that is ultimately grounded in properties – modes of being – of the thing. Finally, for some of them (certainly for Radulphus Brito and quite likely for Boethius of Dacia), the relation of signification and the significate are neither principles of theoretical grammar nor principles of logic, but only the material cause of these principles, insofar as they provide the matter – the word – that receives the modes of signifying and the second intentions as forms.69 This, I submit, can be seen as yet another 69

Note, however, that this position seems to be anticipated by Ps.-Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 24): ‘…logica non est de sermone relato ad rem tantum, sed de sermone

156

chapter 4

attempt by these masters of Arts to move the notion of signification to a domain that is grammatically and logically neutral, namely that of semiotics, a domain presupposed by grammatical and logical explanations but not directly used in their proper accounts, i.e. the accounts of grammaticality and of the methods of knowledge production, respectively. relato ad rem sub debitis modis significandi. Hos modos considerat grammaticus per se et primo’. Nevertheless for this author the determination of an assertions’ truth-value depends on logical principles; cf. Ps. Kilwardby, Sup. Prisc. mai. (ed. Fredborg et al., 26).

Conclusion The study of the three discussions that are directly concerned with the notion of signification – the one about the immediate significate of words, the one about the signification of empty terms and the one about the different roles played by the significate in logic and grammar – allows us to draw conclusions at the doctrinal, methodological and sociological levels. From a doctrinal perspective, the thirteenth-century tendency towards a narrower and more coherent use of the notion of signification merits emphasis. This development complements the intent to solidify the scientific status of the disciplines of the trivium – particularly of logic and grammar – by providing them with mutually exclusive subject matters and explanatory principles based on external reality. Some masters of Arts from the last quarter of the century show this inclination most clearly by keeping the notion of signification essential to the explanation of the institution of linguistic signs and reducing its role in purely logical and grammatical accounts. Both Boethius of Dacia and Radulphus Brito outrank their contemporaries by their remarkable awareness of the ambiguous use of the notion of signification (it was indeed used in linguistic problems of different nature), and their subsequent attempt to overcome this difficulty by narrowing its use to the mere semiotic level. Roger Bacon also shows a similar awareness, which leads him to introduce the subsidiary notion of re-imposition. As a result, Bacon deals with a coherent notion of signification that is instrumental at different levels. However, as we have seen, the introduction of the notion of re-imposition by Bacon has the unfortunate consequence of compromising the possibility of accounting for the efficacy of human communication. The situation we are left with at the turn of the fourteenth-century can be summed up as follows: i) A word is instituted through the somewhat arbitrary imposition of an utterance on the essence itself, with the purpose of human communication, thereby acquiring its relation of signification. ii) Grammar is mainly concerned with grammatical constructions and their fundamental explanatory principles – the modes of signifying (modi significandi). iii) Logic is mainly concerned with the methods of knowledge production and their fundamental explanatory principles – the second intentions (secundae intentiones). iv) Both the modes of signifying and the second intentions presuppose the word’s signification (i.e. they presuppose the word as their material part) but are not identical to it, so that both grammar and logic presuppose a semiotic account of the institution of linguistic signs, but are theoretically independent

© koninklijke brill nv, leiden, ���� | doi 10.1163/9789004300132_007

158

CONCLUSION

from it. Finally, v) logic is not per se concerned with the question of propositional verification.1 The development of these medieval discussions about the notion of signification seems to outweigh, by the attempt to give the notion a more coherent treatment, the twentieth-century discussions about the notion of meaning; for in the latter we are still left with the impression of a reigning confusion as to what the notion of meaning is supposed to account for and how it is to be construed with respect to those problems. Not surprisingly, then, the Stanford Encyclopedia of Philosophy’s entry on theories of meaning starts with the following unenthusiastic remark: The term ‘theory of meaning’ has figured, in one way or another, in a great number of philosophical disputes over the last century. Unfortunately, this term has also been used to mean a great number of different things.2 Nevertheless, the question remains: What was the fortune of this daring thirteenth-century attempt in the fourteenth-century development of grammar and logic? It is already well known that the grammatical doctrine of the modi significandi was severely attacked by some fourteenth-century scholars until it gra­ dually disappeared from the fourteenth-century scholarly picture. Famously, William of Ockham rejects the modes of signifying as superfluous, in that they unnecessarily multiply the number of notions – and of beings – that are needed to account for the grammatical features of human language.3 The doctrine of the modi significandi appears to be incompatible with the commitment to explanatory and ontological austerity of fourteenth-century nominalism; and with the destruction of thirteenth-century theoretical grammar, medieval grammar also seems to fade away as a scientific discipline.4 The thirteenth-century logic of intentions was also at the centre of heated discussions during the first quarter of the fourteenth century. One of the main ideas put to the fore for discussion was the late thirteenth-century claim that

1 For a similar claim in 20th-century philosophy of language, see A.J. Ayer, ‘The Criterion of Truth’, Analysis 3 (1935), 28–32. 2 J. Speaks, ‘Theories of Meaning’, Stanford Encyclopedia of Philosophy, http://plato.stanford .edu/entries/meaning/. 3 Cf. Biard, Logique et théorie du signe, p. 239. 4 For a thorough evaluation of the fourteenth-century rejection of theoretical grammar, see Biard, Logique et théorie du signe, pp. 238–288.

Conclusion

159

logic deals with second intentions.5 Nevertheless, the details of this discussion and its impact on the development of logic between Radulphus Brito and John Buridan are still in need of an exhaustive study of the extant sources. After the work of Jan Pinborg in the 1970s,6 little has been done to fill this void. Thus, it is still difficult to assess the extent to which fourteenth-century nominalism innovated or improved upon late thirteenth-century logic. To be sure, there is some continuity between thirteenth-century logical commentaries and those by John Buridan. For instance, in Buridan’s commentary on Aristotle’s Perihermeneias,7 in his question ‘whether every name signifies something’,8 the Parisian master engages several of the issues discussed in the present study, namely in Part 1. In his discussion, Buridan considers two difficult cases: i) the case of the names of fictional entities, such as ‘chimera’ and ii) the case of a name that happens to lose its extension, such as ‘rose’ when no rose exists. Regarding these two cases, Buridan commits himself to basic tenets of thirteenth-century linguistics, e.g.: i) The significative word is instituted through the conventional act of imposing an utterance on a determinate content – a concept in Buridan’s case – with the purpose of human communication.9 ii) A word, e.g. ‘chimera’, signifies the concept on which it was imposed, and, moreover, through the signification of this concept, the word also signifies the concept’s content (e.g., ‘chimera’ signifies whatever is understood by its eventual listener, i.e. the mental compound of a lion’s head, a bull’s body and a dragon’s tail).10 Also, iii) even when no rose exists, ‘rose’ signifies a simple specific concept falling under the category of substance – the 5

6 7 8 9

10

Cf. Bartholomew of Bruges’ sophisma on the subject matter of logic, edited in S. Ebbesen and J. Pinborg, ‘Bartholomew of Bruges and his Sophisma on the Nature of Logic’, cimagl 39 (1981), 1–80. See, e.g., J. Pinborg, ‘Zum Begriff der Intentio secunda. Radulphus Brito, Hervaeus Natalis und Petrus Aureoli in Diskussion’, cimagl 13 (1974), 49–59. John Buridan, Questiones longe super librum Perihermeneias, ed. R. Van der Lecq (Nijmegen: Ingenium Publishers, 1983) (Artistarium 4). John Buridan, In Perih. (ed. Van der Lecq, 7:28): ‘Utrum omne nomen significat aliquid.’ Cf. John Buridan, In Perih. (ed. Van der Lecq, 8:21–9:8): ‘Notandum est quod omnis vox est significativa… Secundo etiam dicendum est quod omnis vox literalis, vel saltem consimilis, est significativa ad placitum, quia non repugnat quod imponatur ad significandum et quod significet… Notandum est quod tamen de vocibus significativis ad placitum nos vocamus consueta locutione illam vocem significativam quae communitati alicuius ydyomatis imposita est ad significandum, secundum aliquam certam significationem et communiter notam illis de isto ydyomate…’. Cf. John Buridan, In Perih. (ed. Van der Lecq, 10:3–8): ‘Per hoc apparet quod haec vox “chimaera” non solum significat aliquid, immo etiam significat aliquid aliud a se, quia significat conceptum, ut dictum est, qui est alius a voce. Item. Dico ultra quod haec vox

160

CONCLUSION

concept of a rose – and, consequently, signifies roses that no longer exist but that the listener attends to in her mind.11 Furthermore, this signification does not play a central role in the verification of propositions involving those terms, because the proposition’s truthvalue depends on the supposition of its terms and not on their signification. ‘A chimera is understood’ is false, because there are no chimeras for which the term ‘chimera’ can stand. In general: …is false every affirmative proposition in which one of the terms – either the subject or the predicate – does not stand for (supponit pro) one or several things.12 For the same reason, when no rose exists, a similar proposition in the present tense, with ‘rose’ as a subject and without a term that ampliates (i.e. extends) the supposition of ‘rose’ to past roses, is also false. Hence, just as in the logic of intentions the notion of signification does not play a role in propositional verification, in Buridan’s logic it does not seem to play that role either. However, the parallel only goes so far; for, contrary to e.g. Radulphus Brito, Buridan seems to acknowledge that the general principles of propositional verification are a logical problem per se, the property of supposition being the subsidiary notion thereof. It seems, then, that Buridan maintains some of the main tenets of thirteenth-century semiotics, his notion of signification being mainly used

11

12

“chimaera” non solum significat conceptum animae, immo cum hoc aliquid aliud quod isto conceptu concipitur vel aliqua alia quae illo conceptu concipiuntur…’. Cf. John Buridan, In Perih. (ed. Van der Lecq, 12:19–33): ‘Tertia conclusio principalis ponenda est quod rosa significatur per hoc nomen “rosa” quamvis etiam nulla sit rosa, quia hoc nomen “rosa” impositum est ad significandum ad placitum. Ideo significat scienti impositionem et non solum significat seipsam vel conceptum, sicut arguebatur de hoc nomine “chimaera”. Ideo aliquid significat ultra conceptum. Et non aliud quam rosam vel rosas. Non enim posses assignare quid aliud significaret; igitur significat rosas… Sed huic nomini “rosa” correspondet conceptus simplex specificus de praedicamento substantiae. Ideo non potest assignari quod significet aliquid nisi rosam, etsi concessum sit quod hodie nulla sit rosa.’ John Buridan, In Perih. (ed. Van der Lecq, 9:23–28): ‘Et similiter haec est falsa: “chimaera intelligitur”…posito (sicut nos ponimus communiter) quod impossibile sit esse chimaeram…quia omnis propositio affirmativa est falsa in qua aliquis terminorum, scilicet subiectum vel praedicatum, non supponunt pro aliquo vel pro aliquibus. Nec etiam istam regulam intendo nunc probare, sed suppono eam. Modo constat quod iste terminus “chimaera”, qui erat subiectum in dictis propositionibus, pro nullo supponit, si impossibile sit chimaeram esse. Ergo etc.’

Conclusion

161

at this level. However, he puts forth a terministic logic (to the detriment of the late thirteenth-century logic of intentions) that reintroduces the problem of propositional verification into the logical domain. At any rate, it seems important to reiterate that future research on the logical discussions that took place during the first twenty years of the fourteenth century will be most welcome by historians of medieval logic and semantics, the history of fourteenth-century logic from Ockham and Buridan onwards having already been extensively and satisfactorily studied. John Buridan’s use of the property of supposition brings us to the methodological conclusions warranted by this study; for it also justifies excluding an analysis of thirteenth-century terministic logic. Simply put, the development of thirteenth-century logic that took place within the framework of, at least, the Parisian faculty of Arts is largely independent of the parallel development of thirteenth-century terministic logic. Although masters such as Martin of Dacia, Boethius of Dacia and Radulphus Brito are undoubtedly acquainted with the theories of supposition put forth by thirteenth-century scholars such as Peter of Spain and William of Sherwood, they make next to no use of them in their logical discussions, even in the obvious cases where the notion of supposition would have proven to be useful, most notably in the problems of propositional verification. Even more surprisingly, this absence is also the case for the Franciscan theologians Roger Bacon and Peter John Olivi, who are obviously acquainted with the property of supposition but make no use of it in the discussions with which this study was concerned. The doctrinal sources obscure rather than proclaim whether this is due to poor interest in introducing foreign, albeit helpful, theoretical tools into the discussions, or to a rigid system of scholarly practices that did not allow such integration. Be that as it may, the difficulty was overcome by the time of Ockham and Buridan, both of whom impressively articulate thirteenth-century logical tenets with the heretofore poorly used terministic logic.13 On the other hand, the absence of theological themes from the discussions we have examined also explains our exclusion of theological discussions of a linguistic nature, e.g. the one concerning divine nomination. The extent of the interaction between theology, logic and grammar in the thirteenth-century is indeed difficult to assess. However, while a fair number of thirteenth-century theological treatises appear to be influenced by some of the discussions that took place within the faculties of Arts, there is little evidence of a direct influence of the theological discussions on the logical and grammatical ones we 13

Cf. Biard, Logique et théorie du signe.

162

CONCLUSION

have entertained, Augustine’s otherwise highly influential De doctrina christiana included.14 Let me finish with a few words of sociological nature. The chronological consideration of the discussions here analysed also bears witness to the fascinating evolution of the master of Arts’ role in the thirteenth-century intellectual environment. In early masters, such as Nicholas of Paris, we can recognise commentators whose main role is to clarify the content of the authoritative literature, scarcely stirring the philosophical problems these texts raise. By contrast, in later masters like Radulphus Brito we see independent thinkers who fully engage the philosophical problems earlier masters only briefly discussed, thereby climbing considerably beyond the mere clarification of authoritative literature. With the emergence of European universities, the thirteenth century saw logic rapidly acquire scientific independence. This process showed a climax in fourteenth-century nominalism – the missing link of modern research on the history of medieval logic notwithstanding. Thus, the creation of the first European universities in the thirteenth century represents an important point of inflexion in the revival of somewhat independent, speculative thinkers, with a marked interest in well-formulated philosophical problems that, even compared with the sophistication of our modern linguistics, were impressively given solid scientific support and treatment. 14

This would also be a breaking point of fourteenth-century logic with respect to the late thirteenth-century; for, according to Joël Biard, fourteenth-century logic would be largely based on Augustinian semiotics; cf. Biard, Logique et théorie du signe, pp. 9–20.

Bibliography Abbreviations ahdlma al cimagl csel CPhD cup gg gl oup pims pl



Archives d’histoire doctrinale et littéraire du moyen âge Aristoteles latinus Cahiers de l’Institut du Moyen-Âge Grec et Latin Corpus scriptorum ecclesiasticorum latinorum Corpus Philosophorum Danicorum Medii Aevi Cambridge University Press Grecae grammaticae Grammaticae latinae Oxford University Press Pontifical Institute of Medieval Studies Patrologia latina

Primary Sources: Unedited

Anonymous Oxford, Notulae supra librum Periarmenias, MS Oxford, Bodleian Library Can. misc. 403, ff. 31ra–55vb. Boethius of Dacia, Sophisma Omnis homo de necessitate est animal, MS Brugge, Stadsbibliotheek 509, ff. 87v–91v; Firenze, Biblioteca Medicea Laurenziana Pl. 12 sin. 3, ff. 63ra–64rb (unpublished edition by S. Ebbesen). Nicholas of Paris, Rationes super Perihermeneias, MS Vaticano, Biblioteca Apostolica Vat. lat. 3011, ff. 21vb–34vb; cf. MS Münich, Bayerische Staatsbibliothek Clm. 14460, ff. 62ra–100v (anon.) ———, Rationes super Praedicamenta, MS Münich, Bayerische Staatsbibliothek Clm. 14460, ff. 42ra–62ra. Pseudo-Johannes Pagus, Scriptum super librum Perihermeneias, MS Padova, Biblioteca Universitaria 1589, ff. 82r–93r; cf. MS Palermo, Biblioteca Communale 2 Qq.D.142, ff. 37ra–51va (anon.). Radulphus Brito, Quaestiones in Isagogen, in Quaestiones subtilissimae Magistri Rodulphi Britonis super arte veteri, Venice: 1499. Robert Kilwardby, Notulae super Perihermeneias, MSS Cambridge, Peterhouse 206, ff. 13va–21ra; Madrid, Biblioteca Universitaria 73, ff. 44ra–66va; Venezia, Biblioteca Marciana L.VI.66, 2r–18v (partial edition in Lewry, Robert Kilwardby’s Writings).

164

Bibliography

Primary Sources: Edited

Albert the Great, De anima, ed. C. Stroick, OMI, Opera Omnia, Editio Coloniensis 7.1, Münster: Aschendorff, 1968. ———, Liber Perihermeneias, ed. A. Borgnet, Opera Omnia 1, Paris: Vivès, 1890. ———, Super Porphyrium De quinque universalibus, ed. M. Santos Noya, Opera Omnia, Editio Coloniensis 1.1a, Münster: Aschendorff, 2004. Anonymous Oxford, Notulae supra librum Perihermeneias, partial edition in MoraMárquez, ‘Anonymus Oxford’. Anonymus Alani, Sophisma Omnis homo est de necessitate animal, edited in de Libera and Gazziero, ‘Le Sophisma “Omnis homo de necessitate est animal”’. Apollonius Dyscolus, Appolloni Dyscoli scripta minora a Richardo Schneidero edita continens, ed. G. Uhlig, gg II.1, Leipzig: Teubner, 1878. ———, De constructione libri quattuor, ed. G. Uhlig, gg II.2, Leipzib: Teubner, 1910. ———, The Syntax of Apollonius Dyscolus, tr. F.W. Householder, Amsterdam: John Benjamins, 1981. Aristotle, Analytica, ed. W.D. Ross, Oxford: oup, 1964. ———, Categoriae et Liber de interpretatione, ed. L. Minio-Paluello, Oxford, oup, 1949. ———, Categories, On interpretation (ed. and tr. H.P. Cooke) Prior Analytics (ed. and tr. H. Tredennick), Loeb Classical Library 325, Cambridge MA: Harvard University Press, 1938. ———, De interpretatione vel Periermenias: Translatio Boethii, ed. L. Minio-Paluello, al II, 1–2, Bruges/Paris: Desclée de Brouwer, 1965. ———, De l’âme, ed. A. Jannone; tr. E. Barbotin, Paris: Belles Lettres, 1966. ———, Les réfutations sophistiques, introduction, traduction et commentaire par L.A. Dorion, Histoire des doctrines de l’antiquité classique, Paris/Laval: Vrin/Presses de l’Université de Laval, 1995. ———, Metaphysics. A Revised Text with Introduction and Commentary by W.D. Ross, 2. vols., Oxford: oup, special edition 1997. ———, Topica et Sophistici Elenchi, ed. W.D. Ross, Oxford: oup, 1958. ———, Topiques (I–IV), ed. and tr. J. Brunschwig, Paris: Belles Lettres, 1967. Auctores incerti, Quaestiones super Sophisticos Elenchos, ed. S. Ebbesen, CPhD VII, Copenhagen: GAD, 1977. Auctoritates Aristotelis, Les auctoritates Aristotelis. Un florilège médiéval. Etude historique et édition critique, ed. J. Hamesse, Philosophes Médiévaux XVII, Louvain: Publications Universitaires, 1974. Augustine, De magistro, in E. Bermon, La signification et l’enseignement. Le De magistro de Saint Augustin, 2007. ———, De doctrina christiana, in La doctrine chrétienne, Oeuvres de Saint Augustin 11/2, Paris: Institut d’Etudes Augustiniennes, 1997.

Bibliography

165

———, De trinitate VIII–XV, in La trinité. Livres VIII–XV, Oeuvres de Saint Augustine 16, Paris: Institut d’Etudes Augustiniennes, 1955. Boethius, In Categorias Aristotelis libri quattuor, ed. J.-P.Migne, pl 64, Paris: Migne, 1860. ———, Commentarii in librum Aristotelis ΠΕΡΙ ΕΡΜΗΝΕΙΑΣ, ed. C. Meiser, 2 vols., Leipzig: Teubner, 1877–80. ———, In Porphyrii Isagogen commentarium, ed. G. Schepps and S. Brandt, csel XXXXVIII, Vindobonae: F. Tempsky, 1906. ———, De divisione, ed. J. Magee, Philosophia Antiqua 77, Leiden: Brill, 1998. Boethius of Dacia, Modi significandi sive Quaestiones super Priscianum maiorem, ed. J. Pinborg and H. Roos, CPhD IV, Copenhagen: GAD, 1969. ———, Quaestiones super librum Topicorum, ed. N.J. Green-Pedersen, CPhD VI.1, Copenhagen: GAD, 1976. Cicero, De inventione, in De l’invention, ed. and tr. G. Achard, Paris: Les Belles Lettres, 1994. Damascenus, De fide orthodoxa. Versions of Burgundio and Cerbanus, ed. E.M. Buytaert, New York: The Franciscan Institute, 1955. Diogenes Laertius, Vitae philosophorum, 2 vols., ed. H.S. Long, Oxford: Clarendon Press, 1964. Johannes Pagus, Syncategoremata, ed. Braakhuis, De 13de Eeuwse Tractaten, 1979. John Buridan, Questiones longe super librum Perihermeneias, ed. R. Van der Lecq, Artistarium 4, Nijmegen: Ingenium Publishers, 1983. John Duns Scotus, Ordinatio, in Opera omnia I–VII, ed. C. Balic et al., Vatican City: Vatican Polyglot Press, 1950–1973. ———, Quaestiones in duos libros Perihermeneias, ed. R. Andrews et al., Opera philosophica 2, St. Bonaventure, NY: The Franciscan Institute, 2004. John of Dacia, Summa grammatica, in Johannis Daci Opera, ed. A. Otto, CPhD I.1–2, Copenhagen, GAD, 1955. Martin of Dacia, Modi significandi, ed. H. Roos, in Martini de Dacia opera, CPhD II, Copenhagen: GAD, 1961. ———, Quaestiones super Peri hermeneias, ed. H. Roos in Martini de Dacia opera, CPhD II, Copenhagen: GAD, 1961. Michael of Marbais, Summa de modis significandi, ed. L.G. Kelly, Grammatica Speculativa, Stuttgart: Frommann-Hoolzboog, 1995. Nicholas of Paris, Notulae super Perihermeneias, partial edition in H. Hansen and A.M. Mora-Márquez, ‘Nicholas of Paris on Aristotle’s Perihermeneias 1-3’. ———, Syncategoremata, ed. Braakhuis, De 13de Eeuwse Tractaten. Peter of Auvergne, Quaestiones in Peri hermeneias, q. 5–6, partial edition in Ebbesen, ‘Animal est omnis homo’, 150–155. Peter Helias, Summa super Priscianum, ed. L. Reilly, 2 vols., Studies and Texts 113, Toronto: pims, 1993.

166

Bibliography

Peter John Olivi, Tractatus de verbo, in Pasnau, ‘Petri Iohannis Olivi’s Tractatus de verbo’. ———, Quaestiones in secundum librum Sententiarum, ed. B. Jansen, Bibliotecha Franciscana Scholastica 4–6, Quarachi: Collegium. S. Bonaventurae, 1922–1926. ———, Tractatus, ed. L.M. de Rijk, Assen: Van Gorcum, 1972. ———, Quaestiones logicales, ed. S.F. Brown, in ‘Petrus Ioannis Olivi: “Quaestiones logicales”; Critical text’, Traditio 42 (1986), 335–388. Peter of Spain, Syncategoremata, ed. L.M. de Rijk, Leiden: Brill, 1992. Plato, Cratylus, ed. L. Méridier, Paris: Les Belles Lettres, 1931. Priscian, Institutiones grammaticae libri XVIII, 2 vols., ed. M. Hertz, gl II–III, Leipzig: Teubner, 1855–1859. Ps. Kilwardby, Commentum super Priscianum maiorum, ed. Fredborg et al. in ‘The commentary on “Priscianus Maior” ascribed to Robert Kilwardby’, 1975. Radulphus Brito, Quaestiones super Perihermeneias q. 3, partial edition in Pinborg, ‘Bezeichnung in der Logik des XIII. Jahrhunderts’, pp. 275–281, 1971. ———, Quaestiones super Isagogen, partial edition in Pinborg, ‘Radulphus Brito on Universals’, 1980. ———, Quaestiones super Priscianum minorem, ed. H.W. Enders and J. Pinborg, Stuttgart/Bad Cannstatt: Frommann/Holzboog, 1980. ———, Quaestiones super Analytica Posteriora, partial edition in Ebbesen, ‘Talking about what is no more’, 1987. ———, Quaestiones super Metaphysicam, partial edition in Ebbesen, ‘Words and Signification’, 107–114, 2000. ———, Quaestiones De anima, edition of book I and part of book II in S. de Boer, ‘Radulphus Brito’s Commentary on Aristotle’s De anima’, Vivarium 50.3–4 (2012), 245–353. Roger Bacon, Communia naturalium, ed. R. Steele, in Opera hactenus inedita Rogeri Baconi, Oxford: Clarendon Press, 1909–1913. ———, De signis, ed. Fredborg et al. in ‘An Unedited Part of Roger Bacon’s Opus Maius: De signis’, 1978. ———, Compendium studii theologiae, ed. T.S. Maloney, in Compendium of the Study of Theology, STGM 20, Leiden: Brill, 1988. Siger of Brabant, Ecrits de logique, de morale et de physique, ed. B. Bazán, Philosophes Médiévaux XIV, Louvain: Publications Universitaires, 1974. ———, Quaestiones in Metaphysicam, ed. A. Mauer, Philosophes Médiévaux XXV, Louvain-la-neuve: Editions de l’Institut Supérieur de Philosophie, 1983. Simon of Faversham, Quaestiones super Perihermeneias, in Opera logica I, ed. P. Mazarella, Padova: CEDAM, 1957. Thomas Aquinas, In duodecim libros Metaphysicorum Aristotelis expositio, ed. M.R. Cathala and R.M. Spiazzi, Turin: Marietti, 1950.

Bibliography

167

———, Expositio in librum Peri hermeneias, ed. R.-A. Gauthier, Opera omnia 1*/1, Rome/Paris: Commissio Leonina/Vrin, 1989. Walter Burley, Commentarius in librum Perihermeneias, in S.F. Brown, ‘Walter Burley’s Middle Commentary on Aristotle’s Perihermeneias’, Franciscan Studies 33 (1973), 45–134. William of Sherwood, Syncategoremata, ed. J.R. O’Donnell, in ‘The Syncategoremata of William of Sherwood’.

Secondary Literature Ackrill, J.L., Aristotle. Categories and De interpretatione, Oxford: Clarendon Press, 1963. Arens, H., Aristotle’s Theory of Language and Its Tradition, Amsterdam: John Benjamins, 1984. Ashworth, E.J., ‘“Can I speak more clearly than I understand?” A Problem of Religious Language in Henry of Ghent, Duns Scot and Ockham’, Historiographia linguistica 7 (1980), 29–38. ———, ‘Signification and Modes of Signifying in Thirteenth-Century Logic: A Preface to Aquinas in Analogy’, Medieval Philosophy and Theology 1 (1991), 39–67. ———, Les théories de l’analogie du XIIe au XVIe siècle, Paris: Vrin, 2008. ———, ‘Terministic Logic’, in Pasnau (ed.), The Cambridge History, vol. 1, pp. 146–158, 2010. Aubenque, P., ‘Sens et unité du traité aristotélicien De l’interprétation’, in S. Husson (ed.), Interpréter le De interpretatione, Paris: Vrin, 2009, pp. 37–50. Ayer, A.J., ‘The Criterion of Truth’, Analysis 3 (1935), 28–32. Bäck, A., Aristotle’s Theory of Predication, Leiden: Brill, 2000. Baratin, M., La naissance de la syntax a Rome, Paris: Editions de minuit, 1989. ———, ‘Sources philosophiques de Priscian: classement stoïcien des prédicats’, in Baratin et al. (eds.), Priscien, pp. 139–150, 2009. Baratin, M. and F. Desbordes, L’analyse linguistique dans l’antiquite classique. 1. Les théories, Paris: Editions Klincksieck, 1981. Baratin M., B. Colombat and L. Holtz (eds.), Priscien. Transmission et refondation de la grammaire de l’antiquité aux modernes, Studia Artistarum 21, Turnhout: Brepols, 2009. Barney, R., Names and Nature in Plato’s Cratylus, New York/London: Routledge, 2001. Belardi, W., ‘Riconsiderando la seconda frase del De intepretatione’, Studi et saggi linguistici 21 (1981), 79–83. Bermon, E., La signification et l’enseignement. Texte latin, traduction française et commentaire du De magistro de saint Augustin, Paris: Vrin, 2007. Biard, J., Logique et théorie du signe au 14e siècle, Paris: Vrin, 1989. Bobzien, S., ‘Aristotle’s De interpretatione 8 is about ambiguity’, in D. Scott (ed.), Maieusis. Essays in Ancient Philosophy in Honour of Myles Burnyeat, New York: Oxford University Press, 2007, pp. 301–321.

168

Bibliography

Boureau, A. and S. Piron (eds.), Pierre de Jean Olivi (1248–1298). Pensée scolastique, dissidence spirituelle et société, Paris: Vrin, 1999. Braakhuis, H.A.G., De 13de Eeuwse Tractaten over Syncategorematische Termen, II: Uitgave van Nicolaas van Parij’s Sincategoremata, 2 vols., Meppel: Krips Repro, 1979. Brancacci, A., ‘Les mots et les choses: La philosophie du langage chez Démocrite’, in Joly (ed.), Philosophie du langage et grammaire, pp. 9–28, 1986. Brumberg-Chaumont, J., (ed.), L’Organon dans la translatio studiorum à l’époque d’Albert le Grand, Turnhout: Brepols, 2013. Buckner, E. and J. Zupko, Duns Scotus on Time and Existence. The Questions on Aristotle’s ‘De interpretatione’, Washington DC: Catholic University of America Press, 2014. Cameron, M. and J. Marenbon (eds.), Aristotelian Logic East and West, 500–1500: On Interpretation and Prior Analytics in Two Traditions, Vivarium 48.1–2 (2010). ——— (eds.), Methods and Methodologies: Aristotelian Logic East and West, 500–1500, Investigating Medieval Philosophy 2, Leiden: Brill, 2011. Cessalli, L., Le réalisme propositionnel: sémantique et ontologie des propositions chez Jean Duns Scot, Gauthier Burley, Richard Brinkley et Jean Wyclif, Paris: Vrin, 2007. ———, ‘Meaning and Truth’, in Conti (ed.), A Companion, pp. 87–134, 2013. Cesalli, L., A. de Libera and F. Goubier, ‘Does Loving Every Mean Loving Every Every, even Non-existent Ones? Distribution and Universals in the Opus puerorum’, in Fink et al. (eds.), Logic and Language, pp. 305–336, 2013. Charles, D., Aristotle on Meaning and Essence, Oxford: Clarendon Press, 2000. Chiesa, C., ‘Symbole et signe dans le De interpretatione’, in Joly (ed.) Philosophie du langage et grammaire, pp. 203–218, 1986. Christensen, J., An Essay on the Unity of Stoic Philosophy, Copenhagen: Museum Tusculanum Press, 2012. Conti, A. ‘Significato e verità in Walter Burley’, Documenti e studi sulla tradizione filosofica medievale 11 (2000), 317–350. ———, (ed.), A Companion to Walter Burley, Brill’s Companions to the Christian Tradition 41, Leiden: Brill, 2013. Crivelli, P., Aristotle on Truth, Cambridge: cup, 2004. ———, ‘Aristotle on Signification and Truth’, in G. Anagnostopoulos, A Companion to Aristotle, Oxford: Blackwell, 2009, pp. 81–100. Da Silva, J.F., Robert Kilwardby on the Human Soul. Plurality of Forms and Censorship in the Thirteenth Century, Investigating Medieval Philosophy 3, Leiden: Brill, 2012. De Libera, A., ‘Roger Bacon et le problème de l’appellatio univoca’, in H.A.G. Braakhuis et al. (eds.), English Logic and Semantics: From the end of the twelfth century to the time of Ockham and Burleigh. Acts of the 4th European Symposium on Medieval Logic and Semantics, Leiden-Nijmegen: Ingenium Publishers, 1981, pp. 193–221.

Bibliography

169

———, César et le phénix: Distinctiones et sophismata parisiens du XIIIe siècle, Pisa: Centro di cultura medievale della Scuola Normale Superiore, 1991. ———, ‘Roger Bacon et la référence vide. Sur quelques antécédents médiévaux du paradoxe de Meinong’, in J. Jolivet et al. (eds.), Lectionum varietates. Hommage à Paul Vignaux (1904–1987), Paris: Vrin, 1991, pp. 85–120. ———, ‘Référence et quantification: Sur la théorie de la distributio au XIIIe siècle’, in A. de Libera, A. Elamrani-Jamal and A. Galonnier (eds.), Langages et philosophie: Hommage à Jean Jolivet, Paris: Vrin, 1997, pp. 177–200. ———, L’art des généralités. Théories de l’abstraction, Paris: Aubier, 1999. ———, ‘Omnis homo de necessitate est animal. Référence et modalité selon l’Anonymus Erfordensis Q. 328 (Pseudo-Robert Kilwardby)’, ahdlma 69 (2002), 201–237. ———, La référence vide, Paris: Presses Universitaires de France, 2002. ———, Métaphysique et noétique. Albert le Grand, Paris: Vrin, 2005. ———, ‘Faire de nécessité loi: Théories de la modalité dans le sophisma “Omnis homo de necessitate est animal” du codex parisisnus 16135, ff. 11rb–12rb’, ahdlma 76 (2009), 179–233. De Libera, A. and L. Gazziero, ‘Le Sophisma “Omnis homo de necessitate est animal” du Parisinus Latinus 16135, ff. 99rb–103vb’, ahdlma 75 (2008.1), 323–368. De Nonno, M., ‘Ars Prisciani Caesariensis: problemi di tipologia e di composizione’, in M. Baratin et al. (eds.), Priscien, pp. 249–278, 2009. De Rijk, L.M., Aristotle Semantics and Ontology. Volume I: General Introduction. The Works on Logic, Philosophia Antiqua XCI/I, Leiden: Brill, 2002. Donati, S., ‘Apparentia and Modi essendi in Radulphus Brito’s Doctrine of the Concepts: The Concept of Being’, in Fink et al. (eds.), Logic and Language, pp. 337–355, 2009. Dutilh Novaes, C., ‘Medieval Theories of Supposition’, in H. Lagerlund (ed.), Encyclopedia of Medieval Philosophy, Berlin: Springer, 2011, pp. 1229–1236. Ebbesen, S., ‘The Dead Man is Alive’, Synthese 40 (1979), 43–70. ———, ‘The Present King of France wears Hypothetical Shoes with Categorical Laces: Twelfth-century writers on well-formedness’, Medioevo 7 (1981), 91–113 (repr. in Ebbesen, Collected Essays, vol. 2, pp. 15–30). ———, ‘The Chimaera’s Diary – edited by Sten Ebbesen’, in S. Knuuttila and J. Hintikka (eds.), The Logic of Being, Dordrecht: Reidel, 1986, pp. 115–143 (repr. in Ebbesen, Collected Essays, vol. 1, pp. 35–57). ———, ‘Talking about what is no more’, cimagl 55 (1987), 135–168. ———, ‘Boethius as an Aristotelian Commentator’ in R. Sorabji, (ed.), Aristotle Transformed. The Ancient Commentators and their Influence, London: Duckworth, 1990, pp. 373–391. ———, ‘Porphyry’s Legacy to Logic: a Reconstruction’, in Sorabji (ed.), Aristotle Transformed, pp. 141–172, 1990.

170

Bibliography

———, ‘Animal est omnis homo: Questions and sophismata by Peter of Auvergne, Radulphus Brito, William of Bonkes and Others’, cimagl 63 (1993), 145–208, at pp. 150–155. ———, (ed.) Sprachteorien in Spätantiken und Mittelalter, vol. 3 of P. Schmitter (ed.), Geschichte der Sprachtheorie, Tübingen: Gunter Narr Verlag, 1995. ———, ‘Boethius of Dacia: science is a serious game’, Theoria 66 (2000), 145–158 (repr. in Ebbesen, Collected Essays, vol. 2, pp. 153–162). ———, ‘Words and Signification in 13th-century Questions on Aristotle’s Metaphysics’, cimagl 71 (2000), 71–114. ———, ‘Radulphus Brito: The last of the great masters, or Philosophy and Freedom’, in J.A. Aertsen and A. Speer (eds.), Geistesleben im 13. Jahrhundert, Miscellania Mediaevalia 27, Berlin: De Gruyter, 2000, pp. 231–251 (repr. in Ebbesen, Collected Essays, vol. 2, pp. 179–196). ———, ‘Radulphus Brito on the Metaphysics’, in J. Aertsen, K. Emerey and A. Speer (eds.), Nach der Verurteilung von 1277: Philosophie und Theologie an der Universität von Paris im letzten Viertel des 13. Jahrhunderts; Studien und Texte, Miscellanea Mediaevalia 28, Berlin: De Gruyter, 2001, pp. 456–492 (repr. in Ebbesen, Collected Essays, vol. 2, pp. 197–208). ———, ‘Questions and Sophismata: Tracking Peter of Auvergne’, in H.A.G Braakhuis and al. (eds.), Aristotle’s Peri hermeneias in the Latin Middle Ages: Essays on the Commentary Tradition, Groningen: Ingenium, 2003, pp. 31–49. ———, ‘The Man who Loved Every: Boethius of Dacia on logic and metaphysics’, The Modern Schoolman 82 (2005), 235–250 (repr. in Ebbesen, Collected Essays, vol. 2, pp. 163–178). ———, ‘Theories of Language in the Hellenistic Age and in the Twelfth and Thirteenth Centuries’, in D. Frede and B. Inwood (eds.), Language and Learning: Philosophy of Language in the Hellenistic Age, Cambridge: cup, 2005, pp. 299–319 (repr. in Ebbesen, Collected Essays, vol. 1, pp. 79–98). ———, ‘By Necessity’, in V. Hirvonen et al. (eds.), Mind and Modality, pp. 141–151, 2006. ———, Collected Essays of Sten Ebbesen, 2. vols., Aldershot/Farnham: Ashgate, 2008–2009. ———, ‘Priscian and the Philosophers’, in Baratin et al. (eds.), Priscien, pp. 85–108, 2009. Ebbesen, S. and F. Goubier, A Catalogue of 13th-Century Sophismata, 2. vols., Paris: Vrin, 2010. Ebbesen, S. and J. Pinborg, ‘Bartholomew of Bruges and his Sophisma on the Nature of Logic’, cimagl 39 (1981), 1–80. Ebbesen, S., D. Bloch, J. Fink, H. Hansen and A.M. Mora-Márquez, History of Philosophy in Reverse. Reading Aristotle through the Lenses of Scholars from the Twelfth to the Sixteenth Centuries, Scientia Danica. Series H. Humanistica 8. vol. 7, Copenhagen: The Royal Danish Academy of Sciences and Letters, 2014.

Bibliography

171

Fink, J., H. Hansen and A.M. Mora-Márquez (eds.), Logic and Language in the Middle Ages. A Volume in Honour of Sten Ebbesen, Investigating Medieval Philosophy 4, Leiden: Brill, 2013. Fredborg, K.M., ‘Notes on the Glosule and its reception by William of Conches and Peter Helias’, in Rosier-Catach (ed.), Arts du langage et théologie, pp. 453–484, 2011. Fredborg, K.M., L. Nielsen and J. Pinborg, ‘An Unedited Part of Roger Bacon’s Opus Maius: De signis’, Traditio 34 (1978), 75–136. Fredborg, K.M., N.J. Green-Pedersen, L. Nielsen and J. Pinborg, ‘The commentary on “Priscianus Maior” ascribed to Robert Kilwardby’, cimagl 15 (1975), 1–146. Frede, M., ‘The Origins of Traditional Grammar’, in Essays in Ancient Philosophy, Minessota: Minessota Archive Editions, 1987, pp. 338–362. Goubier, F., ‘Influences prédicatives et conséquences référentielles: un aspect de l’approche terministe de la première moitié du XIIIe siècle’, cimagl 71 (2000), 37–70. ———, ‘Les syncatégorèmes au XIIIe siècle’, Histoire Epistémologie Langage 25/II (2003), 85–113. Gourinat, J.B., La dialectique des stoïciens, Paris: Vrin, 2000. Hansen, H., and A.M. Mora-Márquez, ‘Nicholas of Paris on Aristotle’s Perihermeneias 1-3’, cimagl 80 (2011), 1–88. Hirvonen, V., T.J. Holopainen and M. Tuominen (eds.), Mind and Modality. Studies in the History of Philosophy in Honour of Simo Knuuttila, Leiden: Brill, 2006. Hunt, R.W., ‘Studies on Priscian in the Eleventh and Twelfth Centuries I: Petrus Helias and his Predecessors’, Medieval and Renaissance Studies 1 (1943), 194–231. Isaac, J., Le Perihermeneias en occident de Boèce à Saint Thomas, Paris: Vrin, 1953. Joly, H., Philosophie du langage et grammarie dans l’antiquité, Cahiers de philosophie ancienne 5, Bruxelles: Editions Ousia, 1986. Kelly, L.G., The Mirror of Grammar: Theology, philosophy and the modistae (Amsterdam: John Benjamins, 2002). Kneepkens, C.H., ‘The Priscianic Tradition’, in Ebbesen (eds.), Sprachtheorien, pp. 239– 264, 1995. ———, ‘Significatio generalis and significatio specialis: Notes on Nicholas of Paris’ Contribution to Early Thirteenth-Century Linguistic Thought’, in Ebbesen et al. (eds.), Medieval Analyses, pp. 17–43, 1999. ———, ‘Grammar and Semantics in the Twelfth Century: Petrus Helias and Gilbert de la Porrée on the Substantive Verb’, in M. Kardaun and J. Spruyt (eds.), The Winged Chariot: Collected Essays on Plato and Platonism in Honour of L.M. de Rijk, Leiden: Brill, 2000, pp. 237–275. ———, ‘A Note on Articulatio and University Grammar’, in Fink et al. (eds.), Logic and Language, pp. 221–238, 2013.

172

Bibliography

Kretzmann, N., ‘Medieval Logicians on the Meaning of the Propositio’, Journal of Philosophy 67 (1970), 767–787. ———, ‘Plato on the correctness of names’, American Philosophical Quarterly 8 (1971), 126–138. ———, ‘Aristotle on Spoken Sounds Significant by Convention’, in J. Corcoran (ed.), Ancient Logic and its Modern Interpretation, Dordrecht: Reidel, 1974, pp. 3–21. Lagerlund, H. and P. Thom (ed.), A Companion to the Philosophy of Robert Kilwardby, Brill’s Companions to the Christian Tradition 37, Leiden: Brill, 2013. Law, V., The History of Linguistics in Europe. From Plato to 1600, Cambridge: cup, 2003. Lewry, P.O., Robert Kilwardby’s Writings on the Logica Vetus with Regard to their Teaching and Method, Oxford: Ph.D. dissertation, 1978. ———, ‘Robert Kilwardby on Meaning: A Parisian Course on the “Logica Vetus”’, in Beckmann et al. (eds.), Sprache und Erkenntis im Mittelalter, Berlin: De Gruyter, 1981, pp. 376–384. Long, A.A. and D. Sedley, The Hellenistic Philosophers. Vol. 1, Cambridge: cup, 1987. Longeway, J.L., Simon of Faversham’s Questions on the Posterior Analytics: a ThirteenthCentury View of Science, Ithaca: Cornell University, 1977; PhD dissertation. Luhtala, A., Grammar and Philosophy in Late Antiquity. A study of Priscian’s sources, Studies in the History of Language Sciences, Amsterdam: John Benjamins, 2005. Magee, J., Boethius on Signification and Mind, Philosophia Antiqua 52, Leiden: Brill, 1989. Maloney, T.S., ‘The Semiotics of Roger Bacon’, Medieval Studies 45 (1983), 120–154. ———, ‘Roger Bacon on Equivocation’, Vivarium 22 (1984), 85–112. Marenbon, J., Boethius, Oxford: oup, 2003. ———, ‘Gilbert of Poitiers’s Contextual Theory of Meaning’, in Fink et al. (eds.), Logic and Language, pp. 49–64, 2013. Markus, R.A., ‘St. Augustine on Signs’, Phronesis 2 (1957), 60–83. Marmo, C., ‘Ontology and Semantics in the Logic of Duns Scotus’, in U. Eco and C. Marmo (eds.), On the Medieval Theory of Signs, Amsterdam: John Benjamins, 1989, pp. 143–193. ———, Semiotica e linguaggio nella Scolastica: Parigi, Bologna, Erfurt 1270–1330; La semiotica dei modisti, Rome: Istituto Storico Italiano per il Medioevo, 1994. ———, ‘A Pragmatic Approach to Language in Modism’, in Ebbesen (ed.), Sprachtheorien, pp. 169–183, 1995. ———, ‘The Semantics of the Modistae’, in Ebbesen et al. (eds.) Medieval Analyses, pp. 83–104, 1999. ———, La semiotica del XIII secolo tra teologia e arti liberali, Milan: Bompiani, 2010. Martin, C.J., ‘Imposition and Essence: What’s new in Abelard’s theory of meaning’, in T. Shimizu and C. Burnett (eds.), The Word in Medieval Logic, Theology and Psychology, Turnhout: Brepols, 2009, pp. 173–214.

Bibliography

173

———, ‘The Development of Logic’, in Pasnau (ed.), Cambridge History. vol. 1, pp. 129–132, 2010. Montanari, E., La Sezione linguistica del Peri hermeneias di Aristotele. 2 vols., Studi et testi 5 and 8, Florence: Università degli studi di Firenze, 1988. Mora-Márquez, A.M., ‘A New Perspective on the Origin of the Debate on Signification at the End of the 13th Century’, in C.T. Thörnqvist and B. Bydén (eds.), Papers on Aristotelian Logic and Metaphysics Presented to the Danish-Swedish Network for the Aristotelian Tradition in the Middle Ages (2009–2011) (Toronto: pims, forthcoming). ———, ‘Pragmatics in Peter John Olivi’s Account of Signification of Common Names’, Vivarium 49 (2011), 150–164. ———, ‘Radulphus Brito on Common Names, Concepts and Things’, in Fink et al. (eds.), Logic and Language, pp. 357–372, 2013. ———, ‘(Review of) Simon Noriega-Olmos, A Commentary on De interpretatione 16a3–18. (Beträge zur Altertumskunde 303) pp. x + 185. Berlin and Boston: De Gruyter, 2013’, The Classical Review/FirstView Article (April 2014), 1–2. ———, ‘Anonymus Oxford, Commentary on De interpretatione 1 (MS Oxford, BodlL Can. misc. 403, ff. 31ra–34vb)’, cimagl 83 (2014), 135–206. ———, ‘Martinus Dacus and Boethius Dacus on the Signification of Terms and the Truth-value of Assertions’, Vivarium 52 (2014), 23–48. ———, ‘Boethius of Dacia (1270s) and Radulphus Brito (1290s) on the Universal Sign “Every”’, Logica Universalis (2015), DOI 10.1007/s11787-014-0112-6. Noriega-Olmos, S., Aristotle’s Psychology of Signification. A Commentary on De interpretatione 16a3–18, Beträge zur Altertumskunde 303, Berlin: De Gruyter, 2013. O’Donnell, J.R., ‘The Syncategoremata of William of Sherwood’, Mediaeval Studies 3 (1941), 46–93. Panaccio, C., ‘Grammar and Mental Language in the Pseudo-Kilwardby’, in Ebbesen et al. (eds.), Medieval Analyses, pp. 397–414, 1999. Pasnau, R., ‘Petri Iohannis Olivi Tractatus de Verbo’, Franciscan Studies 53 (1993), 121–153. ———, Theories of Cognition in the Later Middle Ages, Cambridge: cup, 1997. ———, ‘Olivi on Human Freedom’, in Boureau et al. (eds.), Pierre de Jean Olivi, pp. 15–26, 1999. ——— (ed.), The Cambridge History of Medieval Philosophy. 2 vols., Cambridge: cup, 2010. Pépin, J., ‘Sumbola, Sèmeia, Homoiômata: A propos de De intepretatione 1.16a3–8 et Politique VIII 5, 1340a6-39’, in J. Wiesner (ed.), Aristoteles. Werk and Wirkung. Aristoteles und seine Schule, t. 1, Berlin: De Gruyter, 1985, pp. 22–44. Perler, D., ‘Duns Scotus on Signification’, Medieval Philosophy and Theology 3 (1993), 97–120.

174

Bibliography

———, ‘Duns Scotus’s Philosophy of Language’, in T. Williams (ed.), The Cambridge Companion to Duns Scotus, Cambridge: cup, 2003, pp. 161–192. ———, Théories de l’intentionnalité au moyen âge, Paris: Vrin, 2003. Pinborg, J., Die Entwicklung der Sprachtheorie im Mittelalter, Beiträge zur Geschichte der Philosophie und Theologie des Mittelalters XLII/2, Münster: Aschendorff, 1967. ———, ‘Bezeichnung in der Logik des XIII. Jahrhunderts’, in A. Zimmermann (ed.), Der Begriff der repraesentatio im Mittelalter, Miscellanea Mediaevalia 8, Berlin: De Gruyter, 1971, pp. 238–281. ———, ‘Zum Begriff der Intentio secunda. Radulphus Brito, Hervaeus Natalis und Petrus Aureoli in Diskussion’, cimagl 13 (1974), 49–59. ———, ‘Radulphus Brito on Universals’, cimagl 35 (1980), 56–142. Pini, G., ‘Species, Concept and Thing: Theories of Signification in the Second Half of the Thirteenth Century’, Medieval Philosophy and Theology 8 (1997), 21–52. ———, ‘Signification of Names in Duns Scotus and Some of his Contemporaries’, Vivarium 39 (2001), 20–51. Pironnet, F. and J. Spruyt, ‘Sophismata’, Stanford Encyclopedia of Philosophy, http:// plato.stanford.edu/entries/sophismata/, 2011. Putallaz, F.X., Insolenté liberté. Controverses et condamnations au XIIIe siècle, Fribourg/ Paris: Presses Universitaires/Le Cerf, 1995. Putnam, H., ‘The Meaning of “Meaning”’, in Mind, Language and Reality. Philosophical Papers. Vol. 2, Cambridge: cup, 1975, pp. 215–271. ———, ‘Meaning, Other People and the World’, in Representation and Reality, Cambridge MA: MIT Press, 1988, pp. 19–41. Read, S. ‘Medieval Theories: Properties of Terms’, Standford Encyclopedia of Philosophy, http://plato.stanford.edu/entries/medieval-terms/, 2015. Robert, A., Penser la substance: étude d’une question médiévale, Nantes: University of Nantes, 2005; unpublished PhD dissertation. Rosier-Catach, I., La grammaire spéculative des modistes, Lille: Presses Universitaires de France, 1983. ———, La parole comme acte. Sur la grammaire et la sémantique au XIIIe siècle, Paris: Vrin, 1994. ———, ‘Henri de Gand, le De dialectica d’Augustin et l’institution des noms divins’, Documenti e studi sulla tradizione filosofica medievale 6 (1995), 145–253. ———, ‘Res significata et modi significandi: les implication d’une distinction médiévale’, in Ebbesen (ed.) Sprachteorien, pp. 135–168, 1995. ———, ‘Modisme, pré-modisme, proto-modisme: vers une définition modulaire’, in Ebbesen et al. (eds.), Medieval Analyses, pp. 45–81, 1999. ———, La parole efficace. Signe, rituel, sacré, Paris: Seuil, 2004. ———, ‘Grammar’, in Pasnau (ed.), The Cambridge History, vol. 2, pp. 196–218, 2010.

Bibliography

175

———, (ed.), Arts du langage et théologie aux confins des XIe–XIIe siècles. Textes, maîtres, débats, Studia artistarum 21, Turnhout: Brepols, 2011. Sedley, D., Plato’s Cratylus, Cambridge: cup, 2003. ———, ‘Plato’s Cratylus’, Stanford Encyclopedia of Philosophy (2006), http://plato .stanford.edu/entries/plato-cratylus/. Shields, C., Order in Multiplicity: Homonymy in the Philosophy of Aristotle, Oxford: Oxford University Press, 1999; paperback edition, 2001. Speaks, J., ‘Theories of Meaning’, Stanford Encyclopedia of Philosophy (2014), http:// plato.stanford.edu/entries/meaning/. Spruit, L., Species intelligibilis. From Perception to Knowledge. t. 1: Classical Roots and Medieval Discussions, Leiden: Brill, 1994. Suto, T., Boethius on Mind, Grammar and Logic. A Study of Boethius’ Commentary on Peri hermeneias, Philosophia Antiqua 127, Leiden: Brill, 2012. Tabarroni, A., ‘Omnis phoenix est: quantification and existence in a new sophismata collection (MS Clm 14522)’, in S. Read (ed.), Sophisms in Medieval Grammar and Logic, Dortrecht: Kluwer, 1993, pp. 185–201. Thom, P., ‘On the Pervasiveness of Being’, in V. Caston and D.W. Graham (eds.), Presocratic Philosophy: Essays in Honour of Alexander Mourelatos, Aldershot: Ashgate, 2002, pp. 293–301. Toivanen, J., Perception and the Internal Senses. Peter of John Olivi on the Cognitive Functions of the Sensitive Soul, Investigating Medieval Philosophy 5, Leiden: Brill, 2013. Valente, L., Logique et théologie. Les écoles parisiennes entre 1150 et 1220, Paris: Vrin, 2008. Wheeler, M., ‘Semantics in Aristotle’s Organon’, Journal of the History of Philosophy 37.2 (1999), 191–226. Whitaker, C.W.A., Aristotle’s De interpretatione. Contradiction and Dialectic, Oxford: oup, 1996.

Index of Subjects abstraction 32n64, 51n34, 69 accident 3, 39–40, 57–58, 125 action and undergoing of action 14n12, 15, 63, 101, 113–115, 113n20–21, 114n24, 117, 127, 134–135, 138 adjective 65, 116n32 adverb 13n10, 114n24, 115, 124 affection, of the soul 13–14, 13n10, 34, 52, 110, 111n8 affirmation 19–21, 23, 121, 149 arrangement (ordo/ordinatio) 23, 25, 111, 128, 133, 135, 148 article 112–115, 114n24 articulation 15n16, 52, 124, 128, 155 assertion (enuntiatio) 2, 12, 14, 18–23, 67, 71–76, 78, 83, 84n109, 90–92, 94–96, 98–100, 111, 113, 118–121, 121n43, 130–139, 139n34, 147n55, 148–150, 150n60, 154–155, 155n69 accidental 56, 82, 91 essential 66, 82–83, 95, 100, 103–104 indefinite 84 modal 92 quantified 84–85, 85n111, 154 with empty subject 76, 77n84, 80–81, 84, 95 attribute (rhêma) 14, 14n12, 15n16, 17–21, 17n21, 19n25, 72, 83, 84n109, 99, 104, 111, 117–121, 121n43, 149 being (esse) 18, 41, 44, 51, 58n47, 59–61, 64, 85, 96–99, 120, 126, 131–134 actual and dispositional 80, 95, 95n133, 96–98, 103 categories of 71–73, 126 imagined 48n29 mental 43, 48, 76, 95, 102 modes of 59–61, 68, 97, 139–145, 150n60, 152, 155, 158 of essence and of existence 96–99, 99n144 bivalence, principle of 19, 83–84 body 28, 41, 87n114, 88, 113, 113n20, 116n32, 159 case (casus) 65, 117, 129, 131–132 cause efficient 43, 47, 83, 113n21 final 49, 124–125

material 155 of logic and grammar 92–94, 127–129, 131, 134, 138, 140, 146, 149–154, 150n60 chimera 24n40, 159–160 cognition 50, 54, 68n66 communication 2–6, 12–13, 13n8, 15, 17–18, 17n21, 21, 25, 30, 35, 49–52, 63, 65–67, 91–92, 100, 105, 110, 143n46, 157, 159 composition 18, 55–56, 72–75, 77n84, 79–86, 93, 95, 97, 104, 119–120, 124, 124n4, 131, 146, 149, 150n60 concept 4, 17n21, 24, 28–75, 83, 85n110, 94, 101, 118, 124, 127, 131, 140–141, 143, 143n48, 145, 159–160 conjunction 79n95, 112–115, 114n24 contingent 94, 121n43 convention/conventional (ad placitum)  14–16, 15n16, 16n18, 18, 25, 27–28, 44, 49–50, 67, 118, 127, 159 copula 79, 79n95, 87n114, 96 corporeal/incorporeal 27, 27n49, 45, 113, 113n20, 116, 116n31 consignification 112, 114, 124, 127–128, 144–145 co-understanding 144–145 dialectics/dialectical 12, 19, 21, 112, 114, 147n55, 148, 150n60 disposition/dispositional 14n12, 66, 73, 80, 87, 95–98, 95n133, 103 division 17–18, 22, 55, 61, 72–74, 83, 95, 109n3, 110–112, 110n6, 114, 116n32, 118, 126, 136, 139–140, 145–147, 149, 150n60 enthymeme 16 equivocation/equivocal 20, 40–41, 66, 76n83, 80, 87n114, 92, 95, 99, 100, 103, 118, 149 essence 32, 38, 49, 51–60, 51n34, 58n47, 70, 76, 92, 95–99, 99n144, 101–103, 105n159, 151, 157 existence 19, 25, 27n49, 34, 52, 56–57, 68–69, 75, 80–81, 83, 84n109, 94–99, 99n144, 101–105, 120, 126, 129–131, 145, 151, 151n64 expression 13–14, 13n10, 22, 24, 26, 33, 50n33, 110–111, 118, 120, 130n18, 138

177

Index of Subjects faculty of Arts 3–6, 53n38, 58, 109, 161 of the soul 17n21, 43, 51n34, 53n38, 68 feature accidental 69, 116, 129, 132, 140, 143, 145 essential 31–34, 57, 63, 98, 116, 132 linguistic 22, 28, 53n38, 66, 68, 70, 73, 79, 85n110, 114n25, 115, 115n27, 117–119, 121, 125, 129, 130, 155, 158 proper 110, 110n6, 112–113, 126 figure ( figura) 88 form 17n21, 68 essential or accidental 126, 132 intelligible 68 multiplicity of 105n159 genus 3, 32–33, 40, 55, 59–60, 76, 82, 105n159, 117, 128, 154 God 6, 6n10, 24n40, 53n38 good 19, 73, 120, 138 grammarian 28, 117n32, 126–129, 133–137, 141, 146, 150–154 grammaticality 2, 4, 106, 109n3, 118, 121, 126–134, 139–140, 146, 153–156 Idea 27n49, 37, 69 identity 15n15, 34–35, 52, 96, 129–130, 131, 154 image (imago/phantasma) 11n2, 27–28, 28n51, 45, 48n29, 88–89, 90n121, 101–102 imagination 37, 47 imposition and re-imposition 41–70, 41n14–15, 45n21, 48n29, 74–75, 87, 89–92, 89n117, 117, 124, 124n4, 151, 157 impression 28n51, 32 inference 70, 83–84 inherence 87n114, 103 institution of words 2, 5, 45–46, 45n21, 49, 61, 124–125, 151, 155, 157 intellect 24, 30, 37, 46–50, 51n34, 52–53, 56, 58–60, 58n47, 62, 67–71, 73, 83, 94, 97, 99, 101, 145 intention 49, 136 abstract 59–60 common 127–128, 153 concrete 59–60 first 59–60 logic of 70, 147, 158, 160–161 second 55, 57, 59–61, 66, 74, 152–155, 157, 159 universal 61

intentionality 69n67, 127 interlocutors 12, 17n21, 21 interpretation 3, 17, 22–23 knowledge 61, 63n55, 105, 152 certain 147–148 human 3, 24n40, 25, 55, 68 intellectual 18, 45, 50, 58–59, 69, 98 methods of 146–147, 155–157 probable 147–148 letter (littera) 13, 22, 24, 43, 110n3, 111n8, 114, 128 likeness (similitudo) 13–14, 17n21, 31–32, 31n63, 34, 39–40, 43, 47, 53–56, 69–70, 75, 127, 142 listener 17–18, 20, 46, 55–57, 74, 101, 110–111, 143n46, 159–160 logician 7, 111, 123, 127–129, 133–134, 137–138, 146–150, 152–153 mark (nota) 16n18, 37, 49, 53, 71 meaning 1–2, 12n4, 25, 158 mediation 29 mind 16–18, 24n40, 25–26, 47, 55–57, 62, 63n55, 65, 74, 88–89, 110, 143, 160 modality 59, 65, 77n84, 86, 92, 125–126, 148 mode of being (see ‘being’) of consignifying 127–128, 144 of imposition 89 of signifying 60, 66, 68, 72, 94, 124–130, 137–146, 151–155, 157–158 of understanding 59–60, 67–68, 70–72, 94, 97, 126–127, 129, 139–145 of verbs 117 modistic grammar 139–140 name common 41–42, 50–52, 54, 58n45, 70, 112–113, 116n32, 139 proper 50, 105n159, 112–113, 115, 116n32, 119, 139 compound 119, 119n39, 121n43 nature 14n14, 15, 24, 27, 27n49, 29, 31, 33–34, 43, 48n29, 65–66, 69, 101, 125–126, 153 necessity 77n84, 79–82, 86 negation 19–20, 23, 85, 100, 121, 149 nomination 161

178 notion (ratio) 40, 55, 61, 69–71, 87, 90, 96–98, 103–105, 110, 110n6, 134, 137, 140, 145, 147, 149–150 opinion 18–19, 21, 50n32 participle 65, 115 passion 25, 28, 31, 37, 48, 53, 55, 131 phoneme 114n24 physics 149 predicate 3, 17n21, 41n14, 56, 71–72, 76, 79n95, 80–81, 81n99, 83, 85, 85n111, 95–96, 103, 113, 113n21, 117, 129–130, 132, 136–137, 146, 160 predication 56, 75, 77n84, 82, 82n104, 90–91, 95, 100, 103, 105, 126, 136–138 premise 147n55, 148, 152, 154 preposition 114–115, 114n24 pronoun 112–115, 114n24, 115n26–27 proposition 20, 55, 82, 96, 103, 153–154, 160 propositional context 90 item 26 verification 4, 70, 73, 76, 94–95, 105–106, 148, 154–155, 158, 160–161 psychology 25, 149 quality 3, 14n12, 30, 38, 41, 116, 116n32, 125–126, 130, 132, 134–136, 138, 144 quantification/quantifier 76n83, 77n84, 79, 81–82, 85n110–111, 86 reality 54, 60, 66, 70, 140, 151, 157 recollection 63n55 reference 11, 12n4, 57, 72, 75, 77n84, 78–80 relation 3, 16–17, 23, 29, 32, 41, 47, 49, 56, 63, 65–66, 66n62, 70, 70n69, 85–87, 85n111, 87n114, 96, 101, 103, 129, 131–134, 142–146, 143n46, 154–155, 157 rhetoric 16 sayable (dicible/lekton) 41, 60, 60n48, 71, 113, 113n17, 113n20–21 semantic(s) 1, 12, 12n4, 23, 29, 85n110, 87, 94, 110n4, 113, 119, 161 semiotic(s) 1–2, 4, 13, 21, 27, 34–35, 38, 41, 50, 52, 54, 57, 61, 64, 66, 70, 75–76, 78, 86, 88n116, 92, 95, 99, 106, 122–123, 154, 156–157, 160, 162

Index of Subjects sense 12n4, 15n16, 24, 26, 42, 47, 50, 62, 76n83, 80–81, 92, 101, 103–104, 110–111, 115, 118, 121, 129n13 sentence (oratio) 2, 14, 15n16, 18–19, 22–23, 73, 75, 91–92, 95, 109, 109n3, 111–115, 111n9, 115n26, 117–121, 127–130, 132–134, 132n23, 136, 138–140, 146, 150–151, 153–155 sign 13–14, 13n10, 16, 21, 41, 43–44, 62–63, 63n55, 75–76, 85, 89, 89n116, 92, 119 linguistic 2, 15n16, 61, 63, 105, 123, 157 given and natural 61–62 speaker 16–18, 20, 26, 46–47, 55, 64–67, 89–91, 96, 101, 110, 143n46 s.’s intention 67, 96 species praedicabilis 3, 32–34, 55, 55n41, 59, 65–66, 76, 82, 105n159, 128, 154 intelligibilis 31n63, 37n5, 38n9, 44n20, 45n23, 51n34, 56, 60, 62, 68, 74n78–79, 75 speech external 23, 47, 113, 133 internal 47 parts of 4–5, 65, 109, 109n3, 111–112, 111n9, 113–118, 115n27, 123–126, 124n5, 128–129, 132–134, 132n23, 137, 139–140, 143–146, 152–154 subject of accidents 140–142, 145, 154 of attribution/predication 34, 41, 61, 71–73, 76, 77n84, 79n95, 80–83, 81n99, 82n109, 85, 85n111, 91, 95–96, 100, 113, 116–117, 120, 129–130, 134, 136–139, 146, 160 s.-matter 22, 70, 147, 157, 159n5 substance 3, 33–34, 38, 40–41, 58, 87, 95, 113n21, 116, 116n32, 124–127, 134–136, 138, 159 supposition 85, 85n110, 160–161 syllable 109n3, 114, 128 syllogism 16, 51, 147–148, 148n55, 154 symbol (symbolum/symbolon) 13–16, 13n10, 15n16, 16n18, 25 syntax/syntactic 1, 76n83, 112–113 tense 117, 121n43, 160 tertium adiacens 96, 103–104 theology 53n38, 58, 161

Index of Subjects thought 12–14, 13n10, 16–19, 21, 23–27, 24n40, 29–30, 41n15, 61–62, 101 compound 17n21, 18–20, 28, 133–134, 136 simple 17n21, 18–21, 19n25, 59, 67, 72, 75, 125 object of 31, 90 time 18, 26, 66, 84, 88, 90, 100, 118–120 token/type 15, 15n15, 25, 66, 70, 115n26 translation 3, 7, 13n10, 14, 14n12, 21–22, 22n34, 25–26, 25n42, 109 truth and falsity 4, 18, 73–74, 119, 121, 127–139, 146–150, 153, 155 truth-determination 73, 155

179 truth-value 18–19, 21, 72, 76, 77n84, 121, 121n43, 148, 155, 155n69 universality 52, 59, 61, 65, 68–70 universals 50–51, 51n34, 52, 54–55, 61, 64, 65, 68, 84, 128, 154 univocity/univocal 11–12, 15, 16n18, 41, 66, 87n114, 100 verb (verbum) 4, 14n12, 22–23, 23n35, 26, 28–31, 34, 44, 71–72, 96–98, 103, 109, 113–118, 114n24, 121–130, 133–138, 139n34, 143 verification/verificational 4, 12, 21, 35, 38, 41, 70, 73, 76, 92, 94–95, 105–106, 148, 154–155, 158, 160–161

Index of Modern Authors Arens, H. 28n48 Ashworth, E.J. 6n10, 41n15, 53n38, 85n110 Aubenque, P. 11n1–2 Ayer, A.J. 158n1 Bäck, A. 83n104 Baratin M. 109n1–2, 113n18, 117n32, 117n34 Barney, R. 14n14 Belardi, W. 13n9–10 Bermon, E. 62n50, 63n55 Biard, J. 39n10, 63n56, 158n3–4, 161n13, 162n14 Bobzien, S. 12n6, 20n27 Boureau, A. 70n69 Brancacci, A. 14n14 Brumberg-Chaumont, J. 49n30, 147n53 Cameron, M. 2n2, 4n7 Cesalli, L. 39n10, 77n84, 81n100 Charles, D. 17n21 Chiesa, C. 14n13 Christensen, J. 112n15 Conti, A. 39n10, 42n17 Crivelli, P. 18n24, 84n109, 104n156 Da Silva, J.F. 43n18 De Libera, A. 32, 51n34, 77n84, 87n114 De Nonno, M. 109n3, 121n45 De Rijk, L.M. 11n1 Donati, S. 58n47 Dutilh Novaes, C. 85n110 Ebbesen S. 3n4, 22n30, 27n49, 39n10, 55n41, 58n47, 76n83–84, 92n125, 109n1, 141n41 Fredborg, K.M. 124n5 Frede, M. 113n21 Goubier, F. 76n83, 77n84, 79n95 Gourinat, J.B. 110n6, 112n15–16, 113n17–18, 113n20 Hunt, R.W. 124n5

Isaac, J. 6n9 Kelly, L.G. 123n2, 124n4, 139n35 Kneepkens, C.H. 123n2, 128n12, 132n23, 139n33 Kretzmann, N. 11n2, 13n9–10, 14n14, 25n42, 26 Law, V. 109n3 Lewry, P.O. 42n17, 43n18 Long, A.A. 112n15 Luhtala, A. 116n30 Magee, J. 22n31, 24n40, 26n44, 27n49 Maloney, T.S. 63n56, 87n114 Marenbon, J. 2n2, 4n7, 6n10, 22n31, 41n15 Markus, R.A. 62n50 Marmo, C. 36n1, 44n19, 52n36, 63n56, 123n3, 124n4, 139n35 Martin, C.J. 43n15, 89n117 Montanari, E. 13n9 Mora-Márquez, A.M. 4n5, 5n8, 13n7–8, 39n10, 48n27, 48n29, 58n45, 66n63, 74n76, 77n84, 96n137 Noriega-Olmos, S. 13n8, 17n21 Panaccio, C. 46n24 Pasnau, R. 68n66 Pépin, J. 13n9–10 Perler, D. 36n1, 69n67 Pinborg, J. 48n27, 55n41, 123n2, 124n4, 139n35, 159, 159n5–6 Pini, G. 36n1 Pironnet, F. and J. Spruyt 76n83 Putallaz, F.X. 68n66 Putnam, H. 1, 11n3 Read, S. 77n84, 84n110 Robert, A. 58n46 Rosier-Catach, I. 6n10, 41n15, 53n38, 63n56, 66n63, 88n116, 123n2, 124n4–5, 132n23, 139n35, 140n37–38, 151n64

181

Index of Modern Authors Sedley, D. 14n14, 112n15 Shields, C. 12n6 Speaks, J. 158n2 Spruit, L. 68n66 Suto, T. 22n31, 25–26, 25n43, 26n44–45, 27n49, 29, 29n59

Tabarroni, A. 77n84 Thom, P. 104n156 Toivanen, J. 68n66 Valente, L. 6n10, 41n15 Wheeler, M. 12n4 Whitaker, C.W.A. 11n1, 13n9, 15n16, 119n39

Index of Ancient and Medieval Authors Albert the Great 37, 37n5, 49–52, 49n30–31, 50n32–33, 51n34–35, 54, 70–71, 72n74, 73n75, 75 Alexander of Aphrodisias 21, 27, 30, 32, 36–37 Alexander of Hales 53n38 Anonymus Alani 80–81, 80n97, 81n98, 81n101, 83n106, 84n107, 86n112, 88, 88n115, 93, 98–101, 99n142–144, 99n146, 100n148–149, 101n150, 103–105, 103n154, 104n155, 105n159 Anonymus Erfordensis 77n84, 78n94, 82, 82n103 Anonymus Liberanus 78n94 Anonymus Oxford 5n8, 42n17, 43n18, 44n20, 46, 72n74, 128n12, 133–134, 135n27, 136n28, 139 Apollonius Dyscolus 109n3, 110n4, 112, 114–117, 114n23–25, 115n26, 116n30–31, 118n36 Aristotle 3–5, 3n3, 11–21, 11n2, 12n4–6, 14n11–12, 15n16, 16n17–20, 17n21–22, 18n23–24, 20n26–29, 23, 23n35, 26–31, 28n53, 34, 36–38, 38n8, 39n10, 44, 48n29, 49n31, 50n32, 53, 55, 71–73, 77n84, 82–84, 82n104–105, 84n108–109, 93, 99–100, 99n145, 100n147, 104–105, 104n156, 109, 111, 112n12, 116n32, 117–121, 118n36, 119n37–40, 120n41–42, 121n43–44, 123, 136–137, 139, 139n34, 149, 159 Auctoritates Aristotelis 99n145 Augustine 6, 6n10, 25n43, 61–63, 62n50–54, 66, 162 Avicenna 57 Boethius 2–5, 3n3, 11, 13n10, 21–34, 22n31–34, 23n35–37, 24n38–41, 27n47–49, 28n50, 28n52, 28n54–56, 29n57–58, 30n60–61, 31n62–63, 32n65, 33n66–68, 34n69–70, 36–37, 37n2, 37n4, 38, 38n6, 41n15, 42, 42n16, 48, 48n28, 50n32, 70–71, 75, 87, 87n113, 105, 109, 111n8, 123n1, 133, 133n24

Boethius of Dacia 2, 6, 55n41, 80, 81n102, 86, 92–96, 92n125, 93n127–128, 94n131, 95n132, 99, 104, 104n156, 106, 140–143, 140n36–37, 141n40, 142n42–43, 143n44, 150n60, 155, 157, 161 Bonaventure 53n38 Cicero 26, 26n46 Damascenus 50n33 Diogenes Laertius 110n6, 112n15, 113n19, 113n22 Diogenes of Babylon 110n6, 112n16 Donatus 3n4, 109n3 Henry of Ghent 53n38 Incertus sf 36, 54–57, 55n41–42, 56n43, 57n44, 74 Johannes Pagus 79, 79n96 John Buridan 159–161, 159n7–10, 160n11–12 John Duns Scotus 4, 4n6, 36–39, 36n1, 38n9, 40n11–12, 53n38, 70, 73n76, 74n77–79, 75, 75n81, 78, 78n93, 80, 139 John of Dacia 44n19, 78, 78n90, 143n46 Lysias 15n15 Martin of Dacia 6, 37, 47–49, 48n27–29, 73, 73n76, 75, 137, 137n30, 139–140, 140n37–39, 141n40, 142–143, 146–149, 146n50–51, 147n52, 148n56, 149n57–58, 150n60, 155, 161 Michael of Marbais 78, 78n89, 101, 101n151 Nicholas of Paris 4, 6, 13n7, 41n13, 42n17, 71–72, 71n70–71, 72n73, 73n75, 79, 79n96, 128n12, 133–135, 133n25, 134n26 Peter of Auvergne 36, 52–54, 52n36–37, 53n39, 54n40, 55n41, 57, 139 Peter Helias 5, 124–126, 124n5, 125n6–8, 132n23, 135n27

Index of Ancient and Medieval Authors

183

Peter John Olivi 5–6, 36, 61, 63, 66–70, 66n63, 67n64, 68n65–66, 69n67–68, 70n69, 78, 78n87, 80, 95–99, 96n134– 138, 97n139, 98n140–141, 161 Peter of Spain 79, 79n95–96, 85n110, 161 Plato 14, 14n14, 112n12 Porphyry 3, 3n3, 21, 27, 27n49, 33, 37, 50 Priscian 2–5, 3n4, 38, 38n7, 53n38, 65, 106, 109–112, 109n1–3, 110n4–7, 111n8–10, 112n12–14, 114–118, 115n27, 116n28–32, 117n33–35, 121, 123, 124n5, 125–126, 130, 132n23, 139 Pseudo-Denys 53n38 Pseudo-Johannes Pagus 138–139, 138n32 Pseudo-Kilwardby 44–47, 44n19–20, 45n21–23, 46n24, 47n25–26, 49, 50n33, 77n84, 78, 78n88, 124n4, 126–127, 126n9, 127n10, 128n11–12, 129n13–15, 130–133, 130n16, 130n18, 131n19–21, 132n22, 139, 145, 155, 155n69

80, 94n129, 102, 102n152–153, 106, 137, 137n31, 140, 140n36–37, 143, 143n45, 144n47, 145n48–49, 146–148, 147n54–55, 151, 151n61–64, 152n65, 153n66–67, 154–155, 154n68, 157, 159–162 Robert Kilwardby 6, 42n17, 43, 43n18, 44n20, 46, 49, 71–72, 71n72, 133, 136, 136n29, 139 Roger Bacon 5–6, 36–37, 37n5, 61, 63–68, 63n56, 64n57–59, 65n60–61, 66n62, 70, 78, 78n86, 80, 87–92, 87n113–114, 88n116, 89n117–118, 90n119–120, 91n122–123, 92n124, 96, 99–101, 157, 161

Radulphus Brito 2, 6, 13n7, 36, 39, 39n10, 57–59, 58n45, 58n47, 60n48, 61, 61n49, 66, 68, 75–76, 76n82, 78, 78n85, 78n92,

Walter Burley 39, 39n10, 67n64 William of Ockham 158, 161 William of Sherwood 79, 161

Siger of Brabant 6, 36, 58n47, 78, 78n91, 78n94 Simon of Faversham 6, 36, 78n85, 139, 139n34 Thomas Aquinas 6, 6n9, 41n15, 49n29, 53n38, 68